{
 "name": "RevOps priors",
 "as_of": "2026-11-16",
 "sha256": "00be2bbf7233eaf7437eccaf627e8ea61068d87f0950d950a4ac2fd90b69444b",
 "count": 50,
 "about": "A maintained read on RevOps rules of thumb and whether they still hold now that AI is in the work. Not first-party proof: verdicts come from public, screened sources.",
 "how_to_read": {
  "verdict": {
   "update": "Update it",
   "keep": "Keep it",
   "too-early": "Too early to call"
  },
  "because": {
   "never-sourced": "never sourced",
   "evidence-against": "evidence against it",
   "shifting": "the mechanism has shifted",
   "still-true": "still true",
   "stronger-with-agents": "stronger with agents",
   "hype-against": "the claims against it are hype",
   "evidence-splits": "the evidence splits",
   "mostly-aspiration": "the case to update is mostly aspiration"
  },
  "premise_status": [
   "holds",
   "changed",
   "unproven",
   "fails"
  ],
  "premise_kind": [
   "causal",
   "prevalence"
  ],
  "deciding_premise": "The one premise whose status change would flip the verdict. Its status is about that premise, not the verdict: a failed challenge premise is why a prior is kept.",
  "source_standing": "authority = the originator or an independent authority; position = a named party with a stated interest, cited as their view, not as neutral evidence. Weaker sources are left out.",
  "changes": "No verdict has changed since the first edition; changed verdicts carry verdict_history entries.",
  "method": "Logic before evidence; outcomes over adoption counts; every side searched; sources screened for credibility, vendor interest named."
 },
 "in_focus": {
  "edition": "2026-11-09",
  "picks": [
   {
    "slug": "three-x-pipeline-coverage",
    "why": "Nobody can trace the 3x rule. Plan from the coverage ratio your own history implies.",
    "note": ""
   },
   {
    "slug": "pay-on-bookings",
    "why": "Where revenue arrives as usage, a signature no longer fixes it, so bookings comp needs a consumption measure there.",
    "note": ""
   },
   {
    "slug": "quota-from-capacity",
    "why": "Raising quotas for an AI lift nobody has measured turns a guess into a pay cut.",
    "note": ""
   },
   {
    "slug": "crm-is-the-record",
    "why": "The CRM stays the record of amount, stage and owner. It never held the deal itself.",
    "note": ""
   },
   {
    "slug": "price-per-seat",
    "why": "Agents break the link between users and value, so seats alone stop fitting.",
    "note": ""
   }
  ]
 },
 "priors": [
  {
   "slug": "three-x-pipeline-coverage",
   "title": "Carry 3x pipeline coverage of quota",
   "summary": "Update it: nobody can trace the 3x rule, so plan against the coverage ratio your own history implies, not a fixed multiple.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "never-sourced",
   "because_label": "never sourced",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "pipeline and forecast",
   "altitude": "L1 Axiom",
   "origin": "No traceable origin. Dave Kellogg, 2013: \"I don't know the history.\" A 2021 search for original research by Bob Apollo also came up empty. The number predates sales-force automation.",
   "rationed": "Analysis cost. Before anyone could cheaply model conversion by cohort, one ratio stood in for the whole funnel.",
   "challenge_claim": "3x is outdated. Win rates have fallen toward 20%, so teams need 5x or more, and AI makes pipeline cheap to inflate.",
   "blindspot": "Most \"update\" posts get the new number by inverting an all-opportunity win rate (1 / 19% = 5.3x). That swaps one folk ratio for another, and it is the method Kellogg rejects. The defenders miss that 3x was never derived either.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Bookings come from day-one pipeline at a steady enough rate that one ratio can stand in for the funnel.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "About a third of qualified pipeline closes, so three times quota covers the number.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "Modeling conversion by stage and cohort was too costly, so a single multiple was the practical proxy.",
      "now": "changed",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "An all-opportunity win rate measures how day-one pipeline converts.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Pipeline quality holds as volume rises, so more pipeline buys more bookings.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "One industry multiple fits your motion.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "About a third of qualified pipeline closes, so three times quota covers the number.",
    "deciding_side": "prior",
    "what_follows": "The premise that actually changed is the cost of analysis, and it points away from any fixed multiple, 3x or 5x. The 5x argument reuses the 3x logic with a different denominator, so it inherits the same flaw. What follows is a coverage target computed from your own conversion by stage and cohort."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody knows. In 2013 Dave Kellogg called 3x the rule of \"always, everywhere, every time\" and admitted he didn't know its history. He'd heard \"the rule of three\" before sales-force automation existed. Later searches for an original study turned up nothing. The number has outlived every attempt to source it."
    },
    {
     "heading": "What people say now",
     "text": "The loudest 2026 claim is that 3x is dead because win rates fell, so 5x is the new floor. The math in most of these posts is the same. They take an industry win rate (often 19%, the 2025 Ebsta and Pavilion average new-logo win rate, counted across all opportunities) and invert it. Steve Litzow, a CRO who cites the 19% without a source, puts it plainly: \"The 3× pipeline rule only works at a 33% win rate.\"\n\nKellogg reads the same data the other way. He calls it a pipeline crisis rather than a conversion crisis, arguing that teams miss because they start quarters without enough coverage. He defends tracking coverage while leaving the 3x figure aside."
    },
    {
     "heading": "Does the logic hold",
     "text": "3x rested on three ideas. One ratio could stand in for the funnel. About a third of pipeline closes. And modeling conversion properly cost more than it was worth. The second was never measured, as no study shows a third of qualified pipeline closing, or shows that it doesn't. AI changed the third because conversion by stage and cohort is now cheap to compute.\n\nThe 5x argument keeps the first idea and swaps the second. It assumes an all-opportunity win rate tells you how day-one pipeline converts. It doesn't: different denominator, different question. It also assumes pipeline quality holds as AI makes pipeline cheap to inflate, which nobody has shown. So the new number fails for the same reason the old one did. A conclusion only follows from the premise that changed, and that premise argues for computing your own ratio, not adopting a bigger one."
    },
    {
     "heading": "What both sides miss",
     "text": "Defenders miss that 3x was never derived, so \"it still works\" has nothing underneath it. Critics miss that 5x is the same move with a new input.\n\nThe sharper version is Kellogg's own, from 2013. Once a CRM could count coverage, managers pushed every rep, region and country until each showed 3x, and soon every company had it. He called the result \"a giant self-fulfilling prophecy,\" since institutionalizing a ratio means you will have that ratio, whether or not it converts. AI makes building pipeline nearly free, which makes the ratio easier to game than to read."
    }
   ],
   "open_question": "Does pipeline built with AI convert like the pipeline your coverage history was built on?",
   "sources": [
    {
     "side": "origin",
     "who": "Dave Kellogg, Kellblog",
     "date": "2013-04-19",
     "url": "https://www.kellblog.com/the-self-fulfilling-3x-pipeline-coverage-prophecy/",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Dave Kellogg, Kellblog",
     "date": "2025-11-08",
     "url": "https://www.kellblog.com/how-to-navigate-the-pipeline-crisis/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Steve Litzow, CRO NetAlly (a sales leader; his 19% is uncited)",
     "date": "2026-03-19",
     "url": "https://www.linkedin.com/posts/stevelitzow_most-sales-leaders-think-they-have-a-pipeline-activity-7440411810408710144-2xb-",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "pipeline",
    "forecast",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "forecast-from-bookings",
   "title": "Forecast revenue from bookings",
   "summary": "Update it: usage and credit revenue arrives as customers consume, not when the contract is signed, so forecast that share from consumption and keep bookings for the revenue a contract actually fixes.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "pipeline and forecast",
   "altitude": "L2 Operating model",
   "origin": "No traceable originator. It is the logic of the subscription contract: a signed deal fixes revenue for its term, so the forecast runs from pipeline to bookings to revenue, and comp is paid on the booking. Paying on bookings is a standard comp heuristic with no primary source found.",
   "rationed": "Revenue visibility. Before consumption could be metered and modeled, a signed contract was the only firm signal of what the coming year would bring.",
   "challenge_claim": "Under usage and credit pricing, revenue follows consumption, not the contract, so forecast from consumption and report usage apart from contracted ARR. AI credit pricing is pulling more software companies into that position.",
   "blindspot": "The update side credits AI with a shift that usage-priced public companies made years ago, and forgets that it covers only the usage share of revenue. The defend side misses that a bookings forecast says little about revenue that arrives month by month, and that annualizing usage into ARR hides the difference rather than forecasting it.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A signed contract fixes revenue for its term, so bookings set revenue.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Consumption was hard to measure and model, so the contract was the best signal available.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Buyers want committed, predictable spend, so most revenue stays contracted.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Usage and credit revenue follows consumption, whatever was booked.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Where consumption is metered, it forecasts usage revenue better than bookings do.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "AI pricing moves enough revenue to consumption that bookings stop mattering for most software companies.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "A signed contract fixes revenue for its term, so bookings set revenue.",
    "deciding_side": "prior",
    "what_follows": "The premise that changed is that a contract fixes revenue: for usage revenue it doesn't, and the companies that live with that forecast from consumption. That change predates AI, and it reaches only the usage share, because buyers still want predictable spend. So the prior splits rather than falls: bookings for committed revenue, consumption for usage, reported apart."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody is credited with it, because for a contracted business it was simply how the money worked. A signed subscription fixes revenue for its term. So the forecast runs from pipeline to bookings to revenue, plans are built on bookings, and reps are paid on the booking. Paying on bookings is a standard comp heuristic, and no primary source for it turned up."
    },
    {
     "heading": "What people say now",
     "text": "This is the one prior finance broadly agrees should change, and the CFOs of usage-priced companies have said so for years. Datadog's CFO: \"We continue to believe revenue is a better indication of our business trends than billings or RPO.\" Snowflake's CFO calls product revenue \"the leading indicator of our business.\" CJ Gustafson, writing in Mostly Metrics, a finance newsletter sponsored by Brex, describes Snowflake's daily consumption forecast standing in for the pipeline-to-bookings chain: \"The implication is you can't use bookings to predict revenue.\"\n\nFinance writers also want the label fixed. OnlyCFO, a pseudonymous newsletter Brex also sponsors, warns of \"deception/fraud in 'ARR' numbers,\" and a finance lead at Notion suggests calling usage run-rate revenue. Other companies go the other way and add annualized usage to ARR."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rested on three ideas. A signed contract fixes revenue for its term. Consumption was hard to measure, so the contract was the best signal anyone had. And buyers want committed, predictable spend, so most revenue stays contracted.\n\nFor usage revenue, the first idea has changed because the money arrives as customers consume, whatever was booked. The second changed for companies that meter well. The challenge's core premise, that usage revenue follows consumption, holds, and the companies built on it report on that basis.\n\nWhere the challenge overreaches is scope. The third premise still holds. Buyers still pay for predictability, and a pricing consultancy, SBI, finds more than 80% of SaaS companies still charge on some seat metric. So the conclusion that follows splits the forecast into one for what a contract fixes and another for what consumption decides."
    },
    {
     "heading": "What both sides miss",
     "text": "None of this started with AI. AI credit pricing spreads a practice that infrastructure companies settled long ago. Comp lags it. ICONIQ, a growth investor, finds about half of consumption companies still set quotas on contract value. Chris Degnan, Snowflake's former CRO, sets the bar: \"it's only when you can truly measure consumption that you earn the right to comp on it.\"\n\nThe defend side also has a stronger form than it uses. A company can keep revenue predictable by charging for committed capacity, used or not, rather than for consumption. Predictability can be chosen in pricing as well as won in forecasting."
    }
   ],
   "open_question": "When a customer buys credits up front, is that a booking, a forecast, or neither?",
   "sources": [
    {
     "side": "update",
     "who": "CJ Gustafson, Mostly Metrics (CFO newsletter sponsored by Brex), \"How Snowflake Forecasts Consumption\"",
     "date": "2026-01-06",
     "url": "https://www.mostlymetrics.com/p/how-snowflake-forecasts-consumption-based-revenue",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "David Obstler, CFO Datadog, Q1 2025 remarks",
     "date": "2025-05",
     "url": "https://investors.datadoghq.com/static-files/6019d4df-ff62-494e-be8e-d1821ec56b6f",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Snowflake CFO, Q3 FY26 earnings call",
     "date": "2025-12-03",
     "url": "https://stockanalysis.com/stocks/snow/transcripts/379491-q3-2026/",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "OnlyCFO (pseudonymous finance newsletter sponsored by Brex), \"How to Calculate ARR for AI\"",
     "date": "2026-06-04",
     "url": "https://www.onlycfo.io/p/arr-guide-for-ai",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "ICONIQ (growth investor), quoting Chris Degnan, ex-CRO Snowflake",
     "date": "2026-11-02",
     "url": "https://www.iconiq.com/growth/insights/consumption-pricing-changed-how-customers-pay-has-it-changed-how-you-operate",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "forecast",
    "usage pricing",
    "finance"
   ],
   "bench_notes": []
  },
  {
   "slug": "stage-weighted-forecast",
   "title": "Forecast by stage probability and commit categories",
   "summary": "Too early to call: everyone agrees the stage forecast is weak, but nobody has shown AI forecasting does better, and replacing the commit call drops the accountability it was for.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "pipeline and forecast",
   "altitude": "L3 Practice",
   "origin": "Salesforce's default stage-to-probability mapping (Prospecting 10% up to Negotiation 90%) and the pipeline, best case and commit categories that roll up from it. Dave Kellogg's writing set the practice: commit should be missed about once in two and a half years, and stage probabilities should come from your own history, not the defaults.",
   "rationed": "Analysis cost and activity capture. Reps were the only source of deal status, and there was no cheap way to model conversion from evidence, so a stage and a rep's call stood in for both.",
   "challenge_claim": "AI forecasting replaces the rep's story with a calculation from activity and conversation signals, so stage probabilities and weekly commit calls can go.",
   "blindspot": "The update side measures the forecast only as a prediction, but a commit is also a promise with a named owner, and a model drops the accountability. Most of the update case comes from vendors, and no independent study shows AI forecasting raising accuracy across companies. The defend side keeps default probabilities that were placeholders and never benchmarks.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Reps are the only source of deal status, so a stage and the rep's call are the best available read on a deal.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "Modeling conversion from evidence costs too much, so a fixed probability per stage stands in for it.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "A commit is a promise a named person makes, and the weekly call is where that person is held to it.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "A model built on activity and conversation signals forecasts more accurately than stage weights and the rep's call.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "A forecast is only a prediction, so accuracy is the whole test.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "The rep's call adds nothing a model can't see.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A model built on activity and conversation signals forecasts more accurately than stage weights and the rep's call.",
    "deciding_side": "challenge",
    "what_follows": "Two premises changed: deal signals no longer come only from the rep, and conversion by stage is cheap to compute from your own history. That is enough to retire default stage probabilities. It doesn't reach the commit call, because the claim that a model beats the call is unproven outside vendor work, and the call's other job, holding a named person to a promise, still holds."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Two habits, rolled into one forecast. The CRM assigns each stage a probability, from 10% at prospecting to 90% at negotiation, and weights pipeline by it. Reps also place deals in pipeline, best case or commit. Dave Kellogg's version of the practice is precise. Commit should be missed about once in two and a half years, and stage probabilities should be derived from your own history. The CRM's defaults were placeholders, not benchmarks."
    },
    {
     "heading": "What people say now",
     "text": "Almost no one defends the forecast's accuracy. Funnelcast, a forecasting vendor, studied 22 company-quarters: only 12% of commit and best-case roll-ups ever landed within 5% of the quarter's wins. It measured category roll-ups instead of stage-weighted ones.\n\nThe update side is mostly vendors and advisers. Their pages call the default stage probabilities folklore and promise a calculation in place of the rep's story. In a common middle form, the model sets the baseline, and reps explain every place they disagree with it."
    },
    {
     "heading": "Does the logic hold",
     "text": "The stage forecast rested on three ideas. Reps were the only source of deal status. Modeling conversion from evidence cost too much, so a fixed probability per stage stood in for it. And the commit call held a named person to the number. AI changed the first two. Activity and conversation signals are now captured without the rep, and conversion by stage is cheap to compute from your own history.\n\nThat is enough to retire the default probabilities, which Kellogg told teams to replace with their own history long before AI. It isn't enough to retire the commit call. That case needs a model to forecast better than the call, and the support for that comes from vendors. It also needs the forecast to be nothing but a prediction. A commit is a promise as well, one a named person makes, and a model can't be held to one. So the update follows for the stage weights and stops short of the commit."
    },
    {
     "heading": "What both sides miss",
     "text": "The update side often builds its model on the same rep-entered stages it calls unreliable. A model trained on rep-set stages inherits the reps' guesses, and no independent study shows AI forecasting tools raising accuracy across companies.\n\nDefenders miss that a stage forecast built on the default probabilities was never measured against anything. Defending it defends a placeholder."
    }
   ],
   "open_question": "On the large, rare deals where even vendors admit the model goes blind, is your commit a forecast or a promise?",
   "sources": [
    {
     "side": "origin",
     "who": "Salesforce Help, forecast category and stage mapping",
     "date": "current",
     "url": "https://help.salesforce.com/s/articleView?id=sales.faq_forecasts_category_mapping.htm",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "Dave Kellogg, Kellblog",
     "date": "2017-05-10",
     "url": "https://www.kellblog.com/how-to-train-your-vp-of-sales-to-think-about-the-forecast/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Bill Kantor, Funnelcast (forecasting vendor)",
     "date": "2025",
     "url": "https://www.funnelcast.com/post/forecast-category-considered-harmful",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "forecast",
    "pipeline",
    "crm"
   ],
   "bench_notes": [
    {
     "slug": "two-forecast-category-fields",
     "title": "Two forecast-category fields: one set from stage, one reps keep"
    }
   ]
  },
  {
   "slug": "pay-on-bookings",
   "title": "Pay reps on bookings, with accelerators above quota",
   "summary": "Update it: where contracts commit spend, signature still fixes revenue and bookings comp holds, but where revenue arrives as usage a signature no longer fixes it, so add a consumption measure there, and price in the mispricing that accelerators buy alongside the extra effort.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L2 Operating model",
   "origin": "Standard SaaS comp practice: commission is paid on booked contract value, with accelerators above quota. The Bridge Group's AE benchmarks measure it (2026: median OTE $200K, quota $960K), and ICONIQ's 2025 survey of 205 GTM executives finds accelerators in most AE plans.",
   "rationed": "Cash and predictability. In seat-based SaaS, signature was the moment a year of revenue became committed, so paying on it rewarded the event that fixed revenue.",
   "challenge_claim": "Usage pricing breaks bookings comp: a signature no longer fixes revenue, so reps should be paid on consumption.",
   "blindspot": "The update side jumps from 'bookings no longer fixes revenue' to 'consumption comp works better', which nobody has measured. The defend side counts accelerators as pure upside and leaves out what they cost in mispriced, mistimed deals.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A signed contract is the revenue the company will collect.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Accelerators above quota raise output among top performers.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Gross margins near 80% make a bookings commission cheap to pay.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Where revenue arrives as usage, a bookings plan pays for an event that no longer fixes revenue.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Paying on consumption produces better outcomes than paying on bookings.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A signed contract is the revenue the company will collect.",
    "deciding_side": "prior",
    "what_follows": "The premise that changed is the first one, and only where contracts don't commit spend: there a signature no longer fixes revenue, so a bookings-only plan pays for the wrong event. That supports adding a consumption measure for uncommitted revenue, not replacing bookings everywhere, since no one has shown consumption comp performs better. Accelerators still work, but at a measured price."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Bookings comp is the default SaaS plan rather than anyone's discovery. A rep earns commission on contract value at signature, and accelerators raise the rate once quota is passed. The Bridge Group's AE benchmarks track it: its 2026 summary puts the median AE at $200K on-target earnings against a $960K quota. ICONIQ's 2025 survey of 205 GTM executives finds accelerators in most AE plans.\n\nThe logic was tidy. In seat-based SaaS, signature was the moment a year of revenue became committed. Paying on it rewarded the event that fixed revenue."
    },
    {
     "heading": "What people say now",
     "text": "Usage pricing is the challenge. Alexander Group, a sales comp consultancy, puts consumption pricing at about 40% of technology companies in its own data, and for uncommitted or pay-as-you-go contracts it recommends paying on consumed revenue. For committed spend, bookings remain. ICONIQ's 2026 network data shows the shift under way: about half its companies still set quotas on committed ACV, and 30% now set them on forecast consumption, up 10 points in a year.\n\nThe practitioner caution comes through ICONIQ too. Chris Degnan, Snowflake's former CRO, describes layering consumption on top of bookings rather than replacing it: \"It's only when you can truly measure consumption that you earn the right to comp on it.\" A comp-software vendor adds that many AI companies run lower gross margins than classic SaaS, which would make the commission harder to afford."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on two ideas. A signed contract is the revenue the company will collect, and accelerators make top performers sell past quota. The first has changed for one kind of revenue. Where a contract commits spend, signature still fixes it. Where revenue arrives as usage, it doesn't, and a bookings-only plan pays for an event that no longer decides the number.\n\nThat supports a narrower conclusion than the challenge usually draws. It says to add a consumption measure where revenue is uncommitted. It doesn't show that paying on consumption works better, because no study compares the two on outcomes.\n\nAccelerators hold. A field study of one firm found overachievement commissions keep top performers selling past quota, and quarterly bonuses pace the laggards."
    },
    {
     "heading": "What both sides miss",
     "text": "Accelerators have a price the defenders leave out. Ian Larkin's 2014 study of one enterprise software vendor found reps timing deals around the accelerating schedule, with the extra discounting costing 6 to 8% of revenue. And in a field redesign at a medical-device firm, removing quotas and ceilings raised revenue 9%.\n\nSo the plan's shape matters as much as its basis. A team that moves to consumption comp and keeps a steep accelerator inherits the old timing problem on a new measure."
    }
   ],
   "open_question": "If you can't yet measure consumption well, what are you paying your reps for?",
   "sources": [
    {
     "side": "origin",
     "who": "Bridge Group, 2026 AE report summary (158 B2B companies)",
     "date": "2026-06-23",
     "url": "https://blog.bridgegroupinc.com/2026-ae-compensation-quota-ai-metrics",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Alexander Group (sales comp consultancy), Amoroso and Mindel, own client data",
     "date": "2026-01-20",
     "url": "https://www.alexandergroup.com/insights/tech-sales-compensation-strategies-for-consumption-based-companies/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "ICONIQ, consumption pricing insight (network survey; quotes Chris Degnan)",
     "date": "2026-11-02",
     "url": "https://www.iconiq.com/growth/insights/consumption-pricing-changed-how-customers-pay-has-it-changed-how-you-operate",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Chung, Steenburgh and Sudhir, Marketing Science (one firm's sales force)",
     "date": "2014",
     "url": "https://pubsonline.informs.org/doi/10.1287/mksc.2013.0815",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Ian Larkin, The Cost of High-Powered Incentives, Journal of Labor Economics",
     "date": "2014",
     "url": "https://www.hbs.edu/ris/Publication%20Files/13-073_cbb24c28-9e84-47d9-8a32-f01b73cfda13.pdf",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Misra and Nair, QME (field implementation at one firm)",
     "date": "2011",
     "url": "http://sanjogmisra.com/mkt_salesforce.pdf",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "comp",
    "consumption pricing",
    "accelerators"
   ],
   "bench_notes": []
  },
  {
   "slug": "training-forgotten-in-30-days",
   "title": "Reps forget 87% of sales training within 30 days, so reinforce it",
   "summary": "Update it: the 87% traces to a 1979 claim about skills, with no timeframe, from one unpublished Xerox study, and forgetting has no fixed rate, so drop the number; the advice to practise and coach after training still holds, and AI practice tools make it cheaper.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "never-sourced",
   "because_label": "never sourced",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L3 Practice",
   "origin": "Neil Rackham, \"The Coaching Controversy,\" Training and Development Journal, November 1979: a Xerox study found that without follow-up coaching 87% of the skills change a program produced was lost, while knowledge training showed much smaller loss. No timeframe. The Xerox study was never published. Later restatements turned it into 87% of training content forgotten within thirty days and credited it to Rackham's research.",
   "rationed": "Manager coaching time. Training was a classroom event and reinforcement meant a manager riding along, so the stat justified spending on scarce follow-up.",
   "challenge_claim": "AI role-play and coaching tools can supply the reinforcement managers never had time for. Their vendors usually quote the 87% as the reason to buy.",
   "blindspot": "AI coaching vendors reuse the dead number as their premise, when the case for practice stands without it. Critics who debunk the number sometimes stop there, as if a bad statistic made reinforcement optional.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Skills taught in a classroom fade unless they are practised and coached afterwards.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Reps lose training at a fixed rate: 87% within 30 days.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Reinforcement has to come from managers, whose time is scarce.",
      "now": "changed",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI coaching can do part of a manager's reinforcement work.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Gains from AI practice last and carry into live deals.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Reps lose training at a fixed rate: 87% within 30 days.",
    "deciding_side": "prior",
    "what_follows": "The number fails: forgetting varies enormously with how something was learned, and the 87% was never a rate over 30 days. The causal premise under the advice holds, and the scarcity behind it changed, since AI coaching can now carry part of the practice. So the update removes the statistic, not the reinforcement."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The number has an ancestor, and it says something different. In November 1979 Neil Rackham wrote in the *Training and Development Journal* that a Xerox study had found that, without follow-up coaching, 87% of the skills change a program produced was lost. He added that knowledge training showed much smaller loss. There was no timeframe. The Xerox study was never published; Rackham cited it with Xerox's permission.\n\nThen the wording drifted. By 2011 *The Challenger Sale* was quoted as saying Rackham's research showed 87% of training content forgotten within thirty days. A 2013 newspaper column used thirty days; a 2016 vendor post said twelve weeks and credited Xerox. Each retelling changed the timeframe and turned \"skills change\" into \"content.\""
    },
    {
     "heading": "What people say now",
     "text": "The sharpest critics are the people who traced it. In 2010, Will Thalheimer reviewed 14 studies covering 69 conditions and found forgetting ranging from 0% to 94%. His verdict: \"Rules-of-thumb that show people forgetting at some pre-defined rate are just plain false.\" A 2015 replication of Ebbinghaus's forgetting curve by Murre and Dros reproduced its shape, but with one subject learning nonsense syllables, so its percentages don't transfer to sales skills. Eubrics, itself a vendor, traced the number back and could not find a source for the version in circulation.\n\nMeanwhile AI role-play and coaching vendors use the 87% as the opening line of their pitch."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior bundles a number and a piece of advice. The advice rests on the causal idea that skills taught in a classroom fade unless they are practised and coached afterwards. That holds. Thalheimer's review shows how much learning method changes forgetting. A small 1997 study, with no control group, found training plus coaching raised productivity far more than training alone.\n\nThe number fails. It was never a rate over 30 days, and forgetting has no fixed rate to find.\n\nThe scarcity behind the advice has changed. Reinforcement used to mean a manager's time. A randomized study at two fintech firms (Luo and colleagues, *Journal of Marketing*, 2021) found an AI coach outperformed human managers, most of all for middle-ranked agents, and a combination of AI and human beat either alone. So the update removes the statistic and keeps the reinforcement."
    },
    {
     "heading": "What both sides miss",
     "text": "The same study found bottom-ranked agents overwhelmed by the AI's feedback and top-ranked agents resisting it. AI coaching is not a uniform substitute for a manager. It fits the middle of the team best.\n\nAnd no one has shown that gains from AI practice last or carry into live deals. Better rehearsal is not yet better selling. Whether it changes what happens on a real call, months later, is the measurement the field still needs, and it is a better question than any forgetting rate."
    }
   ],
   "open_question": "If your enablement case still opens with the 87%, what would it say without it?",
   "sources": [
    {
     "side": "origin",
     "who": "Neil Rackham, The Coaching Controversy, Training and Development Journal (ATD scan; cites an unpublished Xerox study)",
     "date": "1979-11",
     "url": "https://assets.td.org/m/3fc6c7940864a273/original/THE-COACHING-CONTROVERSY.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Will Thalheimer, Work-Learning Research, How Much Do People Forget? (14 studies, 69 conditions)",
     "date": "2010-12-14",
     "url": "https://www.worklearning.com/2010/12/14/how-much-do-people-forget/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Murre and Dros, PLOS ONE (one-subject replication of Ebbinghaus)",
     "date": "2015-07-06",
     "url": "https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0120644",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Eubrics (a vendor, tracing its own industry's citations), The sales training statistics nobody can source",
     "date": "undated",
     "url": "https://www.eubrics.com/blog/the-sales-training-statistics-nobody-can-source",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Olivero, Bane and Kopelman, Public Personnel Management (n=31, no control group)",
     "date": "1997",
     "url": "https://journals.sagepub.com/doi/10.1177/009102609702600403",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Luo, Qin, Fang and Qu, Journal of Marketing (randomized field experiments, two fintech firms)",
     "date": "2021",
     "url": "https://journals.sagepub.com/doi/abs/10.1177/0022242920956676",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "enablement",
    "coaching",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "selling-time-is-scarce",
   "title": "Reps spend most of their week not selling",
   "summary": "Keep it: reps still sell for well under half their week, and the claim that AI turns saved hours into selling time is mostly a pitch, because in most teams the hours leak back into admin.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "hype-against",
   "because_label": "the claims against it are hype",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L2 Operating model",
   "origin": "Salesforce's State of Sales survey: reps sell 28% of the time (2022 edition), 30% (2024). The claim against it, that AI lifts selling time from 25% to 75%, is Mark Roberge's (Stage 2 Capital, 2025 to 2026), and it is repeated across RevOps communities.",
   "rationed": "Seller hours. Admin, research and data entry ate most of the week, so any hour returned was assumed to go to buyers.",
   "challenge_claim": "AI agents take over the non-selling work, so selling time triples and output per rep doubles or more, algebraically.",
   "blindspot": "The update side assumes the hours get reinvested and that demand is unlimited; Gartner finds most sales orgs don't reinvest the time. The defend side misses that the minority who redesign the work around the saved hours do see the gains. Both sides argue from self-reported time-use surveys.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Admin, research and data entry take most of a rep's week.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "That work has to be done by the rep, by hand.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "An hour taken off that work goes to buyers.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI takes over enough of the non-selling work to triple selling time.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Hours saved are reinvested in selling.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "There are enough ready buyers to fill the extra selling hours.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "Hours saved are reinvested in selling.",
    "deciding_side": "challenge",
    "what_follows": "What changed is who does the admin: AI now saves sellers hours each week. The challenge needs those hours to reach buyers, which is the old prior's own assumption, and it fails in most teams; the demand to fill them is unproven. So reps still spend most of the week not selling."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The number comes from a vendor survey. Salesforce's State of Sales has reported for years that reps spend under a third of their week selling: 28% in the 2022 edition, 30% in 2024. Salesforce sells the automation that is meant to fix it, so the number has always come with an interest attached. It became the baseline for every argument about rep time."
    },
    {
     "heading": "What people say now",
     "text": "The claim against it is Mark Roberge's. On LinkedIn he wrote that every sales team has the opportunity to \"double selling time per AE\" with AI, and that \"doubling selling time should, algebraically, double productivity per rep.\" In interviews he puts it as selling time rising from 25% to 75% of the week, now one of the most repeated lines in RevOps.\n\nThe best data runs the other way. Gartner surveyed 210 sales leaders: AI saves sellers about 4.8 hours a week, but 72% of sales organizations show low reinvestment of that time into high-value work. Trade coverage put it plainly: the hours AI frees up are \"quietly leaking back into admin and data entry.\"\n\nSalesforce's own 2026 edition reports 40% selling time, with prospecting and data entry still counted as non-selling work. That is the largest move the series has shown, and it comes from the vendor whose tools are meant to produce it. Ebsta, which sells revenue intelligence, and Pavilion, a private community for GTM leaders, put the average seller under two hours a day of active selling."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior is a short chain. Admin, research and data entry fill most of a rep's week. Reps have to do that work themselves. So any hour taken off it was assumed to go to buyers. The first link still holds in every survey, since even Salesforce's best reading leaves most of the week to work that isn't selling. AI changed the second link. Tools now save sellers close to five hours a week.\n\nRoberge's algebra keeps the third link and leans on the second. It assumes the saved hours flow to buyers, and that buyers are waiting for them. Gartner's data breaks the hours assumption in most teams, where the time leaks back into admin. Nothing in the evidence shows the buyers are waiting. The algebra holds everything else constant, including demand, and Dave Kellogg sees a market short of pipeline, not of rep hours. So the conclusion rests on the one premise of the old prior that failed."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides argue from self-reported time-use surveys, which measure what reps say about their week, not what they did.\n\nDefenders miss the minority it works for. In the same Gartner survey, organizations that do reinvest the saved time are 2.2 times more likely to beat their customer-growth goals. The gain depends on someone redesigning the week, not on the tool."
    }
   ],
   "open_question": "What did the teams that turned AI's saved hours into selling time change about the week?",
   "sources": [
    {
     "side": "origin",
     "who": "Salesforce (sells sales automation), State of Sales (4th edition)",
     "date": "2023",
     "url": "https://www.salesforce.com/news/stories/sales-research-2023/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Mark Roberge, Stage 2 Capital",
     "date": "2026-08-21",
     "url": "https://www.linkedin.com/posts/markroberge_every-sales-team-has-the-opportunity-to-double-activity-7496581943904432128-DDgT",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Gartner, AI time savings and reinvestment survey",
     "date": "2026-05-19",
     "url": "https://www.gartner.com/en/newsroom/press-releases/2026-05-19-gartner-survey-finds-ai-saves-sellers-nearly-five-hours-per-week-yet-seventy-two-percent-of-sales-organizations-fail-to-reinvest-time-in-high-value-activities",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Salesforce (sells sales automation), State of Sales 2026 announcement",
     "date": "2026-02-03",
     "url": "https://www.salesforce.com/news/stories/state-of-sales-report-announcement-2026/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Ebsta (sells revenue intelligence) and Pavilion (private community for GTM leaders), 2025 GTM Benchmarks",
     "date": "2025",
     "url": "https://www.joinpavilion.com/hubfs/Ebsta%20x%20Pavilion%202025%20GTM%20Benchmarks%20Report.pdf",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "capacity",
    "productivity",
    "ai claims"
   ],
   "bench_notes": []
  },
  {
   "slug": "quota-from-capacity",
   "title": "Set quota from measured rep capacity, about 4 to 5x OTE",
   "summary": "Keep it: quota set from measured capacity already absorbs any real AI lift once it shows up in bookings, while raising quotas for a lift nobody has measured turns a guess into a pay cut.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "still-true",
   "because_label": "still true",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L3 Practice",
   "origin": "Quota set at about 4 to 5 times on-target earnings: Tomasz Tunguz's comp models (2013 to 2024) and the Bridge Group's AE benchmarks, which measure 4.6x in 2026. The challenge to it, raising quota by the expected AI productivity lift, appears in 2026 CRO and comp-vendor surveys.",
   "rationed": "Seller hours. Quota encoded what one person could close in a year, and comp was set as a share of it.",
   "challenge_claim": "AI makes each rep more productive, so quotas should rise now to capture the gain, and most revenue leaders already plan to raise them.",
   "blindspot": "The update side prices in a lift it hasn't measured, so a missed assumption becomes a pay cut for every rep. The defend side reads falling attainment as proof that AI doesn't work, when Bridge places the drop on a long-term downward trend.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Quota should encode what one rep has shown they can close in a year.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Capacity is measured first, and quota follows from it.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Pay is a share of quota, so a quota set too high is a pay cut.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI raises output per rep by an amount you can know before it shows up in bookings.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Pricing the lift in early costs little if it arrives late.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Higher attainment at AI-engaged companies comes from the AI.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "AI raises output per rep by an amount you can know before it shows up in bookings.",
    "deciding_side": "challenge",
    "what_follows": "Nothing under the prior has changed, and none of its premises depends on how productive reps are: a real AI lift would raise measured capacity, and quota would follow. The challenge needs the lift known before it is measured, which is unproven, and a missed assumption to be cheap, which fails because pay is a share of quota."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Quota has long been set at about four to five times a rep's on-target earnings. Tomasz Tunguz modelled it that way, and the Bridge Group measures a median of 4.6x in 2026, up from 4.2x in 2024. The ratio encoded what one person could close in a year. Capacity came first, and quota followed from it."
    },
    {
     "heading": "What people say now",
     "text": "The challenge says quotas should rise now to capture the productivity lift AI adds. BCG's 2026 survey of 398 revenue leaders found 91% expect quotas or productivity targets per seller to go up. Only 13% name win rates among AI's top measurable impacts, and only 8% name higher quota attainment. In a survey of 200 incentive-comp professionals by CaptivateIQ, which sells comp software, 43% say their companies already price AI productivity into quotas.\n\nFinance and planning voices defend the old order. The common advice is a base-case plan with no AI lift, and lift added only once it shows up in the numbers. Forma.ai, which sells sales planning software, says AI productivity \"should therefore enter the quota model through evidence.\""
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior is a chain of three ideas. Quota should encode what one rep has shown they can close. So capacity is measured first and quota follows it. And because pay is a share of quota, a quota set too high is a pay cut. None of these depends on how productive reps are, so AI doesn't touch them.\n\nThe challenge changes the order. It asks leaders to price the lift in before it is measured, and that needs two things: a lift you can predict, and a low cost of being wrong. The first is unproven; few revenue leaders can yet point to AI in their win rates or attainment. The second fails on the prior's own logic, since a missed assumption lands in every rep's pay. The challenge's conclusion only follows once the lift is measured, and at that point the old method gets there too."
    },
    {
     "heading": "What both sides miss",
     "text": "The two halves of the story are documented, but the link between them isn't. Quotas went up, and Bridge measures attainment falling to 48% of AEs from 51%. No study ties that drop to AI-priced quotas, and Bridge places it on a long-term downward trend.\n\nThe defenders' evidence has its own catch. Bridge finds 57% of reps at quota in the most AI-engaged third of companies, against 39% in the least. Bridge labels this observational data, and better teams may simply adopt more tools. One CRO's line covers the risk on both sides: \"I don't believe that if I was to double somebody's quota that I give them any chance to be successful.\""
    }
   ],
   "open_question": "Has anyone had to reset comp because the AI lift they priced into quota never arrived?",
   "sources": [
    {
     "side": "origin",
     "who": "The Bridge Group, AE Models, Motions & Metrics 2026",
     "date": "2026-06-22",
     "url": "https://www.bridgegroupinc.com/research/2026-ae-models-motions-metrics",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "BCG, How CROs Are Turning AI Into Sales Growth",
     "date": "2026-10-13",
     "url": "https://www.bcg.com/publications/2026/cros-turning-ai-into-sales-growth",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "CaptivateIQ (comp software), pricing AI into quotas survey",
     "date": "2026-05-12",
     "url": "https://www.captivateiq.com/blog/pricing-in-productivity-into-sales-quotas",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Forma.ai (sales planning software)",
     "date": "2026-07-29",
     "url": "https://www.forma.ai/resources/article/ai-sales-capacity-planning",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "quota",
    "comp",
    "capacity planning",
    "finance"
   ],
   "bench_notes": []
  },
  {
   "slug": "ae-ramp",
   "title": "A new AE ramps in three to six months",
   "summary": "Keep it: AE ramp still takes quarters, now at the long end of three to six months, and the proof that AI shortens it comes from support desks, not sales floors.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "still-true",
   "because_label": "still true",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L3 Practice",
   "origin": "The Bridge Group's AE benchmark series and the capacity models built on it, such as Tomasz Tunguz's, which ramp reps over four to nine months by seniority. Definitions of \"ramped\" vary: first deal, or sustained attainment.",
   "rationed": "Seller hours and account knowledge. A new rep had to learn the product, the buyers and the accounts by doing, and that took quarters.",
   "challenge_claim": "AI coaching, role play and instant answers compress ramp sharply, so capacity models can count new reps as productive much sooner.",
   "blindspot": "The update side borrows its strongest evidence from customer support, where AI-assisted novices caught up with six-month veterans in two. Quota-carrying sales has no equivalent study. The defend side reads the record-long ramp as AI failing, though Bridge attributes it to deal complexity and larger buying committees, and its data doesn't test AI.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A new rep learns the product, the buyers and the accounts by doing, and that takes quarters.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Ramp is paced by the deals a rep has to learn to win.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "What a new rep needs to know reaches them slowly, through colleagues and repetition.",
      "now": "changed",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Faster access to answers and coaching shortens the time to carrying a full number.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The support-desk result carries over to quota-carrying sales.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Faster onboarding means faster ramp.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Ramp is paced by the deals a rep has to learn to win.",
    "deciding_side": "prior",
    "what_follows": "What changed is how fast a new rep can get answers and practice. The premise that ramp is paced by the deals themselves holds, and Bridge ties the longer ramp to bigger, more complex deals. The challenge carries a support-desk result and onboarding speed over to ramp, and neither step has been shown in sales."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Capacity models need a ramp curve, and three to six months became the default. The Bridge Group has measured AE ramp for years, and models such as Tomasz Tunguz's ramp new reps over four to nine months depending on seniority. \"Ramped\" never had one definition, as some teams mean the first deal and others mean sustained attainment."
    },
    {
     "heading": "What people say now",
     "text": "Bridge's 2026 report puts AE ramp to full productivity at 6.2 months, the highest in the study's history, so the prior's range now sits at or past its own top end. \"Full productivity\" is Bridge's label; the page doesn't define the threshold.\n\nThe case that AI shortens ramp rests mostly on one strong study from outside sales. Brynjolfsson, Li and Raymond followed more than 5,000 support agents. With AI assistance, agents two months in performed like agents without it at six months or more, and novices gained the most. It is peer-reviewed, and it covers support work outside quota-carrying sales.\n\nInside sales the evidence is vendor-made. A benchmark co-authored by Fullcast, a planning vendor, and Pavilion, a private community for GTM leaders, reports that \"AI-enabled teams ramp 32.7% faster.\" Mindtickle, an enablement vendor, reports top companies onboarding in 25 days against a 58-day average. Both are correlational."
    },
    {
     "heading": "Does the logic hold",
     "text": "Ramp took quarters for two reasons. A new rep had to learn the product, the buyers and the accounts, and that knowledge reached them slowly, through colleagues and repetition. And the deals set the pace, because a rep can't show a full number faster than they learn to win the deals that make it up.\n\nAI changed the first reason. Answers, coaching and role play now arrive on demand. The challenge assumes this shortens the whole ramp. That holds where the job is mostly knowing the answer, which is what the support-desk study measured. In quota-carrying sales the second reason weighs more, and Bridge ties the record ramp to larger buying committees and more complex deals. The vendor numbers measure onboarding, the part AI sped up, and onboarding is not ramp. So the conclusion overreaches. AI can make a new rep's first months faster without moving the quarter they first carry a full number."
    },
    {
     "heading": "What both sides miss",
     "text": "SDR ramp is moving. Bridge measured it at 3.0 months in 2025, the lowest since 2010, and said that \"may reflect AI-powered onboarding.\" Bridge offers that as a possibility, not a finding.\n\nDefenders read the record ramp as AI failing. Bridge's data doesn't test AI, so it isn't evidence either way. AI tools could be shortening ramp underneath the drift toward bigger deals, and no one has published the cohort data to separate the two."
    }
   ],
   "open_question": "Will quota-carrying sales ever get the support-desk result, where AI let two-month novices perform like six-month veterans?",
   "sources": [
    {
     "side": "defend",
     "who": "The Bridge Group, AE Models, Motions & Metrics 2026",
     "date": "2026-06-22",
     "url": "https://www.bridgegroupinc.com/research/2026-ae-models-motions-metrics",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Brynjolfsson, Li and Raymond, Generative AI at Work (QJE)",
     "date": "2025-02-04",
     "url": "https://academic.oup.com/qje/article/140/2/889/7990658",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Fullcast (sells planning software) and Pavilion (private community for GTM leaders), 2026 GTM Benchmark",
     "date": "2026",
     "url": "https://19965450.fs1.hubspotusercontent-na2.net/hubfs/19965450/2026%20Benchmarks%20Report%20Assets/Fullcast_Pavillion-2026-GTM-Benchmark%20Report.pdf",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "The Bridge Group, SDR Models, Motions & Metrics 2025",
     "date": "2025-02-06",
     "url": "https://www.bridgegroupinc.com/research/2025-sdr-models-metrics-report-the-bridge-group",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Mindtickle (sells enablement software), State of Revenue Enablement 2025",
     "date": "2025-07-15",
     "url": "https://go.mindtickle.com/rs/270-YTL-592/images/State_of_Revenue_Enablement_2025-07-15-25%20FINAL.pdf",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "ramp",
    "enablement",
    "capacity"
   ],
   "bench_notes": []
  },
  {
   "slug": "annual-territory-carve",
   "title": "Carve and balance territories once a year",
   "summary": "Keep it: reassigning accounts still costs revenue, so a planned major realignment on the comp cycle stays right, with small boundary changes in between, and the case for continuous AI rebalancing is vendor marketing with no outcome data.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "still-true",
   "because_label": "still true",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L3 Practice",
   "origin": "Practice, not a coined rule. Alexander Group, 2013: territory design and optimization \"should be an annual process.\" The academic base is Zoltners and Lorimer (JPSSM, 2000), who found most territories badly sized and estimated good alignment adds 2 to 7% to sales. Their own later review never said only once a year: small boundary changes happen all the time, and major realignments less often.",
   "rationed": "Two things: analyst capacity, because alignment is a huge combinatorial problem often left to district managers, and customer relationships, because each reassignment puts an account at risk.",
   "challenge_claim": "AI can rebalance territories continuously in minutes, so the annual carve is a relic and territories should move whenever the data says they're out of balance.",
   "blindspot": "The update side treats analysis as the constraint, when the binding cost is the relationship broken by a reassignment. The defend side reads 'once a year' more strictly than its own source, which already expects small changes between major realignments.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Redrawing territories takes a lot of analysis.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Reassigning accounts costs revenue because it breaks customer relationships.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Comp plans and quotas run annually, so territories change on the same cycle.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Continuous AI rebalancing beats an annual redraw on revenue.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Territories drift out of balance within a year by more than a reassignment costs.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Reassigning accounts costs revenue because it breaks customer relationships.",
    "deciding_side": "prior",
    "what_follows": "Analysis cost changed, but it was never what made realignment rare. The relationship cost holds on causal evidence, and the challenge's claim that frequent rebalancing earns more than it breaks has no outcome data. So the cadence holds, and what improves is how a reassignment is handled."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody coined it. Annual territory planning grew up alongside annual comp plans. Alexander Group, a sales consultancy, put it plainly in 2013: territory design and optimization \"should be an annual process.\"\n\nThe research behind it comes from ZS. Zoltners and Lorimer found in 2000, across their work with more than 300 sales forces, that 55% of territories were too large or too small, and estimated that good alignment adds 2 to 7% to sales. But their own 2001 review doesn't say once a year. It says alignments change often, with small boundary changes going on all the time and major realignments happening less frequently. The strict version of the prior is a narrowing of its own source."
    },
    {
     "heading": "What people say now",
     "text": "The challenge comes from software vendors, which advertise redrawing territories many times faster than spreadsheets. The broader pitch is continuous, AI-driven rebalancing that moves accounts whenever the data says a territory is out of balance.\n\nNo credible independent voice makes the case. That is itself the finding. The AI-territory statistics in circulation mostly cite analyst firms without links."
    },
    {
     "heading": "Does the logic hold",
     "text": "The annual carve rested on three ideas. Redrawing took a lot of analysis. Reassigning an account costs revenue, because it breaks a relationship. And comp plans and quotas run annually, so territories follow the same cycle.\n\nThe first has changed. Optimization models have existed since the 1970s, and tools now claim minutes rather than weeks. But analysis was never what made major realignment rare.\n\nThe second holds, on causal evidence. After one optimized realignment, Sinha and Zoltners (2001) tracked about 4,500 reassigned accounts against a control. Accounts buying $50,000 to $100,000 a year bought 20% less. Shi and colleagues (2017), studying a Fortune 500 distributor, found a change of rep cut affected customers' annual sales by 13.2 to 17.6%. That study covers departures rather than realignments, but the mechanism is the same.\n\nThe challenge needs territories to drift out of balance by more than a reassignment costs, and continuous rebalancing to earn more than annual. Neither has outcome data. Cheaper analysis makes moving accounts easier; it doesn't make it cheaper for the customer."
    },
    {
     "heading": "What both sides miss",
     "text": "The relationship cost depends on how it is managed. In the same 2001 realignment, accounts above $100,000 showed no loss, because the outgoing and incoming reps shared the account and its commissions for a period. The lesson for both sides is about the handoff, not the calendar.\n\nSo the practice to keep is the one in the source: a planned major realignment on the comp cycle, small boundary changes when needed, and a shared transition for any account big enough to lose."
    }
   ],
   "open_question": "When an account changes hands, does the outgoing rep stay on it long enough to keep it?",
   "sources": [
    {
     "side": "origin",
     "who": "Alexander Group (sales consultancy), Territory design: the gateway to increased sales productivity",
     "date": "2013-02-12",
     "url": "https://www.alexandergroup.com/insights/territory-design-the-gateway-to-increased-sales-productivity/",
     "standing": "position"
    },
    {
     "side": "origin",
     "who": "Zoltners and Lorimer, JPSSM (ZS experience across 300+ sales forces)",
     "date": "2000",
     "url": "https://www.tandfonline.com/doi/abs/10.1080/08853134.2000.10754234",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Sinha and Zoltners, Interfaces, Sales-Force Decision Models: Insights from 25 Years (realignment test and control)",
     "date": "2001",
     "url": "https://operationalincomestatement.com/roi/wp-content/uploads/2013/09/ZS-SRO-25-year-review-condensed.pdf",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Shi, Sridhar, Grewal and Lilien, Journal of Marketing (Fortune 500 distributor, difference-in-differences)",
     "date": "2017-03",
     "url": "https://journals.sagepub.com/doi/10.1509/jm.15.0358",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "territories",
    "planning",
    "capacity"
   ],
   "bench_notes": []
  },
  {
   "slug": "more-reps-more-revenue",
   "title": "More reps means more revenue",
   "summary": "Too early to call: the case that AI ends headcount-led growth rests on outliers and predictions, the defence rests on budgets that still buy quota carriers, and neither side has measured what an added rep returns.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "mostly-aspiration",
   "because_label": "the case to update is mostly aspiration",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L2 Operating model",
   "origin": "No stated source; it is implicit in every capacity model. The closest canon is Neeraj Agrawal's T2D3 (Battery, 2015), which maps ARR stages to rep counts from seven public companies, and Mark Roberge's capacity model, which warned that leads and ramp are the binding constraints.",
   "rationed": "Seller hours. Each rep could work only so many accounts, so revenue scaled with the number of people selling.",
   "challenge_claim": "AI raises output per rep so much that companies will grow revenue without adding sellers. AI-native outliers run several times the revenue per employee of classic SaaS, so headcount stops being the lever.",
   "blindspot": "The update side benchmarks against outliers; at the median, AI startups run lower revenue per employee than non-AI ones at every stage SVB measured. The defend side misses that adding heads into a volume strategy can lower revenue per seller. Both sides skip the demand term: reps only convert pipeline that exists.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Seller hours are the binding constraint, so revenue scales with the number of people selling.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "There is enough demand to fill each added rep's hours.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI raises bookings per rep at typical companies, not only at the fastest-growing ones.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "AI-native outliers show the revenue per employee the median company will reach.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Once output per rep rises, adding sellers stops being the way to add revenue.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "AI raises bookings per rep at typical companies, not only at the fastest-growing ones.",
    "deciding_side": "challenge",
    "what_follows": "Neither side has outcome evidence for its core premise. That seller hours bind rests on budgets still buying quota carriers, which shows what companies do, not what added reps return. The challenge's deciding premise, that AI raises bookings per rep at the median company, is unproven, and the outlier evidence for it fails at the median. Both sides also lean on a demand premise that neither has measured."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody wrote it down as a law. It sits inside every capacity model as ramped reps times quota times attainment. Battery's T2D3 path, built from seven public companies, mapped each ARR stage to a rep count. Mark Roberge, who built one of the best-known capacity models, still says the math \"almost never holds,\" for the pre-AI reasons of ramp, demand and product fit."
    },
    {
     "heading": "What people say now",
     "text": "The update case leans on AI-native companies. Bessemer, a venture firm that invests in AI startups, reports its fastest cohort at about $1.13M of ARR per employee in the first year. Gartner now says sales productivity gains \"remain constrained by operating models designed to scale primarily by adding more people.\"\n\nThe defence comes from the people signing the budgets. Salesforce, which sells AI agents, told investors that \"supply is capacity\" and that it still has 1,200 more account executives to hire. Gartner's CFO survey has total headcount growth falling for 2026, yet sales still gets one of the largest budget increases."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on two ideas. Seller hours are the scarce input, so each added rep adds roughly a rep's worth of bookings. And there is enough demand to fill those hours. Every capacity model encodes the first and simply assumes the second.\n\nThe challenge needs AI to raise output per rep so far that the same team carries more revenue, and its evidence is the AI-native outliers. The outliers are not the median. SVB, a bank whose clients are venture-backed startups, publishes data showing the median AI startup with *lower* revenue per employee than non-AI peers at every stage from seed to Series D, so the premise that outliers show where everyone is heading fails. The broader premise, that AI lifts bookings per rep at an ordinary company, is still a forecast.\n\nThe prior's own first premise is on no firmer ground. The evidence that seller hours still bind is that budgets still buy quota carriers, and that describes what companies do, not what an added rep returns. So neither side has shown its core premise, and the premise that would settle the question, demand, goes unmeasured by both camps."
    },
    {
     "heading": "What both sides miss",
     "text": "Defenders miss what happens when the added heads go into a volume strategy. Pavilion's CEO Sam Jacobs, summarising the benchmark Pavilion co-published with Fullcast, a planning vendor, describes a cohort that added 38% more sellers and pointed AI at top-of-funnel volume. Its revenue per seller fell 26%. That figure is correlational and the dataset is vendor co-authored, but it cuts against both camps.\n\nNeither side has shown the reversal. No named GTM company has publicly undone an AI-justified sales headcount cut. The cleanest reversal on record, Klarna, was in customer support."
    }
   ],
   "open_question": "Outside the fastest-growing companies, has AI raised bookings per rep anywhere yet, or only the forecasts of it?",
   "sources": [
    {
     "side": "origin",
     "who": "Neeraj Agrawal, Battery Ventures (T2D3)",
     "date": "2015-03-21",
     "url": "https://www.battery.com/blog/helping-entrepreneurs-triple-triple-double-double-double-to-a-billion-dollar-company/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Bessemer Venture Partners (invests in AI startups), The State of AI 2025",
     "date": "2025-08-13",
     "url": "https://www.bvp.com/atlas/the-state-of-ai-2025",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Gartner, AI time savings and reinvestment survey",
     "date": "2026-05-19",
     "url": "https://www.gartner.com/en/newsroom/press-releases/2026-05-19-gartner-survey-finds-ai-saves-sellers-nearly-five-hours-per-week-yet-seventy-two-percent-of-sales-organizations-fail-to-reinvest-time-in-high-value-activities",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Silicon Valley Bank (banks venture-backed startups; PitchBook data), State of the Markets H2 2025",
     "date": "2025-08",
     "url": "https://www.svb.com/globalassets/trendsandinsights/reports/sotm/state-of-the-markets-h2-2025-final.pdf",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Miguel Milano, Salesforce (Q2 FY27 earnings call)",
     "date": "2026-08-26",
     "url": "https://s205.q4cdn.com/626266368/files/doc_financials/2027/q2/Salesforce-Q2-FY27-Earnings-Transcript.pdf",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Sam Jacobs, CEO Pavilion (relaying the benchmark Pavilion co-published with Fullcast)",
     "date": "2026-03-07",
     "url": "https://www.linkedin.com/posts/samfjacobs_ai-isnt-creating-one-future-for-go-to-market-activity-7436066887387602944-kU_t",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "capacity",
    "headcount",
    "planning"
   ],
   "bench_notes": []
  },
  {
   "slug": "revops-ratio",
   "title": "Staff one ops person for every 12 to 15 reps",
   "summary": "Too early to call: the ratio was always a description of what teams run, never a tested staffing rule, and the claim that GTM engineers shrink it rests on job posts and title counts, not output.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "mostly-aspiration",
   "because_label": "the case to update is mostly aspiration",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "capacity and comp",
   "altitude": "L2 Operating model",
   "origin": "SellingBrew / MindBrew Research Team, 2017: \"most successful teams have one sales ops person for every ten to fifteen salespeople,\" heard anecdotally and confirmed with a member survey whose method isn't given. Adam Schoenfeld's PeerSignal measured it in 2024 on org data for 2,500 B2B software companies: about 12 reps (AE plus SDR) per RevOps person.",
   "rationed": "Ops analyst hours. CRM administration, territories, quotas, comp and reporting were largely manual, and the work grew with the number of sellers and plans.",
   "challenge_claim": "GTM engineers automate the ops work, so one technical hire covers what several ops people did and the ratio falls.",
   "blindspot": "Both sides argue from prevalence. Defenders cite a ratio that describes what teams run and never showed it drives results; critics cite job posts and title counts, which say what's being hired, not what it produces.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Ops headcount has to grow with the number of sellers.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Most teams run about one ops person for every 12 reps.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "GTM engineers do the work of several ops roles, so fewer ops people can support the same sellers.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "GTM engineer roles are growing fast from a small base.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "Ops headcount has to grow with the number of sellers.",
    "deciding_side": "prior",
    "what_follows": "Nothing causal has changed, because nothing causal was ever established: the ratio describes practice, and whether ops headcount must track sellers was never tested. The challenge rests on hiring data, which shows a new title growing, not a new output per head. Until someone measures output, neither ratio is a staffing rule."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The ratio started as an anecdotal rule of thumb of one sales ops person for every ten to fifteen salespeople, with no published method behind it.\n\nAdam Schoenfeld's PeerSignal measured it properly in 2024, on org data for 2,500 B2B software companies: about 12 reps per RevOps person overall, with fewer ops people per seller above 1,000 employees and \"wide variance.\" PeerSignal doesn't tie the ratio to results. Like the rule of thumb, it describes what teams run."
    },
    {
     "heading": "What people say now",
     "text": "The challenge comes from the rise of the GTM engineer, a technical role that builds the automation ops people used to do by hand. James McKay, an agency founder writing for Pavilion, argues GTM engineers \"are eating RevOps' lunch,\" citing falling RevOps job ads. Cristian Frunze, publishing his company Kenso's own profile data, counts 121 people with the title at the end of 2023 and 791 by March 2026, while sales and marketing headcount stayed flat.\n\nThe scale is still small. Kyle Poyar, using job-post data, found one GTM engineer posting for every 14 RevOps postings and every 35 sales ops postings."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on the causal idea that ops work scales with the number of sellers, so ops headcount has to grow with them. That was plausible when territories, quotas, comp and reporting were mostly manual. But it was never tested. PeerSignal shows ops and sellers rise together across companies, with wide variance, which is correlation across firms, not a staffing law. The ratio is a fair description of practice and nothing more.\n\nThe challenge needs the reverse causal idea that one GTM engineer does the work of several ops roles, so the same sellers need fewer ops people. The evidence for it is hiring data. A title growing fast from a small base tells you what companies are trying while saying nothing about what each hire produces. No study compares output per ops head with and without GTM engineers. So the challenge rests on an unproven premise, and so does the prior it challenges."
    },
    {
     "heading": "What both sides miss",
     "text": "Neither side measures the thing a staffing ratio is for. A useful number would tie ops capacity to something a revenue leader cares about, such as plan changes shipped, data quality, or time to a reliable forecast. Without that, 1:12 and any leaner ratio are both descriptions of what someone chose to hire.\n\nThe same assumption that the work follows headcount sits under \"more reps, more revenue.\" When automation changes how much work each seller generates for ops, the ratio stops being a useful shortcut either way, and a team has to count the work itself."
    }
   ],
   "open_question": "If your ops team shrank by a third tomorrow, what would break first, and would a GTM engineer fix it?",
   "sources": [
    {
     "side": "defend",
     "who": "Adam Schoenfeld, PeerSignal (Keyplay org data, 2,500 companies)",
     "date": "2024-06-01",
     "url": "https://www.peersignal.org/p/revops-benchmarks",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Kyle Poyar, Growth Unhinged, Do you need a GTM engineer? (Sumble job-post data)",
     "date": "2025-09-10",
     "url": "https://www.growthunhinged.com/p/do-you-need-a-gtm-engineer",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "James McKay (agency founder), Pavilion blog, GTM Engineers Are Eating RevOps' Lunch",
     "date": "2025-07-29",
     "url": "https://www.joinpavilion.com/blog/gtm-engineers-are-eating-revops-lunch",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Cristian Frunze, founder of Kenso (publishing Kenso's own profile data), GTM engineer origins",
     "date": "2026-10-19",
     "url": "https://ken.so/research/rise-of-the-gtm-engineer",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "revops",
    "headcount",
    "gtm engineering"
   ],
   "bench_notes": []
  },
  {
   "slug": "touches-and-volume",
   "title": "It takes 8 to 12 touches, and volume is a lever",
   "summary": "Update it: the 8-to-12 touch count was never sourced and cheap volume has stopped working, but fewer, more relevant touches still do, so revise the cadence rather than drop outbound.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L3 Practice",
   "origin": "Sales-training folklore. \"8 to 12 touches\" and \"80% of sales need five follow-ups\" have no study behind them that anyone has found; the closest attribution, to RAIN Group, appears only in secondary sources. The Bridge Group measures an average SDR cadence of 11.3 attempts, which describes what teams run, not what works.",
   "rationed": "Writing cost. Each message took a person's time, so a sequence signalled effort, and more sends meant more reach at a known reply rate.",
   "challenge_claim": "AI makes personalised volume free, so teams can send more, better-tailored messages and lift replies, or, on the other side, outbound is dead.",
   "blindspot": "Volume-first teams miss that free writing removed the effort signal and that mailbox providers now penalise bulk senders. \"Outbound is dead\" posts miss that the consensus is narrower: cheap volume stopped working, fewer and more relevant touches still do. Both sides quote reply-rate trends that no single source has measured on a fixed denominator.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Each message cost a person's time, so a long sequence signalled effort.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "More sends bought more reach at a reply rate you could count on.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Replies keep arriving deep into a sequence, so eight to twelve steps pay off.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI-written personalisation lifts replies at volume.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Reply rates have collapsed, so outbound no longer works.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "A short, relevant sequence gets most of the replies a long one would.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "deciding_premise": "More sends bought more reach at a reply rate you could count on.",
    "deciding_side": "prior",
    "what_follows": "Two premises changed at once: writing became free, which removed the effort signal, and mailbox rules began to penalise bulk sending. Both point to fewer, more relevant touches. \"Send more with AI\" reuses the old volume logic with a cheaper writer, and \"outbound is dead\" rests on a collapse nobody has measured on a fixed denominator."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody can find the study. \"Eight to twelve touches\" and \"80% of sales need five follow-ups\" circulate in sales training, and the closest attribution reaches a research firm only through secondary blogs. The Bridge Group, which sells sales-development consulting, measures an average of 11.3 attempts per SDR cadence. That describes what teams run. It says nothing about what works."
    },
    {
     "heading": "What people say now",
     "text": "Vendor blogs now argue over whether outbound is dead, and the narrower point they keep landing on is that cheap volume stopped working. The strongest update comes from senders' own data, some of it against their interest. Sales.co, which sells outbound, tested AI-written first lines across 1.28 million contacts: AI personalisation replied at 0.53%, and a plain template with no personal line at 1.00%. Hunter, which sells outreach tools, looked at 31 million emails sent in 2025 and found three emails the best sequence length: 6.8% replied, against 3.3% for a single email.\n\nThe rules changed too. Since February 2024, Gmail has required senders of more than 5,000 messages a day to authenticate, offer one-click unsubscribe and keep reported spam below 0.3%."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior made sense when writing cost something. A person wrote each message, so a long sequence signalled effort, and more sends bought more reach at a reply rate you could count on. The touch count was never sourced, but the logic behind volume was sound.\n\nTwo of those premises have moved. Free writing removed the effort signal. Gmail's rules mean volume can now cost you placement instead of buying reach. And the step data, all of it from vendors, puts most replies early. In data from Instantly, which sells cold-email sending software, 58% of replies come from the first step. If that holds in your data, steps nine through twelve were never doing much.\n\nThe two loud updates draw opposite conclusions from the same change. \"Send more with AI\" keeps the old volume premise and swaps in a cheaper writer, and the template test above shows what that buys. \"Outbound is dead\" needs a collapse in reply rates that the data doesn't show. What follows is the narrower claim: fewer, more relevant touches."
    },
    {
     "heading": "What both sides miss",
     "text": "The collapse story is weaker than it sounds. Instantly, whose figure is the one most quoted, reports a 3.43% average reply rate and calls it \"stable despite growing volume.\" Published 2025 reply rates span roughly tenfold depending on the denominator. Nobody has a multi-year series from one source on a fixed definition."
    }
   ],
   "open_question": "Do fewer, more relevant touches bring reply rates back, or has the channel settled at a lower floor?",
   "sources": [
    {
     "side": "origin",
     "who": "The Bridge Group (sells sales-development consulting), 2023 Sales Development Report",
     "date": "2023",
     "url": "https://blog.bridgegroupinc.com/hubfs/resources/TBG%202023%20Sales%20Development%20Report.pdf",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Ryan Doyle, Sales.co (sending-tool vendor)",
     "date": "2026-09-27",
     "url": "https://sales.co/research/personalisation-no-longer-matters",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Instantly (sells cold-email sending software), Cold Email Benchmark Report 2026",
     "date": "2026-01-12",
     "url": "https://instantly.ai/cold-email-benchmark-report-2026",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Google, Gmail email sender guidelines",
     "date": "2024-02-01",
     "url": "https://support.google.com/a/answer/81126",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Hunter (outreach-tool vendor), The State of Email Outreach",
     "date": "2025",
     "url": "https://hunter.io/the-state-of-cold-email",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "outbound",
    "cadence",
    "deliverability"
   ],
   "bench_notes": []
  },
  {
   "slug": "buyers-70-percent-through",
   "title": "Buyers are 70% through the journey before they talk to sales",
   "summary": "Update it: buyers do arrive later and with AI-built shortlists, but the 70% figure stretches a 2012 study past what it measured, so keep the direction and drop the number.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "CEB's Marketing Leadership Council with Google, The Digital Evolution in B2B Marketing (2012): buyers are 57% of the way through a purchase before first serious engagement with sales. The respondents were customers of 22 large suppliers, and 70% was the top of the range, not the average.",
   "rationed": "The seller's information advantage. Buyers had to talk to a rep to learn pricing, fit and how a product compared.",
   "challenge_claim": "Buyers now research inside AI assistants and arrive with a shortlist already ranked, so the first sales conversation comes later than ever and the seller's job shifts to validating.",
   "blindspot": "The update side recycles the 70% figure as if it were a measured average, and uses it to sell replacements for the MQL. The defend side, which says big shifts come from buyer behaviour and not seller tools, underrates how much buyer research has moved into AI. Both read stated preferences in surveys as behaviour.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Buyers had to talk to a rep to learn pricing, fit and how products compared.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "Buyers do most of their research before they engage sales.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "The share done before sales can be pinned at about 70%.",
      "now": "fails",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Buyers now research inside AI assistants and arrive with a ranked shortlist.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "The first sales conversation comes later than ever.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "The seller's job shifts to validating what the buyer found.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "The share done before sales can be pinned at about 70%.",
    "deciding_side": "prior",
    "what_follows": "The seller's information advantage is the premise that changed: buyers can now learn fit and comparisons from AI, which strengthens the direction of the prior. The number never had a measurement under it, and \"later than ever\" rests on one vendor's data, which shows the seller's share of the journey rising. Keep the direction, drop the number."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "A real study, retold beyond what it found. In 2012 CEB and Google reported that business buyers were 57% of the way through a purchase before their first serious engagement with sales. The respondents were customers of only 22 large suppliers. The study measured first serious engagement, which differs from first contact, and 70% was the top of its range. The 70% version is the one that spread."
    },
    {
     "heading": "What people say now",
     "text": "The direction has strengthened. Forrester's 2026 buying study has 94% of business buyers using AI somewhere in the purchase, and twice as many as a year earlier name generative AI as more meaningful than any other source. 6sense, which sells intent data, reports that 94% of buying groups rank vendors before first contact, and 77% buy from the vendor they preferred at the end of their selection phase. In its data the buyer's research-to-seller split moved from 70/30 to 60/40, and buyers still averaged 16 interactions with the winning vendor.\n\nThe rep's role is shifting rather than vanishing. Gartner found that 69% of buyers prefer to validate AI-generated insights with a sales rep, even as 67% say they'd prefer a rep-free experience."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on the seller losing an information advantage. Buyers once had to talk to a rep to learn pricing, fit and how products compared, and the prior says they now do most of that on their own first. The 2012 study measured that direction. It did not measure 70%. The average was 57%, and 70% was the top of the range.\n\nAI pushes that premise further. Buyers can now ask an assistant what they used to ask a rep, and most buying groups rank vendors before anyone calls. That strengthens the direction. It doesn't make \"later than ever\" follow. In 6sense's own data the research-to-seller split moved toward sellers. So the AI-era claim carries forward the prior's direction, which holds, and its number, which never did. What follows is the direction with a new mechanism and no fixed figure."
    },
    {
     "heading": "What both sides miss",
     "text": "The stat is now a tool for selling the replacement. One software CEO's case for killing the MQL opens: \"By the time an MQL hits your CRM, the buyer is already 70% through their journey.\" The number carries more weight in that sentence than the 2012 study can hold.\n\nDefenders who say buyer behaviour hasn't really changed miss the AI data. Updaters miss that Gartner's figures, like 6sense's, are buyers' stated preferences in a survey, and stated preferences don't always match behaviour. The same buyers who say they want no rep still have 16 interactions with the one they pick."
    }
   ],
   "open_question": "How do you see, let alone credit, a shortlist your buyer built inside an AI chat your systems never touch?",
   "sources": [
    {
     "side": "origin",
     "who": "CEB Marketing Leadership Council and Google, The Digital Evolution in B2B Marketing",
     "date": "2012",
     "url": "https://www.thinkwithgoogle.com/_qs/documents/3928/b2b-digital-evolution_articles_1.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Forrester (John Buten), B2B Buyers Make Zero-Click Buying Number One",
     "date": "2026-01-22",
     "url": "https://www.forrester.com/blogs/b2b_buyers_make_zero_click_buying_number_one/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "6sense (intent-data vendor), 2025 Buyer Experience Report",
     "date": "2025-11-12",
     "url": "https://6sense.com/news/the-timeline-for-influencing-b2b-buyers-is-shrinking-insights-from-6senses-2025-buyer-experience-report",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Gartner, buyers validating AI-generated insights",
     "date": "2026-05-20",
     "url": "https://www.gartner.com/en/newsroom/press-releases/2026-05-20-gartner-survey-finds-sixty-nine-percent-of-b-two-b-buyers-turn-to-sales-reps-to-validate-ai-generated-insights",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "demand",
    "buyer journey",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "buying-committee-six-to-ten",
   "title": "A buying committee is six to ten decision makers",
   "summary": "Update it: groups with veto power still decide B2B purchases, but no single size holds across purchases, so size the group to the purchase, count sign-offs and influencers separately, and work for consensus rather than headcount.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "Gartner's undated sales page \"The B2B buying journey\": \"The typical buying group for a complex B2B solution involves six to 10 decision makers.\" It gives no sample or method; a 2020 archive capture carries the line and a 2024 capture of the same page no longer does. The nearest sourced ancestor is CEB (bought by Gartner in 2017): Schmidt, Adamson and Bird, HBR 2015, found on average 5.4 people formally signing off on each purchase.",
   "rationed": "Seller attention and access. A rep can work only a few relationships per deal, so the number told reps how far to widen beyond a single champion and told marketing how many personas to build for.",
   "challenge_claim": "Committees keep growing, and AI purchases pull in groups twice the usual size, so sellers have to thread wider, and AI makes each extra thread cheap enough to do it.",
   "blindspot": "Both sides argue about headcount. Gartner, whose page made six to ten famous, now finds that consensus, not size, predicts a good deal, and the multi-threading data that gives the number its use shows association, not cause.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Purchases are decided by groups with veto power, not by one executive.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "One typical committee size holds across purchases, so it is worth planning around.",
      "now": "fails",
      "kind": "prevalence"
     },
     {
      "premise": "Reaching more of the group raises the odds of winning.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI-era purchases have larger committees.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "Committees are shrinking as IT consolidates the decision.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "AI lowers the cost of each extra thread, so sellers should go wider.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "One typical committee size holds across purchases, so it is worth planning around.",
    "deciding_side": "prior",
    "what_follows": "The causal core holds: groups with veto power decide, and getting them to agree is the job. What fails is the idea of one typical size, which research has tied to the purchase for decades. Both challenges swap one number for another, so what follows is no fixed range at all: size the group to the purchase and count sign-offs and influencers separately."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The line most people quote comes from a Gartner sales page, \"The B2B buying journey\": \"The typical buying group for a complex B2B solution involves six to 10 decision makers.\" The page gives no date, sample or method. An archived capture from April 2020 carries the line. A capture of the same page from May 2024 no longer does.\n\nThe nearest sourced ancestor is CEB, which Gartner bought in 2017. In HBR in 2015, Schmidt, Adamson and Bird drew on four CEB surveys of more than 5,000 stakeholders. On average, 5.4 people formally signed off on each purchase, and the likelihood of purchase fell from 81% with one decision maker to 31% with six."
    },
    {
     "heading": "What people say now",
     "text": "Current counts disagree in both directions. Forrester's 2026 survey of about 18,000 buyers finds generative AI purchases involve 14 members against 7 for other purchases. G2, surveying its own software buyers, says committees of five to eight are giving way to three or four. Gartner itself now says five to 16 people across as many as four functions.\n\nMulti-threading is the practice built on the number. If the group is six to ten, reach more of them. The data behind it comes from vendors' own customers. Gong reports won deals have about twice the buyer contacts. Airspeed, across 133,068 closed deals at 116 companies, finds win rates rising from 24.3% with one contact to 46.4% with five or more, and the pattern survives when it holds the number of calls constant."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on the two ideas that groups with veto power decide purchases and that one typical size holds across them. Multi-threading adds a third idea, that reaching more of the group wins more often.\n\nThe first holds. Gartner's 2025 survey of 632 buyers found 74% of buyer teams in unhealthy conflict, and groups that reached consensus were 2.5 times as likely to report a high-quality deal. The second fails. Academic work has tied size to the purchase for decades. A 2005 meta-analysis found the buying center grows with the importance, complexity and uncertainty of the purchase, and a 1992 study put the mean near four. Today's counts disagree even within one firm, and they count different things, including sign-offs, decision makers and influencers.\n\nThe challenges don't repair this. \"Bigger now\" holds for AI purchases, on one analyst survey. \"Smaller now\" rests on a vendor survey with no definition of a member. Both replace one number with another. What follows is to size the group to the purchase and count who signs separately from who influences."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides argue about headcount, while Gartner, the firm whose page made six to ten famous, now says agreement is what predicts the deal. In the same survey, content tailored to the individual had a 59% negative effect on group consensus, and content relevant to the group had a 20% positive one. That cuts against the easy AI version of multi-threading, where cheap threads mean you should go wider. No outcome study tests it.\n\nThe multi-threading data is association. Healthy deals may pull in colleagues rather than the reverse, and Ebsta's vendor data finds win rates rising with relationships only up to a point, then falling sharply past 16. \"Reach the group\" holds up. \"More threads cause wins\" is unproven."
    }
   ],
   "open_question": "When a deal stalls, is the group too small, or has it just not agreed?",
   "sources": [
    {
     "side": "origin",
     "who": "Gartner, \"The B2B buying journey\" sales page (undated; Wayback capture)",
     "date": "2020-04-21",
     "url": "http://web.archive.org/web/20200421105517/https://www.gartner.com/en/sales/insights/b2b-buying-journey",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "Schmidt, Adamson and Bird (CEB), HBR, Making the Consensus Sale",
     "date": "2015-03",
     "url": "https://hbr.org/2015/03/making-the-consensus-sale",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Barbara Winters, Forrester, State of Business Buying 2026",
     "date": "2026-01-21",
     "url": "https://www.forrester.com/blogs/state-of-business-buying-2026/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Tim Sanders, G2 Buyer Behavior Report (G2's own buyer survey)",
     "date": "2025-05-14",
     "url": "https://company.g2.com/news/buyer-behavior-in-2025",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Delainey Kirkwood, Gartner sales survey press release",
     "date": "2025-05-07",
     "url": "https://www.gartner.com/en/newsroom/press-releases/2025-05-07-gartner-sales-survey-finds-74-percent-of-b2b-buyer-teams-demonstrate-unhealthy-conflict-during-the-decision-process",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Helene Dötsch, Airspeed (vendor, its customers' deal data)",
     "date": "2026-09-01",
     "url": "https://www.goairspeed.com/blog/what-133068-closed-deals-say",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "buying committee",
    "multi-threading",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "gate-content-for-leads",
   "title": "Gate your best content to generate leads",
   "summary": "Update it: the gate trades reach for names, and the names are mostly weak leads; now the gate also hides the asset from the answer engines where buyers start their research, so ungate the summary and gate only the deep asset, if anything.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "evidence-against",
   "because_label": "evidence against it",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L3 Practice",
   "origin": "No single originator; gating whitepapers for names predates the term content marketing. The canonical public statement is the 2011 debate between David Meerman Scott and Mike Volpe, then VP of marketing at HubSpot. Volpe: marketers have lead goals, so \"a business has to use forms to generate leads.\"",
   "rationed": "Visibility into who was reading. Before account identification and intent data, a form was the only way to learn a reader's name, and sales needed a lead count to work against.",
   "challenge_claim": "Ungate it: forms collect false or useless contacts, and answer engines can't read behind a form, so a gated asset is invisible where buyers now start their research.",
   "blindspot": "The update side's best before-and-after is one company whose ABM program and measurement changed at the same time, and no one has shown that being cited in AI answers produces pipeline. The defend side counts how many teams gate and how many leads they say it brings, which says nothing about whether those leads buy.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A form is the only way to know who is engaging with your content.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "People who fill in the form are leads worth a sales follow-up.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "The reach lost to the gate costs less than the leads it captures.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Answer engines cannot read gated content, so a gated asset is invisible where buyers now research.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Buyers increasingly start their research in AI assistants.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "Being cited in AI answers produces pipeline.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "People who fill in the form are leads worth a sales follow-up.",
    "deciding_side": "prior",
    "what_follows": "The verdict rests on a premise that failed before AI: the people a gate captures are mostly weak leads, so the names were never worth the reach. AI adds a new, mechanical cost, since a crawler can't read behind a form, but whether being read by answer engines produces pipeline is unproven, so it raises the price of the gate without deciding the case."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Gating whitepapers for names predates the phrase content marketing, and no one coined it. Its canonical public statement is a 2011 debate between David Meerman Scott, a marketing author who has advised HubSpot since 2007, and Mike Volpe, then VP of marketing at HubSpot. Volpe argued that marketers have lead goals, so \"a business has to use forms to generate leads.\"\n\nScott answered that ungated content gets 20 to 50 times the downloads. The only data he published behind it is one 2009 example from AIIM: 5,138 ungated ebook downloads in a month against 513 for gated research, about 16 times. The wider range is his claim, not a finding."
    },
    {
     "heading": "What people say now",
     "text": "Forrester has argued against blanket gating since 2016. It called gating a dating game that pushes buyers to submit false information and yields \"too many useless leads.\" Its 2019 advice was to gate only high-value assets such as analyst reports, never case studies.\n\nThe newer argument is mechanical. Answer engines don't fill in forms, so a gated asset can't be read or cited where buyers now research. A Forrester analyst put it this way in 2025: content that isn't crawlable is \"not in the conversation.\"\n\nThe defense is mostly adoption. PathFactory, a vendor, found 83% of marketers gate and 65% say gating brings in a significant share of their leads. That describes how common the practice is without showing whether it pays."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on three ideas. A form was the only way to learn who was reading. The people who fill it in are leads worth a sales call. And the reach the gate costs is worth less than the names it captures.\n\nThe first has changed because account identification and intent tools now see engagement without a form. The second fails. Forrester found gates yield false details, and HockeyStack, a vendor comparing 87 companies on its own data, found content leads convert to sales-qualified at under a quarter of the rate of hand-raises, a gap the entry on the MQL covers. The third is barely tested.\n\nThe answer-engine challenge holds where it is mechanical, since a crawler can't read behind a form and Forrester's 2026 survey finds 94% of buyers using AI in buying. Whether being cited produces pipeline is unproven. So the update rests on the failed second premise, which predates AI. AI raises the price of the gate. It doesn't decide the case."
    },
    {
     "heading": "What both sides miss",
     "text": "The strongest ungating story is one company, told by its CMO. Chuck Moxley's team at Blue Triangle, a digital-experience software company, spent $15,000 on a gated campaign that drew 3,400 visits and 106 forms; ungated, the same ebook reached 175 engaged readers on $2,000. Half the gated emails were bad. A year after ungating everything, demo requests were up 265% and pipeline up 242%. But an ABM program and new measurement started at the same time, so it isn't a test. No controlled comparison of gated and ungated content on pipeline exists.\n\nDefenders miss that the form always traded reach for names, and the names were weak. What survives is a narrow rule to ungate the summary and the main findings while gating the deep asset. Forrester lands there, and so does Carolyn Shelby, an SEO consultant at Yoast, which sells SEO software."
    }
   ],
   "open_question": "If answer engines are where buyers read first, what is a gate still protecting?",
   "sources": [
    {
     "side": "origin",
     "who": "David Meerman Scott (marketing author and speaker, HubSpot adviser) vs Mike Volpe (HubSpot, sells marketing software), debate on Scott's blog",
     "date": "2011-01-19",
     "url": "https://www.davidmeermanscott.com/blog/2011/01/debate-totally-free-content-vs-requiring-registration.html",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Steven Casey, Shanta Samlal-Fadelle et al., Forrester, Unlock Content Gates",
     "date": "2016-08-31",
     "url": "https://www.forrester.com/report/unlock-content-gates-to-support-self-educating-buyers/RES134301",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Lisa Gately, Forrester",
     "date": "2025-05-19",
     "url": "https://www.forrester.com/blogs/shifting-search-behaviors-demand-smarter-content-strategies/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Ann Gynn on Chuck Moxley (CMO of Blue Triangle, reporting his own team's results), Content Marketing Institute",
     "date": "2026-03-04",
     "url": "https://contentmarketinginstitute.com/content-marketing-strategy/cmo-ungated-everything",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Katie Jones, PathFactory (vendor benchmark survey), via Demand Gen Report",
     "date": "2025-10-22",
     "url": "https://www.demandgenreport.com/industry-news/feature/how-to-prove-content-roi-in-the-ai-era-an-interview-with-pathfactorys-katie-jones/50638/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Carolyn Shelby (SEO consultant, principal SEO at Yoast, which sells SEO software), Search Engine Land",
     "date": "2025-09-04",
     "url": "https://searchengineland.com/gating-ai-hide-show-why-461535",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "content",
    "lead generation",
    "answer engines"
   ],
   "bench_notes": []
  },
  {
   "slug": "product-led-growth",
   "title": "Lead with a free tier or trial, and bring in sales later",
   "summary": "Update it: the product-first, sales-later sequence holds and AI may strengthen it, but inference cost ended the near-free free user, so free-first becomes a capped trial or credits first.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "Blake Bartlett says OpenView, a VC, coined \"product-led growth\" in 2016 (Sacra interview, 2023): the product acquires the customer through self-service, and \"Sales still very much exists in PLG, but it's optional, not foundationally necessary.\" Self-service has a ceiling, and sales steps in past it. OpenView's own PLG page no longer carries the definition.",
   "rationed": "Sales headcount, and the near-zero marginal cost of serving software, which made a free tier a cheap acquisition channel.",
   "challenge_claim": "AI products carry real cost per use, so an indefinite, full-featured free tier burns money; free becomes time-boxed, credit-capped or a reverse trial.",
   "blindspot": "The update side shows the practice changing but has no data on whether capped free tiers convert or retain as well as the ones they replace. The defend side counts AI's high self-serve share of spend, which is adoption, as if it showed the old free tier still works.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Serving a free user costs almost nothing.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "Users can reach value on their own, before talking to anyone.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Self-serve usage later feeds a sales-assisted expansion.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Inference cost makes an indefinite, full-featured free tier uneconomic, so free becomes time-boxed, credit-capped or a reverse trial.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "A capped or time-boxed free tier converts and retains as well as an indefinite one.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Cheap self-serve AI products churn too fast for free-first to pay off.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Serving a free user costs almost nothing.",
    "deciding_side": "prior",
    "what_follows": "One premise changed: serving a free user now has a real inference cost. The other two, that users reach value alone and that self-serve feeds sales later, hold and may be stronger with AI. So the sequence stays and the free tier is revised: metered trial or credits first, sales later. Whether the capped version converts as well is unproven."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Blake Bartlett says his firm, OpenView, a VC, coined \"product-led growth\" in 2016. The product does the acquiring through a free tier or a trial, self-service to value, and sales later. Bartlett is careful that this isn't anti-sales. \"Sales still very much exists in PLG, but it's optional, not foundationally necessary.\" Self-service has a ceiling, and sales steps in past it.\n\nThe prior rationed two things. Sales headcount, because the product carried acquisition. And the marginal cost of software, which was close enough to zero that giving the product away was a cheap channel."
    },
    {
     "heading": "What people say now",
     "text": "The update comes from people running AI products. Wes Bush, who founded ProductLed and sells PLG advisory, puts it bluntly: \"AI products have real COGS,\" and \"a bad free model can burn hundreds of thousands in COGS before you realize it's broken.\" His conclusion is narrow: \"Free isn't dead. But indefinite access with full features? That's gone.\" Elena Verna, who runs growth at Lovable and wrote in a post sponsored by Stripe Billing, describes a free tier of five credits a day, 30 a month at most.\n\nThe defense is that the motion is stronger than ever. Verna herself says \"AI is collapsing time-to-value down to first use.\" Menlo Ventures, a VC, surveyed about 500 enterprise buyers and puts product-led purchases at 27% of AI application spend, against 7% in traditional software, and notes Cursor reached $200 million in revenue before hiring enterprise sellers."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on three ideas. Serving a free user costs almost nothing. Users can reach value on their own. And self-serve usage later feeds a sales-assisted expansion.\n\nAI changed the first. Every use of an AI product runs inference, so a free user carries real cost; the price-per-seat entry in this series makes the same point about paid users. The second holds and may be stronger, since time to value is shrinking. The third holds on case evidence. Products that grew through self-serve built sales once they reached large accounts.\n\nSo the challenge's conclusion follows, but only for the free tier. An indefinite, full-featured free tier gets expensive, and teams move to time-boxed access, credit caps or reverse trials. The sequence, product first and sales later, is untouched. What the challenge doesn't have is evidence that the capped version works as well. No study compares credit-capped and indefinite free tiers on conversion or margin for AI products.\n\nMenlo's share figure is adoption. It shows buyers taking the product-led path, though it leaves unproven whether any particular free tier pays."
    },
    {
     "heading": "What both sides miss",
     "text": "Retention, and it cuts both ways. ChartMogul, a billing vendor, with Kyle Poyar, found AI-native gross retention of 23% below $50 a month and 70% above $250, in buckets of about 50 companies it calls directional. Cheap, self-serve AI products lose most of their customers. Against that, a16z, a VC, argues from its portfolio that products with generous free tiers see converted customers retain better at month three, without publishing figures.\n\nNobody has outcome data on companies that cut their free tier for inference cost. The acv-sets-the-motion entry covers the other half, noting that if AI makes a sales touch cheaper while free users get costlier, both push sales assistance down-market."
    }
   ],
   "open_question": "Does a credit-capped free tier convert and keep customers as well as the free tier it replaced?",
   "sources": [
    {
     "side": "origin",
     "who": "Blake Bartlett, OpenView (VC; invested in product-led companies), interviewed by Jan-Erik Asplund, Sacra",
     "date": "2023-06-27",
     "url": "https://sacra.com/research/blake-bartlett-openview-future-product-led-growth/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Wes Bush, ProductLed (sells PLG advisory), PLG predictions for 2026",
     "date": "2025-12-23",
     "url": "https://productled.com/blog/plg-predictions-for-2026",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Elena Verna (growth at Lovable; post sponsored by Stripe Billing), Elena's Growth Scoop",
     "date": "2025-06-06",
     "url": "https://www.elenaverna.com/p/why-ai-startups-are-blowing-past",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Menlo Ventures (VC), State of Generative AI in the Enterprise 2025, survey of about 500 enterprise buyers",
     "date": "2025-12-09",
     "url": "https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise/",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Santiago Rodriguez and Alex Immerman, a16z (VC), AI retention benchmarks",
     "date": "2025-09-10",
     "url": "https://a16z.com/ai-retention-benchmarks/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Kyle Poyar for ChartMogul (billing vendor), SaaS retention: the AI churn wave",
     "date": "2025-12-10",
     "url": "https://chartmogul.com/reports/saas-retention-the-ai-churn-wave/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "product-led growth",
    "free tier",
    "pricing"
   ],
   "bench_notes": []
  },
  {
   "slug": "mql-hands-off-demand",
   "title": "The MQL is how marketing hands demand to sales",
   "summary": "Keep it: the hand-off from marketing to sales still works, and hand-raises convert far better than content leads, but \"the MQL is dead\" mostly attacks the score, and whether scores have lost their predictive power is too early to call.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "hype-against",
   "because_label": "the claims against it are hype",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "The SiriusDecisions Demand Waterfall (dated 2002 or 2006 depending on the source): inquiry to MQL to SAL to SQL, with a scoring threshold and a sales SLA. SiriusDecisions re-architected it in 2012 and in 2017 replaced the lead with the demand unit, the buying group.",
   "rationed": "The seller held the information advantage, so a form fill was a reliable sign that someone wanted to talk.",
   "challenge_claim": "Buyers research with AI and arrive with a shortlist, so lead scoring and MQL thresholds measure nothing. Replace them with account-level or buying-group signals.",
   "blindspot": "The update side often leans on the \"buyers are 70% through the journey\" stat. That stat comes from a 2012 study of 22 suppliers' customers, and 70% was the top of its range. The canon itself replaced the lead in 2017, so the replacement is older than the AI argument made for it.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Marketing and sales need a defined hand-off, with a threshold and a follow-up agreement.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "A hand-raise, such as a demo request, means a buyer wants to talk.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Scored page views and downloads show who else is ready to buy.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "One person filling in a form stands in for the buying company.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Buyers research where scoring can't see, so behaviour scores have lost their predictive power.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "So the hand-off itself should be replaced.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "AI is what made the lead the wrong unit.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A hand-raise, such as a demo request, means a buyer wants to talk.",
    "deciding_side": "prior",
    "what_follows": "The hand-off holds, and the hand-raise now has outcome support: hand-raise MQLs move to SQL at about four times the rate of pools that include content leads. Whether scores have lost predictive power since AI moved research off the site is unmeasured, and model-based scoring still lifts conversion in recent studies, so retiring the score is too early to call. Either way, the challenge runs from a weaker score to a dead hand-off, and that step doesn't follow."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The SiriusDecisions Demand Waterfall gave marketing and sales a shared ladder of inquiry, marketing-qualified lead, sales-accepted and sales-qualified. It came with a scoring threshold and a service-level agreement for follow-up. In 2017 the same firm replaced the lead with the demand unit, the buying group, because one person filling in a form had stopped describing how companies buy."
    },
    {
     "heading": "What people say now",
     "text": "\"The MQL is dead\" is one of the most repeated lines in marketing ops. The usual case: buyers research on their own, increasingly inside AI assistants, and arrive with a shortlist, so a score built on page views and downloads measures nothing.\n\nThe pushback comes from the people who run the hand-off, and it circulates in marketing-ops forums: if you have a demo form on your site, you're still using MQLs, and the problem is how most teams implement the MQL, not the acronym."
    },
    {
     "heading": "Does the logic hold",
     "text": "The MQL bundles four ideas. Marketing and sales need a defined hand-off. A hand-raise means someone wants to talk. Scored behaviour, page views and downloads, shows who else is ready. And one person stands in for the company buying. The fourth failed before AI did anything, when the canon itself moved to the buying group, so that replacement is older than the argument now made for it.\n\nThe second idea now has outcome support. HockeyStack, which sells attribution, compared 87 B2B software companies over late 2023 and early 2024. Companies that counted only demo, pricing and contact requests as MQLs moved 21.55% of them to SQL. Companies that also counted ebook and webinar leads moved 4.93%. The hand-raise pools closed at a higher rate too, at about half the cost per SQL.\n\nThe third idea is where the challenge aims, and the case is weaker than it sounds. The claim is that buyers now research inside assistants the score never sees, so scores have lost their power. Nobody has measured that loss, and there is no trend in score-based conversion since 2023. The HockeyStack gap is a level at one point in time, not a decline. And model-based scoring still works in recent academic tests. Across three B2B service industries, a scoring model lifted conversion well above an intuition baseline. Retiring the score is too early to call.\n\nWhat doesn't follow at all is the step from a weaker score to a dead hand-off. A demo request means what it always meant. A hand-raise and a score crossing are different events that share one label, and most of the case against the MQL is a case against the score."
    },
    {
     "heading": "What both sides miss",
     "text": "Updaters miss who is making the case. Kerry Cunningham, one of the two people who unveiled the buying-group model in 2017, now writes the sharpest \"MQL is dead\" pieces from 6sense, a vendor that sells the replacement. The strongest supporting stat usually cited, that buyers are 70% through their journey before contacting sales, comes from a 2012 study of 22 suppliers' customers.\n\nThe defenders miss something too. \"It's the implementation\" is a fair defence only if someone actually fixes the implementation."
    }
   ],
   "open_question": "In your won deals, what did the first real signal look like: a score, a hand-raise, or something your systems never recorded?",
   "sources": [
    {
     "side": "origin",
     "who": "Forrester (SiriusDecisions), Demand Unit Waterfall launch",
     "date": "2017-05-17",
     "url": "https://www.forrester.com/blogs/meetthenewestsiriusdecisionsdemandwaterfall/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Kerry Cunningham, 6sense (sells intent and ABM software)",
     "date": "2025-08-31",
     "url": "https://6sense.com/science-of-b2b/the-tyranny-of-the-benchmark-how-mql-standards-institutionalized-irrelevance",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Emir Atli, HockeyStack Labs (sells attribution), The Final Verdict: Demand Gen vs. Lead Gen",
     "date": "2025-12-15",
     "url": "https://www.hockeystack.com/lab-blog-posts/the-final-verdict-demand-gen-vs-lead-gen",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Wu, Andreev and Benyoucef, PRISM, Journal of Personal Selling & Sales Management",
     "date": "2026",
     "url": "https://www.tandfonline.com/doi/full/10.1080/08853134.2026.2633231",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "demand",
    "lifecycle",
    "marketing ops"
   ],
   "bench_notes": []
  },
  {
   "slug": "speed-to-lead",
   "title": "Respond to inbound leads within the hour",
   "summary": "Keep it: outcome evidence still shows conversion falling fastest in the first hour after an inquiry, so reply within the hour and never the next day, but drop the five-minute rule and its 100x and 21x figures, which stop at qualification.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "still-true",
   "because_label": "still true",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L3 Practice",
   "origin": "Oldroyd, McElheran and Elkington, HBR 2011, \"The Short Life of Online Sales Leads\": firms that tried to reach a web lead within an hour were nearly seven times as likely to qualify it as firms that tried an hour later. The 2007 Lead Response Management study by Oldroyd and Elkington, on InsideSales.com's own dialer data from six companies, sharpened it to five minutes: 100 times the contact rate and 21 times the qualification rate of a call at thirty. Elkington was CEO of InsideSales.com.",
   "rationed": "Rep attention at the moment of inquiry. A fast reply needed someone watching the queue, so few teams managed it.",
   "challenge_claim": "AI responders answer every inbound lead in seconds, so speed is table stakes for everyone and the relevance of the reply is the new edge.",
   "blindspot": "The update side reads adoption as if it settled the question: whether rivals reply fast says nothing about whether a fast reply converts. The defend side quotes 100x and 21x as if they predicted revenue, when that data stops at qualification and the best outcome study finds tens of percent, not multiples.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A buyer's attention to an inquiry fades soon after it, so an early reply converts better than a late one.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Replying fast is costly, because it needs a person watching the queue.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "The lift is sized in multiples, 100x to contact and 21x to qualify, and it carries through to revenue.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI responders make an instant reply cheap for any team.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Once everyone is fast, speed stops mattering and the relevance of the reply decides.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "An instant automated reply captures the same lift as a fast human reply.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A buyer's attention to an inquiry fades soon after it, so an early reply converts better than a late one.",
    "deciding_side": "prior",
    "what_follows": "The causal premise holds on outcome evidence: conversion falls fastest in the first 10 to 20 minutes and flattens after about an hour. AI changed the cost of speed, not whether speed works, so a reply within the hour stays worth paying for whatever competitors do; being table stakes is a reason to keep a practice, not to drop it. What goes is the five-minute precision and the 100x and 21x sizes, and whether an instant AI reply converts like a fast human one is unproven."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The rule has two versions. In 2011 Harvard Business Review published \"The Short Life of Online Sales Leads,\" built on 1.25 million leads at 42 firms: firms that tried to reach a lead within an hour were nearly seven times as likely to qualify it as firms that waited an hour longer. A 2007 paper by the same authors, on InsideSales.com's own dialer data from six companies, had sharpened the window to five minutes, with contact about 100 times and qualification about 21 times as likely as a call at thirty. David Elkington, a co-author of both, ran InsideSales.com, which sold the dialer. Both studies stopped at qualification."
    },
    {
     "heading": "What people say now",
     "text": "The challenge says AI answers every form in seconds, so speed has stopped being an edge. Sales trainers now repeat on LinkedIn that once everyone can be fast, fast is no advantage.\n\nThe sharper challenge comes from 6sense, which sells intent and ABM software and gains when attention moves away from form-fill speed. It argues the speed data shows selection, not persuasion: leads reached fast differ from leads reached slowly, and in its 2026 buyer study the eventual winner was on the day-one shortlist 92% of the time. Even 6sense concedes that a buyer who asks to talk should hear back promptly."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior makes the causal claim that a buyer's attention fades soon after an inquiry, so an early reply converts better. Two ideas came with it. Replying fast was costly, and the lift was large, measured in multiples.\n\nThe causal claim now has outcome evidence that answers the selection point. Musalem and colleagues (*Operations Research*, 2023) used eight months of quote data from an auto insurer, with a method built to separate the effect of waiting from the kind of lead that waits. Purchase conversion fell 31% when the first call moved from 5 to 30 minutes, a real effect in the tens of percent rather than 100x or 21x. The drop was steepest in the first 10 to 20 minutes and flattened after about 70. That is consumer insurance, but the vendor data that includes B2B agrees on the shape. VanillaSoft, which sells sales-engagement software, found in its 2018 data that wins were far more likely within the hour than after it. Velocify, which sold lead-distribution and dialer software, reported the advantage \"greatly reduced after the first hour.\"\n\nAI changed the cost of speed while leaving its effectiveness intact, so the challenge's conclusion doesn't follow. If a reply within the hour converts better, a team keeps paying for it whatever its rivals do. Table stakes is a reason to keep a practice, not to drop it."
    },
    {
     "heading": "What both sides miss",
     "text": "Defenders keep the wrong numbers, figures that stop at qualification. Two vendor datasets even find the very fastest band is not the best, as VanillaSoft's wins peaked in the 10-to-60-minute window. \"Within the hour, never the next day\" is well supported. \"Five minutes beats fifteen\" is not.\n\nCritics assume an instant AI reply carries the same lift as a fast human one. No study compares them on inbound. Field experiments in consumer and small-business lending cut both ways. An AI first call raised loan applications when a human followed up and failed otherwise, and disclosing a bot before the call cut purchases sharply."
    }
   ],
   "open_question": "Does an instant AI reply convert like a fast human one, or does the lift need a person on the other end?",
   "sources": [
    {
     "side": "origin",
     "who": "David Elkington and James Oldroyd, Lead Response Management Report (InsideSales.com dialer data, 6 companies; InsideSales sold the dialer)",
     "date": "2007-11-01",
     "url": "http://content.marketingsherpa.com/heap/DG07SFSlides/LeadResponseManagementReport.pdf",
     "standing": "position"
    },
    {
     "side": "origin",
     "who": "Oldroyd, McElheran and Elkington (CEO, InsideSales.com), HBR, The Short Life of Online Sales Leads",
     "date": "2011-03",
     "url": "https://hbr.org/2011/03/the-short-life-of-online-sales-leads",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Musalem, Olivares and Yung, Operations Research (independent academic study, auto insurance)",
     "date": "2023-05",
     "url": "https://pubsonline.informs.org/doi/10.1287/opre.2022.2418",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "VanillaSoft (sells sales-engagement software) and University of Ottawa Telfer School, lead conversion study (preliminary)",
     "date": "2018",
     "url": "https://vanillasoft.com/resources/lead-conversion-study-telfer",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Kerry Cunningham and Sara Boostani, 6sense (sells intent and ABM software), The Speed to Lead Fallacy",
     "date": "2026-11-02",
     "url": "https://6sense.com/science-of-b2b/the-speed-to-lead-fallacy",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "inbound",
    "speed to lead",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "sdr-ae-specialization",
   "title": "Split the work: SDRs book meetings, AEs close",
   "summary": "Too early to call: nobody has measured whether SDR-booked meetings convert better or more cheaply than AE-sourced or AI-booked ones, so both sides argue the split's productivity case without data, while its other job, training future AEs, has real outcome support.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "Aaron Ross and Marylou Tyler, Predictable Revenue (2011): split prospecting (SDR) from closing (AE). The claim against it, \"the SDR is dead\", took off in 2025, led by AI-SDR vendors and some sales-tech executives.",
   "rationed": "Seller hours and writing cost. Prospecting took so many hours that it paid to give it to a cheaper, junior role, and research and writing for each touch were expensive.",
   "challenge_claim": "AI agents now research, write and book meetings, so the SDR seat disappears within a year or two and AEs take the pipeline directly from agents.",
   "blindspot": "The update side ignores what the seat was for besides meetings: it is where future AEs learn to sell, and promoted insiders tend to outperform outside hires. The defend side treats a flat SDR ratio and slow AI-SDR uptake as proof the split works, when they only show what the market is doing.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Researching and writing each touch took so many hours that prospecting paid to hand to a cheaper role.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "A junior seller can book meetings well enough to free AEs to close.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The SDR seat is where future AEs learn to sell.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Agents can do the whole first touch with no human in the loop.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The fall in SDR hiring is caused by AI.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "AEs can take pipeline straight from agents and lose nothing the seat produced.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A junior seller can book meetings well enough to free AEs to close.",
    "deciding_side": "prior",
    "what_follows": "AI changed the cost of research and writing, which changes the work inside the seat. The deciding question, whether junior-booked meetings free AEs at a good enough conversion and cost, has no outcome test either way, and neither does an agent-run first touch, so the productivity case is too early to call. The seat's second job does hold: insiders promoted into a role outperform outside hires in other fields, so removing the seat costs a bench, and that cost belongs in the decision."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The split is about fifteen years old. Aaron Ross's *Predictable Revenue* (2011) separated prospecting from closing, so a junior SDR booked meetings and an AE closed them. It became standard in software sales, and the Bridge Group has tracked it since."
    },
    {
     "heading": "What people say now",
     "text": "The case against it got loud in 2025, under the slogan that the SDR role is dead. Salesloft's CRO said he wasn't sure the BDR role would \"exist post 2026.\" AI-SDR vendors went further and sold replacement outright.\n\nSome of those vendors have since retreated. 11x's CEO: \"I'm the CEO of the company that coined AI SDRs. And I'm here to tell you AI SDRs don't work.\" His point is that the replacement framing fails, not the software. Artisan retired its \"stop hiring humans\" slogan in September 2026 and still sells an AI BDR.\n\nThe numbers both sides trade describe the market without measuring the outcome. Bridge's 2025 report has the SDR-to-AE ratio \"consistent since 2018\" and AI SDRs at 1% of companies surveyed. Kyle Poyar's job-post data has SDR postings down 21%. None of it says whether the split produces better pipeline."
    },
    {
     "heading": "Does the logic hold",
     "text": "The split rests on three ideas. Prospecting took so many hours, mostly in research and writing, that it paid to give it to a cheaper role. A junior seller could book meetings well enough to free AEs to close. And the seat trained the next AEs.\n\nAI changed the first because research and writing for a touch now cost almost nothing. That changes the work inside the seat, but the split's productivity claim lives in the second idea, and nobody has tested it. No public dataset compares the win rate or cost per won deal of SDR-booked meetings with AE-sourced ones. The figures comparing AI-booked with human-booked meetings all trace to vendors or unnamed tests. The challenge's own premise, that agents can run the whole first touch, is just as unmeasured. Both sides lean on the same missing number.\n\nThe third idea is different, because it has outcome support from other fields. Among more than 100,000 commissioned salespeople at one national retailer, internal hires sold more than external hires, mostly because the best insiders stayed (Rissing and Benson, *Organization Science*, 2021). In investment banking, people promoted into a job outperformed external hires for about two years and were paid about 18% less (Bidwell, 2011). The only SDR-specific data, a small Bridge Group sample from 2017, found 26% of internally promoted SDRs failed as AEs, against 41% of SDRs who took an AE seat elsewhere, and failure fell to 6% for SDRs with 16 months or more in the seat. A plan that removes the seat has to replace that bench."
    },
    {
     "heading": "What both sides miss",
     "text": "Updaters credit AI with a decline that Salesloft's own CRO blamed on the end of cheap money. If the cause is the cost of capital, a shrinking SDR headcount says little about what agents can do.\n\nDefenders miss that the job inside the title is changing. Where AI has worked, it has been on inbound qualification, or as research feeding human reps."
    }
   ],
   "open_question": "If agents do the first touch, where do your future AEs come from?",
   "sources": [
    {
     "side": "origin",
     "who": "Aaron Ross, Predictable Revenue",
     "date": "2011-09-12",
     "url": "https://pebblestorm.com/2011/09/12/my-new-book-predictable-revenue/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Mark Niemiec, CRO Salesloft (sells sales-engagement software), Belkins podcast",
     "date": "2025-09-15",
     "url": "https://www.youtube.com/watch?v=g2go2IUHFts",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Prabhav Jain, CEO 11x (sells AI SDRs)",
     "date": "2026-03-02",
     "url": "https://www.linkedin.com/posts/jainprabhav_im-the-ceo-of-the-company-that-coined-ai-activity-7434285950316445696-CFY9",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Rissing and Benson, Organization Science (via Cornell ILR summary), internal vs external hires among 109,063 commissioned salespeople",
     "date": "2021-01-13",
     "url": "https://www.ilr.cornell.edu/news/research/study-tests-why-internal-hires-outperform-external-hires",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Matthew Bidwell, Administrative Science Quarterly, Paying More to Get Less (investment banking)",
     "date": "2011",
     "url": "https://doi.org/10.1177/0001839211433562",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Matt Bertuzzi, The Bridge Group (sells sales-development consulting), failure rate of SDR-to-AE promotions",
     "date": "2017-08-17",
     "url": "https://blog.bridgegroupinc.com/failure-rate-of-sdr-to-ae-promotions",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "outbound",
    "sdr",
    "org design"
   ],
   "bench_notes": []
  },
  {
   "slug": "multi-touch-attribution",
   "title": "Multi-touch attribution can apportion credit across touches",
   "summary": "Too early to call: the case that attribution is dead rests on consumer-media evidence nobody has tested on B2B opportunity influence, and buyers shortlisting inside AI chats opens a gap that neither side can measure yet.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "Bizible, now Adobe Marketo Measure, set the B2B models: U-shaped splits credit 50/50 between first touch and lead creation; W-shaped gives 30% each to first touch, lead creation and opportunity creation, and 10% to the rest (Justin Norris, Marketo Nation, 2018). The IAB's 2017 multi-touch primer noted last-touch still dominated despite its known limits. The weights are conventions, not estimates, and sources define W-shaped differently.",
   "rationed": "Measurement cost. A fixed rule over tracked touches gave every channel a share of pipeline without running an experiment.",
   "challenge_claim": "Multi-touch attribution is dead: signal loss from cookies and iOS breaks touch tracking, holdout tests show it over-credits the lower funnel, and the replacement is incrementality testing plus marketing-mix modeling.",
   "blindspot": "The \"dead\" case is built on consumer paid media and applied to B2B opportunity influence, which is what marketing ops actually runs. The defenders miss the new hole: a shortlist built inside an AI chat leaves no touch to credit, under any model.",
   "logic": {
    "prior_premises": [
     {
      "premise": "The touches that shape a deal can be tracked and tied to a person and an account.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "A fixed weighting by touch position approximates what each touch contributed.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Marketing needs a per-channel read of what drove pipeline to defend and allocate budget.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Correlated credit overstates touches that would have happened anyway.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Consumer-media signal loss and holdout results carry over to B2B opportunity influence.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Incrementality tests and mix models can replace attribution for B2B budget decisions.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Consumer-media signal loss and holdout results carry over to B2B opportunity influence.",
    "deciding_side": "challenge",
    "what_follows": "Two things weakened the first premise: tracking lost signal, and buyers moved research into AI tools where no touch is recorded. The challenge is right that correlated credit flatters, but its conclusion rests on a transfer from consumer media to B2B that nobody has tested, and on replacements unproven for B2B. The deciding premise is unproven, so it is too early."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Bizible, now Adobe Marketo Measure, gave B2B marketing its standard models. U-shaped splits credit between the first touch and lead creation. W-shaped gives 30% each to first touch, lead creation and opportunity creation, and spreads the last 10% over everything else. The weights are conventions, not estimates, and the sources don't agree on what W-shaped even means. The IAB's 2017 primer noted that last-touch still dominated despite its known limits."
    },
    {
     "heading": "What people say now",
     "text": "Trade-press columns now call multi-touch attribution (MTA) a waste of time and money. Measured, which sells incrementality testing, goes further: \"a more sophisticated version of a flawed idea is still a flawed idea.\" Its own geo holdouts found lower-funnel remarketing over-credited by about 2.6 times.\n\nThe defense comes from attribution vendors. CaliberMind: \"there's no alternative.\" Between the two sits a softer consensus around triangulation across attribution, mix modeling and experiments, which commentators treat as settled."
    },
    {
     "heading": "Does the logic hold",
     "text": "Attribution rested on three ideas. The touches that shape a deal can be tracked and tied to a person and an account. A fixed weighting approximates what each touch contributed. And marketing needs a per-channel read of what drove pipeline. The second was never measured; it was a convention from the start. The third still holds, which is why teams keep running models they don't fully trust.\n\nTwo things moved the first idea. Tracking lost signal, and buyers moved their research into AI tools. Forrester finds 94% of business buyers now use AI in the buying process.\n\nThe challenge is right that correlated credit flatters touches that would have happened anyway. But its evidence comes from consumer media, with cookies, iOS and retail holdouts, and its conclusion is applied to B2B opportunity influence. That transfer hasn't been tested. Nor has the claim that mix models and holdouts can carry B2B budget decisions. The premise that decides the question is unproven."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides argue about how to split credit among touches you can see. The newer problem is the touch you can't. If a buyer builds the shortlist in a chat with an AI assistant, the first recorded touch may come from someone who has already narrowed the field. Position-based models credit whatever came next. Holdouts and mix models measure channels you spend on, and it isn't clear what a holdout of an AI chat would even look like. Nobody has a measurement for this yet, including the camp declaring MTA dead."
    }
   ],
   "open_question": "How do you credit pipeline the buyer shortlisted inside an AI chat?",
   "sources": [
    {
     "side": "update",
     "who": "Terence Einhorn, Measured (sells incrementality testing)",
     "date": "2026-06-03",
     "url": "https://www.measured.com/faq/multi-touch-attribution-is-dead-heres-what-replaced-it/",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Mary Batchelder, CaliberMind (sells attribution), State of Marketing Attribution 2025",
     "date": "2025-06-23",
     "url": "https://calibermind.com/guides-reports/state-of-marketing-attribution-report-2025/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "John Buten, Forrester, B2B Buyers Make Zero-Click Buying Number One",
     "date": "2026-01-22",
     "url": "https://www.forrester.com/blogs/b2b_buyers_make_zero_click_buying_number_one/",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "attribution",
    "marketing ops",
    "measurement",
    "buyer journey"
   ],
   "bench_notes": []
  },
  {
   "slug": "third-party-intent",
   "title": "Third-party intent shows who is in market",
   "summary": "Too early to call: the critique that a surge lands after the shortlist is set has an analyst behind it, but the newer claim that AI research starves intent data has no measurement yet, and the defenses argue for using intent differently rather than showing that it predicts.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L3 Practice",
   "origin": "The intent vendors' own product definitions. Bombora's Company Surge compares an account's research on a topic over the last three weeks against a 12-week baseline, from a co-op of publisher sites, and says a surge shows an account \"indicating active demand\". 6sense scores accounts into buying stages that indicate \"the likelihood of an account opening or progressing in the next 90 days.\"",
   "rationed": "Visibility into anonymous research. Buyers read before they called anyone, and pooling publisher traffic by company was the only way to see who was reading.",
   "challenge_claim": "Buyers now research and shortlist inside AI assistants that emit no intent signal, so third-party surges miss them or arrive after the shortlist is already set.",
   "blindspot": "The update side leans on an AI coverage loss nobody has measured, when its stronger argument, that the shortlist forms before the surge, does not need AI at all. The defend side answers by widening what intent is for, which concedes the timing point instead of answering it.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Research before a purchase happens on sites a publisher co-op can observe and tie to an account.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "A rise in that research marks an account that is in market.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The surge comes early enough for a seller to get on the shortlist.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "A growing share of buyer research happens inside AI assistants.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "Research inside AI assistants leaves no trail a co-op can see, so intent coverage falls.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "So the surge now arrives after the shortlist is set.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Research inside AI assistants leaves no trail a co-op can see, so intent coverage falls.",
    "deciding_side": "challenge",
    "what_follows": "The deciding premise is whether AI research removes the signal, and nobody has measured intent coverage before and after buyers moved to assistants; the one analyst summary on it says co-op signals have held up. The better-evidenced critique, that the shortlist forms before the surge, predates AI and leans toward failing the prior's timing premise, but no one has measured when surges land against when shortlists form. Neither side's conclusion follows yet."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Third-party intent rests on a co-op. Bombora, which sells the data, collects research activity through a tag on the sites in its co-op and flags a surge when an account's reading on a topic over the last three weeks rises above its 12-week baseline. 6sense, which sells an intent platform, scores accounts into buying stages that predict an opportunity opening within 90 days. These are descriptions of what the products do. No independent validation of the claim that a surge predicts a purchase is on record, and the vendors' own figures drift from one document to the next."
    },
    {
     "heading": "What people say now",
     "text": "The sharpest critique is about timing. Forrester, which sells research rather than intent data, warns that late-stage signals \"may represent prospects who have already narrowed their options to a shortlist of vendors.\" 6sense's own 2026 buyer research points the same way, against its own product: 92% of winning vendors were on the initial shortlist, and buyers spend about 60% of a 10.3-month cycle in selection before contacting a seller.\n\nThe newer critique adds AI, noting that buyers research in assistants that emit no intent signal. Surveys agree research is moving there and disagree on how far. 6sense finds only 26% use language models to discover vendors; a G2 survey finds 51% of software buyers now start research in a chatbot more often than in Google. No one has published a measurement of intent coverage, match rates or surge frequency before and after that shift. The claim describes a mechanism without offering a measurement.\n\nThe defenses are softer. Forrester argues intent has uses beyond spotting in-market accounts, and Gartner tells demand leaders to triangulate across several intent sources. A Forrester report from April 2026, read only in summary, says co-op and bidstream providers have \"shown resilience in maintaining predictive signals,\" with the walled-garden sources most at risk. None shows the signal predicts."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on three ideas. Research before a purchase happens where a co-op can see it. A rise in that research marks an account in market. And the rise comes early enough to act on, before the shortlist is set. The vendors make the case for all three; nobody independent has tested them.\n\nThe challenge says AI changed the first idea. That needs research to have moved into assistants, which surveys support, and the move to have cut what co-ops see, which nobody has measured. So the premise that decides the AI claim is unproven.\n\nThe timing critique is stronger, and it doesn't need AI. If the shortlist forms before a buyer talks to anyone, a surge can show who is buying and still arrive too late to change whom they buy from. The evidence leans that way, but nobody has measured when surges land against when shortlists form."
    },
    {
     "heading": "What both sides miss",
     "text": "The prior bundles two claims, that this account is in market and that you've learned it in time to matter. The timing evidence touches only the second. Defenders answer by retreating to other uses, which concedes the second without defending the first. The question nobody has data on is whether a surge, however late, still sorts accounts better than not having one."
    }
   ],
   "open_question": "Did your won deals surge before your first call, or after the buyer had already picked you?",
   "sources": [
    {
     "side": "origin",
     "who": "Bombora (sells intent data), Company Surge intent",
     "date": "n.d. (read 2026-11-10)",
     "url": "https://bombora.com/intent/",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "6sense (sells an intent platform), Predictive Buying Stages",
     "date": "2026-11-03 (page updated)",
     "url": "https://support.6sense.com/docs/predictive-buying-stages",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Forrester, B2B marketers: are you getting everything you can out of intent data?",
     "date": "2025-07",
     "url": "https://www.forrester.com/blogs/b2b-marketers-are-you-getting-everything-you-can-out-of-intent-data/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "6sense (sells intent), 2026 Buyer Experience Report press release",
     "date": "2026-11-02",
     "url": "https://www.businesswire.com/news/home/20261102445967/en/6sense-Research-Finds-That-Vendors-Need-to-Be-Known-and-Preferred-Before-the-Buying-Journey-Begins",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Brett Kahnke, Forrester, The Impact Of AI-Powered Search On The Intent Data Market (summary)",
     "date": "2026-04-08",
     "url": "https://www.forrester.com/report/the-impact-of-ai-powered-search-on-the-intent-data-market/RES193421",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Alan Antin and Chris Chandler, Gartner, Combine Multiple Intent Signals to Find Real Tech Buying Activities",
     "date": "2025-04-15",
     "url": "https://www.gartner.com/en/documents/6361911",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "intent data",
    "demand",
    "abm"
   ],
   "bench_notes": []
  },
  {
   "slug": "abm-tiers",
   "title": "Tier ABM accounts one-to-one, one-to-few and one-to-many",
   "summary": "Too early to call: AI may have cut the cost of account research, but no one has shown AI-scaled programs getting one-to-one results, and human attention still has to be rationed, so the open question is the unit and cadence of ranking, not whether to rank.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "mostly-aspiration",
   "because_label": "the case to update is mostly aspiration",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "ITSMA coined account-based marketing in 2003, and Bev Burgess, then at ITSMA, codified it. The three types (Strategic ABM one-to-one, ABM Lite one-to-few, Programmatic ABM one-to-many) were documented in ITSMA's 2016 benchmarking survey, which is gated; the earliest public trace is the 2017 ITSMA and ABM Leadership Alliance study page. Burgess's 2025 handbook expands them to five types run as a portfolio.",
   "rationed": "Marketer hours per account and the cost of custom research and content, with seller attention as a second limit. ITSMA's 2016 survey found a median of four strategic accounts per dedicated marketer; Forrester put a full-time one-to-one marketer at about five.",
   "challenge_claim": "AI makes account research and tailored content cheap, so one-to-one depth can reach many accounts, and fixed tiers are a relic of scarcity.",
   "blindspot": "The update side reads a fall in research cost as a rise in results, when no study links them and Gartner finds content tailored to the individual can hurt group consensus. The defend side leans on self-reported ROI that never compares tiers, and on an annual account list that even the tiers' originator now treats as one option in a portfolio.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Account-specific research and content are expensive, so depth must be rationed to a few accounts.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Human attention from marketers, sellers and executives is scarce and should go where value is highest.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Deeper tiers return more per account.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI-made account research and content give many accounts results close to a one-to-one program.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The opportunity and its buying group, re-ranked as signals change, is a better unit than an annual account list.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Teams already run several tiers at once.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "AI-made account research and content give many accounts results close to a one-to-one program.",
    "deciding_side": "challenge",
    "what_follows": "Only the cost premise changed, and only on an analyst's assertion: no one has published cost or outcome data. The claim that tiers collapse needs AI-scaled research to deliver one-to-one results, which nobody has measured, while the scarcity of human attention that the tiers ration still holds. The narrower change, ranking opportunities rather than fixing an annual list, doesn't depend on AI and is the part worth testing now."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "ITSMA, a firm that sells ABM research and training, coined account-based marketing in 2003, and Bev Burgess, then at ITSMA, codified it. The three tiers came later, with Strategic ABM for one account at a time, ABM Lite for small clusters, and Programmatic ABM for many. They were documented in ITSMA's 2016 benchmarking survey, which is gated. The earliest public trace is a 2017 study page from ITSMA and the ABM Leadership Alliance.\n\nThe tiers rationed marketer hours. That 2016 survey found a median of four strategic accounts per dedicated marketer, and nine for ABM Lite. The originator has since moved on. Burgess's 2025 handbook sets out five types, Strategic, Scenario, Segment, Programmatic and Pursuit, run as a portfolio."
    },
    {
     "heading": "What people say now",
     "text": "The sharpest challenge comes from the analyst who defended the top tier. In 2024 Nora Conklin of Forrester called strategic one-to-one ABM \"successful and durable\" and said scale \"is not what will power profitable growth.\" In 2026 she wrote that \"static segmentation was a solution to scarcity,\" and that insights which \"once relied on manual research are now available at scale.\"\n\nVendors push further, saying AI makes one-to-one research cheap enough to collapse the tiers. Adoption is moving toward mixing them. Momentum ITSMA's 2024 benchmark, sponsored by the vendor Demandbase, finds 61% of teams running two or more types, up from 35% in 2017. That says what teams do without showing what works."
    },
    {
     "heading": "Does the logic hold",
     "text": "The tiers rest on three ideas. Account research and content are expensive, so depth has to be rationed. Human attention is scarce and should go where value is highest. And deeper tiers return more.\n\nAI changed the first in cost, though on an analyst's assertion rather than published data, and the surveys say AI in ABM is early. Average maturity is 2.3 out of 5, and nearly 70% of respondents to a 2025 Demand Gen Report survey find its current effectiveness limited. The second holds, and Conklin keeps it: \"not every potential deal deserves equal attention.\" The third was never shown. ABM's returns are self-reported, and no study compares tiers.\n\nThe challenge needs AI-made research and content at scale to approach one-to-one results. Nobody has measured that, so the case is aspiration. Read closely, even Conklin's 2026 piece keeps prioritization. It changes the unit, from the account to the opportunity and its buying group, and the cadence, from an annual list to segments that move."
    },
    {
     "heading": "What both sides miss",
     "text": "The challenge assumes personalization is the lever. Gartner's 2025 buyer survey points the other way. Content tailored to the individual had a 59% negative effect on buying-group consensus, while content relevant to the group helped. Cheap tailored content is not the same as a one-to-one result.\n\nDefenders assume the top tier pays. The one academic look at a close cousin, key account management in one firm, found key accounts retained better but less satisfied and less profitable by year three. Forrester's Nicky Briggs already cautioned against one-to-one for new logos in 2023.\n\nThe change that holds up today isn't about AI. Forrester and a 2026 study in *Industrial Marketing Management* both point to prioritization that follows the opportunity's complexity and value over a fixed annual list."
    }
   ],
   "open_question": "If AI makes one-to-one research cheap, does the one-to-one result follow, or only more tailored content?",
   "sources": [
    {
     "side": "origin",
     "who": "Bev Burgess, Account-Based Marketing: The Definitive Handbook (Kogan Page)",
     "date": "2025-03-03",
     "url": "https://www.koganpage.com/marketing-communications/account-based-marketing-9781398619135",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "ABM Leadership Alliance and ITSMA (sells ABM research and training), 2017 ABM Benchmark",
     "date": "2017",
     "url": "https://abmleadershipalliance.com/ebook/growth-three-types-abm/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Nora Conklin, Forrester, ABM Tiers Were Made For Yesterday's World",
     "date": "2026-09-25",
     "url": "https://www.forrester.com/blogs/abm-tiers-were-made-for-yesterdays-world/",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Nora Conklin, Forrester, ABM Won, But It's Not Done Changing The Game",
     "date": "2024-05-30",
     "url": "https://www.forrester.com/blogs/abm-won-but-its-not-done-changing-the-game/",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Nicky Briggs, Forrester",
     "date": "2023-05-19",
     "url": "https://www.forrester.com/blogs/five-questions-to-ask-before-committing-to-121-account-based-marketing/",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Sharma and Evanschitzky, Journal of Business & Industrial Marketing (one firm)",
     "date": "2016-03-07",
     "url": "https://doi.org/10.1108/jbim-11-2014-0234",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "abm",
    "account selection",
    "personalization"
   ],
   "bench_notes": []
  },
  {
   "slug": "events-cut-first",
   "title": "Cut events first: they are expensive and the hardest to attribute",
   "summary": "Too early to call: \"hard to attribute\" never meant low return, and in-person meetings show causal value where they have been measured, but the claim that AI makes events worth more than before is a vendor-led pitch with no outcome data yet.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "mostly-aspiration",
   "because_label": "the case to update is mostly aspiration",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L3 Practice",
   "origin": "No traceable origin; it is budget practice. Events are the largest offline line item and the hardest to tie to revenue, so they go first when budgets tighten. The accountability problem is old: Gopalakrishna, Lilien, Williams and Sequeira, Journal of Marketing 1995, \"Do Trade Shows Pay Off?\", found show spend justified by rules of thumb and named accountability as the top challenge.",
   "rationed": "Budget under a flat marketing envelope. Gartner has marketing budgets at 7.8% of revenue in 2026, 18% below four years earlier, and attribution tools favour channels that leave a click trail.",
   "challenge_claim": "AI floods digital channels with generated content and wears down trust in them, so in-person events now return more than before and should be protected, not cut.",
   "blindspot": "The cut side treats hard to attribute as low return, when the measured cases, all before AI and outside software, show in-person meetings doing real work. The protect side's AI-era case is a vendor campaign and commissioned opinion, with no before-and-after data on what events return.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Events are expensive per contact.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "Event impact is hard to measure.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "What can't be attributed returns little, so the hardest-to-attribute line goes first.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Buyers lean on trusted human sources to check what AI tells them.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "In-person meetings do causal work that a digital touch doesn't replace.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "AI-generated outreach lowers the value of digital touches relative to in-person ones, so events now return more than before.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "AI-generated outreach lowers the value of digital touches relative to in-person ones, so events now return more than before.",
    "deciding_side": "challenge",
    "what_follows": "The prior's weak link is its own inference: hard to measure says nothing about return, and where return has been measured, meeting in person does work. That knocks out the reason for cutting events first, but not the case for cutting them. The challenge's step from AI-flooded channels to higher event returns has no outcome data, so neither side's conclusion is shown yet."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "No one coined it. \"Cut events first\" is budget practice. Events are the largest offline line item and the hardest to tie to revenue, so when the budget shrinks they go first. The accountability problem is old. In 1995 the *Journal of Marketing* asked \"Do Trade Shows Pay Off?\" and found show spend justified by rules of thumb, with accountability the top challenge. The same paper measured one firm at one show under controlled conditions and projected a 112% long-term return."
    },
    {
     "heading": "What people say now",
     "text": "CMOs are acting on the prior. Gartner's 2026 survey of 401 CMOs finds them \"rapidly shifting budget from offline to digital channels,\" with AI a driver because it favours channels \"that can be effectively AI-optimized.\" Forrester's 2026 survey of more than 400 event decision-makers finds nearly 70% running fewer events. Small formats under 200 people are growing, and only 18% plan more large ones.\n\nThe counter-claim is that events gain value as AI floods digital channels. Its canonical form is commercial. Cvent, an events vendor, built its 2026 \"Presence Premium\" message around it, backed by a vendor survey in which 85% say AI content makes trust harder to build."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on three ideas. Events are expensive per contact. Their impact is hard to measure. And what can't be attributed returns little, so the hardest line to attribute goes first. The first two hold. Forrester finds over half of teams still use net-new leads as the main event objective, which it says \"drives overly simplistic measurement.\"\n\nThe third is a jump with no finding behind it. Hard to measure says nothing about return, and where return has been measured, meeting in person does work. When SARS kept foreign buyers away from the 2003 Canton Fair, exhibitors reached 5 to 7% fewer export destinations, and the study traces the effect to the cost of finding and matching partners. That is causal evidence, though it comes from pre-AI export trade fairs outside B2B software.\n\nThe challenge needs more than that. Buyers do lean on trusted human sources to check what AI tells them, Forrester finds. But the step to \"events now return more than before\" has no before-and-after or cohort data behind it. Its support is opinion, including a Forrester Consulting study that Cvent commissioned. Both conclusions rest on returns nobody has measured."
    },
    {
     "heading": "What both sides miss",
     "text": "Gartner holds both positions at once. It reports AI pushing CMOs away from events toward channels AI can optimize, and in the same release warns them not to let AI steer budget toward \"the channels and stages of the journey that are easiest to tune.\" Its most AI-mature organizations put a lower share of budget into digital.\n\nNeither side is measuring the event well. Forrester finds teams whose event platform is deeply tied into their stack more satisfied, 62% against 37%, and finds only 38% of CMOs agreeing that events perform well, against about half of organizations overall. The fight over whether to cut is mostly a fight over attribution, which has its own entry."
    }
   ],
   "open_question": "If AI makes every digital touch cheap, is an hour in the same room worth more now, or does it only feel that way?",
   "sources": [
    {
     "side": "origin",
     "who": "Gopalakrishna, Lilien, Williams and Sequeira, Journal of Marketing, Do Trade Shows Pay Off? (one firm, one show)",
     "date": "1995-07",
     "url": "https://doi.org/10.2307/1252121",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Ewan McIntyre, Gartner 2026 CMO Spend Survey press release",
     "date": "2026-06-08",
     "url": "https://www.gartner.com/en/newsroom/press-releases/2026-06-08-gartner-marketing-survey-finds-awareness-and-conversion-account-for-62-6-of-total-media-spend",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Conrad Mills and Rosa Tardio, Forrester, State of B2B Events Q1 2026",
     "date": "2026-06-25",
     "url": "https://www.forrester.com/blogs/the-next-era-of-b2b-events-eight-data-backed-shifts-defining-2026/",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Bai, Ma and Makioka, RIETI, Canton Fair during SARS (difference-in-differences)",
     "date": "2022-08",
     "url": "https://www.rieti.go.jp/jp/publications/dp/22e078.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Reggie Aggarwal, CEO of Cvent (events vendor), The Presence Premium",
     "date": "2026-07-14",
     "url": "https://www.cvent.com/en/blog/events/presence-premium",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "events",
    "budget",
    "attribution"
   ],
   "bench_notes": []
  },
  {
   "slug": "partner-sourced-pipeline",
   "title": "Partner-sourced pipeline beats marketing-sourced pipeline",
   "summary": "Too early to call: every win-rate figure for partner deals is vendor-reported and exposed to selection, the 28 buying moments have no published method, and the claim that AI agents bypass partners is one analyst prediction.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "mostly-aspiration",
   "because_label": "the case to update is mostly aspiration",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "outbound and demand",
   "altitude": "L2 Operating model",
   "origin": "Jay McBain, Canalys (insight dated 2023-12-29, said on podcasts since at least 2022): the modern buyer experiences 28 moments on the way to a purchase, many of them with or through partners. No method for the 28 was found. The win-rate half comes from ecosystem vendors: Crossbeam's network data shows an 11.7% average win-rate lift when partners are involved.",
   "rationed": "Buyer access and trust. Vendors couldn't see most of a buyer's research, and a trusted third party lent credibility a vendor's own marketing couldn't.",
   "challenge_claim": "AI buying agents will take an express route past the moments where partners influence a purchase, so partner value moves from influence before the sale to accountability after it.",
   "blindspot": "Defenders read correlational, self-tagged win rates as proof the partner caused the win, when partners tend to bring warm accounts. The update side rests on one prediction from inside the same analyst group that coined the 28 moments.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Most of the moments that shape a purchase happen outside the vendor's view, many of them through partners.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "Partner-sourced deals win more often because of the partner, not because partners bring deals that were already likely to close.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI buying agents skip the pre-sale moments where partners exert influence.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The channel's share of buying is falling.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "Partner-sourced deals win more often because of the partner, not because partners bring deals that were already likely to close.",
    "deciding_side": "prior",
    "what_follows": "Nothing has changed because nothing was established: the prior's causal claim was never separated from selection, and the update's causal claim is a forecast. The one premise that holds, a small drop in the channel's share of IT spend, says nothing about whether a partner-sourced deal wins because of the partner."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Two claims travel together. Jay McBain of Canalys, an analyst firm covering the IT channel, says the modern buyer passes through 28 moments on the way to a purchase, many of them with or through partners. On podcasts he has said direct marketing might win four of the 28. No published method for the number turned up.\n\nThe second claim, that partner-sourced pipeline beats marketing-sourced, comes from ecosystem vendors. Crossbeam, which sells partner-data software, reports from its own network an 11.7% average win-rate lift when partners are involved. A 2023 survey Crossbeam co-produced has partner deals 53% more likely to close, self-reported by respondents."
    },
    {
     "heading": "What people say now",
     "text": "The challenge is a prediction. Sheena Wee of Omdia, which now owns Canalys, told a 2026 channel conference that AI agents will take an express route past the 28 moments, and that \"the partner's value actually moves from the influence before the sale to the accountability after it.\"\n\nThe market data is mixed. Another Omdia analyst, Alastair Edwards, reports the channel's share of global IT spend fell from 69% to 65% in a year, driven by direct hyperscaler AI deals. He adds that partner influence is still increasing. Both voices come from the same corporate family as the origin."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on two ideas. Most of what shapes a purchase happens out of the vendor's sight, much of it through partners. And partner deals win more because of the partner. The first is a number with no method. The second is the one that matters, and it has never been separated from selection. Partners tend to bring accounts that are already warm and qualified, and partner influence is often self-tagged. A higher win rate on those deals is what you'd see whether or not the partner made the difference.\n\nThe challenge doesn't help. Its causal claim, that agents bypass partner influence, is a forecast. The premise that holds, a four-point drop in channel share, tracks where money flows without showing whether a partner-sourced deal wins because of the partner."
    },
    {
     "heading": "What both sides miss",
     "text": "The closest outcome evidence is from consumer banking. In a 2011 study, Schmitt, Skiera and Van den Bulte found referred bank customers at least 16% more valuable, through better matching and lower churn. A replication confirmed the lower churn but found referred customers not necessarily more valuable. If referrals help mainly through matching, partners may be selecting better deals more than closing them. That would still be worth paying for, but it would change how partner pipeline is credited and forecast.\n\nNobody has run the comparison of partner-sourced and direct deals on accounts of similar warmth that would settle it."
    }
   ],
   "open_question": "If you compared partner-sourced deals with direct deals on accounts of the same warmth, would the lift survive?",
   "sources": [
    {
     "side": "origin",
     "who": "Jay McBain, Rodney Hur and Lisa Lawson, Canalys (no method published)",
     "date": "2023-12-29",
     "url": "https://www.canalys.com/insights/through-channel-marketing-automation",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Evie Nagy, Crossbeam (ecosystem software vendor), own network data",
     "date": "2024",
     "url": "https://insider.crossbeam.com/entry/new-data-involving-partners-in-deals-increases-win-rate-for-nearly-every-ecosystem-size-and-type",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Sheena Wee (Omdia, which owns Canalys), keynote prediction, reported by Doug Woodburn, IT Channel Oxygen",
     "date": "2026-10-13",
     "url": "https://itchanneloxygen.com/partners-told-to-prepare-for-agent-invasion-of-hyperscaler-marketplaces/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Alastair Edwards (Omdia), reported by Channel Dive (an Omdia sister publication)",
     "date": "2026-07-21",
     "url": "https://www.channeldive.com/news/the-channel-owns-less-than-two-thirds-of-global-it-spend/825825/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Schmitt, Skiera and Van den Bulte, Journal of Marketing (German bank, about 10,000 customers)",
     "date": "2011",
     "url": "https://faculty.wharton.upenn.edu/wp-content/uploads/2012/04/Schmitt-Skiera-vandenBulte-2011-Referral-Programs-Customer-Value.pdf",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Armelini, Barrot and Becker, IJRM (replication)",
     "date": "2015-12",
     "url": "https://doi.org/10.1016/j.ijresmar.2015.09.004",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "partners",
    "pipeline",
    "attribution"
   ],
   "bench_notes": []
  },
  {
   "slug": "price-per-seat",
   "title": "Price per seat",
   "summary": "Update it: agents break the link between users and value and inference makes heavy users costly to serve, so seats alone stop fitting; pure usage trails on margin, and hybrid pricing has partial outcome support.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "unit economics",
   "altitude": "L2 Operating model",
   "origin": "No originator; it was the default assumption of the SaaS era. Tomasz Tunguz, relaying Madhavan Ramanujam in 2021, set per-seat companies (Salesforce, Adobe) against per-use ones (Snowflake, Twilio). Kyle Poyar's survey has seats plus flat-rate subscriptions covering about half of companies as late as early 2024.",
   "rationed": "Marginal cost. One more user cost almost nothing to serve, so price could follow users, and users stood in for value.",
   "challenge_claim": "Agents don't sit in seats. When software does the work, per-seat pricing breaks, and usage or outcome pricing replaces it.",
   "blindspot": "The update side is mostly founders of products that agents consume, and it skips the outcome data: pure usage pricing posts the highest NRR but the lowest gross margin. The defend side misses that seat cuts are now buyers' main lever on software spend.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Value tracks the number of people using the product.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Serving one more user costs almost nothing, so pricing by user doesn't hurt margin.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Buyers and their CFOs need spend they can predict.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Agents doing the work break the link between users and value.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Buyers will accept a bill that moves with use.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "Pure usage or outcome pricing replaces seats.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Value tracks the number of people using the product.",
    "deciding_side": "prior",
    "what_follows": "Two of the prior's premises changed: value no longer tracks users, and with inference costs, serving a heavy user is no longer free. That is enough to update seat-only pricing. Pure usage doesn't follow: in cross-sectional benchmarks it has the highest NRR but the lowest gross margin. A fixed fee plus metering has partial support, from the same benchmarks and from tariff research outside software, but no AI-specific outcome data exist yet."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody invented per-seat pricing; it was simply how SaaS was sold. Tomasz Tunguz, then a VC at Redpoint, relaying Madhavan Ramanujam, drew the line in 2021: Salesforce and Adobe price per seat, Snowflake and Twilio per use. Until recently, seats and flat-rate subscriptions covered about half the companies Kyle Poyar surveys."
    },
    {
     "heading": "What people say now",
     "text": "The loudest voices say the seat is dying, and most of them sell products that agents use. One founder: \"Pricing agents by the seat is like pricing electricity by the bulb.\" Another: \"Per-seat assumes a human logs in.\" Poyar, whose newsletter is sponsored by a usage-billing vendor, finds hybrid models at 37% of companies, up from 25% a year earlier, and only 5% of investors prefer seats.\n\nThe market hasn't followed the slogan. SBI, a pricing consultancy, finds more than 80% of 321 SaaS companies still use some seat metric, with outcome pricing at two companies. Bain reviewed more than 30 vendors adding generative AI: about 65% layered an AI meter on top of seats, about 35% raised seat prices, and none moved fully to usage or outcomes. This describes vendor choices rather than results."
    },
    {
     "heading": "Does the logic hold",
     "text": "Seat pricing rested on three ideas. Value tracks the number of people using the product. Serving one more user costs almost nothing. And buyers need spend they can predict. AI changed the first two. An agent can do a seat's work without a login, and inference makes a heavy user costly to serve. ICONIQ, a growth investor, has AI products running gross margins of 41% to 45%, against the 80% SaaS assumed. Those two changes are enough to update a pure seat price.\n\nThe challenge goes further, to pure usage, and the outcome data doesn't carry it there. In Benchmarkit's 2026 survey of 342 companies, usage-only pricing had the highest median NRR, 108% against 98% for seat-only. It also had the lowest total gross margin, 62% against 76% for seats. The survey is vendor-sponsored, cross-sectional and confounded by category, and it doesn't isolate AI products.\n\nWhether buyers accept a bill that moves with use is unproven. Cursor, Replit and GitHub Copilot each pushed users toward metering and met a backlash, and Cursor's CEO wrote that the change \"hurt that trust.\" Yet usage-only firms post the highest NRR, so their buyers do pay more as use grows. No one has published what the switches did to revenue or churn.\n\nHybrid is partly supported. Fixed fee plus usage has the top Rule of 40 scores in Benchmarkit's data. Outside software, a telecom pricing field experiment found a two-part tariff, an access fee plus a usage price, lowered retention and usage against pure pay-per-use yet still made the most profit."
    },
    {
     "heading": "What both sides miss",
     "text": "The backlash was aimed at variance more than at metering. Each of the three companies restored predictability with spend limits, credits or a seat plus a meter, and none went back to seats.\n\nThe defenders miss the pressure from the buying side. BCG, a consultancy, finds 40% of IT buyers name seat reduction as their main lever to cut software spend. When agents do part of the work, the seat count is the first number a buyer's CFO goes after. A pure seat model loses ground even where seats survive as a unit."
    }
   ],
   "open_question": "When agents do the work, what unit will a buyer's CFO agree to predict?",
   "sources": [
    {
     "side": "origin",
     "who": "Tomasz Tunguz, then at Redpoint Ventures (invests in software), \"Per Seat or Per Use Pricing\" (relaying Madhavan Ramanujam)",
     "date": "2021-08-29",
     "url": "https://tomtunguz.com/seat-vs-usage-based-pricing/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Kyle Poyar, Growth Unhinged (sponsored by a usage-billing vendor), \"The state of B2B monetization in 2026\"",
     "date": "2026-05-13",
     "url": "https://www.growthunhinged.com/p/the-state-of-b2b-monetization-in-2026",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Benchmarkit (benchmarking firm) with Aleph (FP&A vendor, co-sponsor), 2026 SaaS & AI-Native Metrics Benchmarks",
     "date": "2026-06",
     "url": "https://www.benchmarkit.ai/2026-saas-ai-native-metrics",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Michael Truell, CEO Cursor (sells an AI coding tool), on the June 2025 pricing change",
     "date": "2025-07-04",
     "url": "https://cursor.com/blog/june-2025-pricing",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Iyengar, Jedidi, Essegaier and Danaher, Marketing Science (telecom pricing field experiment)",
     "date": "2011",
     "url": "https://doi.org/10.1287/mksc.1110.0655",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "pricing",
    "unit economics",
    "finance"
   ],
   "bench_notes": []
  },
  {
   "slug": "magic-number",
   "title": "A magic number above 0.75 means put more into sales and marketing",
   "summary": "Update it: the 0.75 line has no traceable source and nobody has tested whether spending more above any line pays back, so read the magic number as a prompt to look at cohorts, not a trigger to spend.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "never-sourced",
   "because_label": "never sourced",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "unit economics",
   "altitude": "L2 Operating model",
   "origin": "Rory O'Driscoll, Scale Venture Partners, \"Magic Number Math,\" 2010: the annualised change in quarterly subscription revenue divided by the prior quarter's sales and marketing spend. Above 1x is \"a compelling business investment\"; below 0.5x the company hasn't figured out its model. His page doesn't mention 0.75. That line traces only to a 2008 guest post through a secondary source, and the trail ends there.",
   "rationed": "Capital and board attention. With revenue lagging spend, boards needed a quick read on whether the next sales and marketing dollar would come back, and one quarterly ratio stood in for the cohort-level CAC analysis most companies couldn't run.",
   "challenge_claim": "The magic number barely tracks value, and AI-native economics, with thin margins and growth that comes through the product, make the SaaS efficiency ratios misleading.",
   "blindspot": "Defenders keep publishing the line as a benchmark, which is not the same as testing it. Critics lean on the ratio's weak link to valuation, which says little about whether the next dollar of spend pays back, and their AI case has no data behind it.",
   "logic": {
    "prior_premises": [
     {
      "premise": "The 0.75 line was set from evidence of where the next dollar of sales and marketing pays back.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A magic number above the line tells you an extra dollar of sales and marketing will pay back.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Revenue added in a quarter comes from the prior quarter's sales and marketing spend.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "The magic number is a weak guide to go-to-market efficiency, with or without AI.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "AI-native economics, with thin gross margins and fast self-serve growth, make SaaS efficiency ratios misleading.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "The 0.75 line was set from evidence of where the next dollar of sales and marketing pays back.",
    "deciding_side": "prior",
    "what_follows": "The premise that fails is the line itself: the originator used 1x and 0.5x as judgment calls, and 0.75 has no traceable source. Whether spending more above any line pays back was never tested, and the AI-specific case is unproven, so the update rests on the missing source, not on AI. Stop treating 0.75 as a spending trigger; the ratio stays a quarterly signal worth a closer look."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Rory O'Driscoll of Scale Venture Partners, a VC, wrote it down in 2010. Take the change in quarterly subscription revenue, annualise it, and divide by the prior quarter's sales and marketing spend. Above 1x, he called more spend \"a compelling business investment.\" Below 0.5x, the company hadn't figured out its model yet. Credit for the metric is contested, but the ratio is well sourced.\n\nThe line most teams quote, 0.75, is not. It isn't on O'Driscoll's page. One trail leads to a 2008 guest post through a secondary source and ends there. Boards used the ratio because revenue lags spend and they needed a quick read on whether the next dollar would come back. One quarterly number stood in for a cohort analysis most companies couldn't run."
    },
    {
     "heading": "What people say now",
     "text": "The case against it predates AI. McKinsey, a consultancy, looked at about 20 operating metrics across public software companies in 2021 and found the magic number showed \"almost no correlation\" with value multiples. Ray Rike, who sells benchmarks, puts its R-squared against revenue multiples at 0.07 and calls it never a good primary metric. These are the same kind of value studies the LTV:CAC entry in this series uses for payback.\n\nThe AI case is newer and thinner. Bessemer Venture Partners reports its fastest-growing AI companies running gross margins near 25%, often negative, and about $1.13 million of revenue per employee. The argument is that SaaS ratios mislead for companies like that. No source with data makes it, and High Alpha, a VC, surveyed more than 800 companies and found AI-native firms scoring better on the Rule of 40 than AI-enabled ones, 34% against 22%."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior turns a ratio into a spending rule, and it needs three ideas. The line marks where the next dollar pays back. A ratio above the line tells you that it will. And a quarter's new revenue comes from the prior quarter's spend.\n\nThe first fails. The originator drew 1x and 0.5x as judgment calls, and nobody can trace 0.75 to a derivation or a test. The second is unproven. The ratio averages last quarter's spend against revenue change; no source tests whether it predicts the return on the marginal dollar, and its link to value is weak. The third is the built-in lag. For products that grow through self-serve and usage, revenue may move with the product rather than with last quarter's spend. That is plausible, but no source with data shows it.\n\nSo the update doesn't need AI. The line was never sourced, and the causal claim under it was never tested. That leaves the ratio as a signal to dig into cohorts, not a rule to spend."
    },
    {
     "heading": "What both sides miss",
     "text": "Defenders point to the benchmark tables that still publish the ratio. A benchmark describes where companies sit; it doesn't show that spending more above the line paid off.\n\nCritics lean on valuation, which is not the return on sales and marketing. A ratio can track value poorly and still flag a broken motion. Nobody has tested whether companies that raised spend after crossing the line got the return.\n\nIts cousins fare differently. The Rule of 40 tracks value better in the same studies, and the burn multiple has no outcome evidence at all."
    }
   ],
   "open_question": "Has anyone who raised spend after crossing the line checked whether the next dollar came back?",
   "sources": [
    {
     "side": "origin",
     "who": "Rory O'Driscoll, Scale Venture Partners (VC), Magic Number Math",
     "date": "2010-04-20",
     "url": "https://www.scalevp.com/blog/magic-number-math",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "McKinsey (consultancy; Roche, Tandon), SaaS and the Rule of 40",
     "date": "2021-08-03",
     "url": "https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/saas-and-the-rule-of-40-keys-to-the-critical-value-creation-metric",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Ray Rike, SaaS Barometer (sells benchmarks), \"SaaS Magic Number is not magic\"",
     "date": "2024-09-30",
     "url": "https://thesaasbarometer.substack.com/p/saas-magic-number-is-not-magic",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Bessemer Venture Partners (VC), State of AI 2025",
     "date": "2025-08-13",
     "url": "https://www.bvp.com/atlas/the-state-of-ai-2025",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Mollie Kuramoto, High Alpha (VC), AI-native benchmarks survey, 800+ responses",
     "date": "2025-12-12",
     "url": "https://highalpha.com/blog/does-it-pay-to-be-ai-native-growth-rates-for-ai-companies",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "efficiency",
    "sales and marketing spend",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "quarter-end-discounts",
   "title": "Quarter-end discounts close the gap to the number",
   "summary": "Update it: the best deal-level evidence shows quarter-end discounts mostly move deals between quarters at a lower price, costing several percent of revenue, so keep bounded discount authority and stop treating the quarter-end discount as found money.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "evidence-against",
   "because_label": "evidence against it",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "unit economics",
   "altitude": "L3 Practice",
   "origin": "No single coiner. Discount authority that rises with rank, rep to manager to VP, is the standard deal-desk pattern. Jason Lemkin's \"The Confounding Logic of Discounting\" (SaaStr, 2021) is the best-known practitioner statement: reps discount at the end of the month or quarter, so control discounts rather than eliminate them, and mark list up to leave room. Paul Oyer (QJE, 1998) first showed sales bunching at fiscal-year ends under nonlinear pay.",
   "rationed": "Information and managerial attention. Reps knew what a buyer would pay and managers didn't, so firms delegated some pricing and escalated the big exceptions; the quarter-end discount was the lever for a period number that comp plans and public reporting made binding.",
   "challenge_claim": "Discounting is how deals close, so control it with an approval ladder and use it at quarter end to land the number. The newer version says AI pricing guidance and buyer-side negotiating agents will take over the discount decision.",
   "blindspot": "The update side borrows AI salience it hasn't earned: the evidence against quarter-end discounting predates AI, and AI is only a candidate remedy. The defend side treats the approval ladder as the control, when the gaming in the best study happened inside one.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Reps hold customer information managers lack, so bounded rep discretion with escalation prices better than a fixed price.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "End-of-quarter discounts win revenue that would otherwise be lost, rather than moving deals between quarters at a lower price.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "An approval ladder is enough to contain discount leakage.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Model-based price guidance beats seller discretion on realized margin.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Buyer-side AI negotiators remove the quarter-end lever.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Deploying the product beats discounting as the closing move.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "End-of-quarter discounts win revenue that would otherwise be lost, rather than moving deals between quarters at a lower price.",
    "deciding_side": "prior",
    "what_follows": "The premise that fails is the quarter-end one, and it failed on deal data that predates AI: the discounts track the rep's commission stake and mostly pull or push deals rather than win new ones. Bounded rep discretion still holds, so the update is to the quarter-end habit, not to discount authority. AI enters only as a remedy, and only model price guidance has a controlled test behind it."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody coined it. Most deal desks use discount authority that climbs with rank, where a rep can give a little, a manager more, and a VP more again. Jason Lemkin, who runs SaaStr's events and media business and invests in SaaS companies, gave the plain statement of the habit in a 2021 essay. Reps quote list, then discount at the end of the month or quarter, as far as they're pushed. His advice was to control discounting rather than stop it, and to mark list up so there's room.\n\nThe economics trail is older. In 1998 Paul Oyer showed that manufacturing sales run high at fiscal-year end and low at the start, and argued that nonlinear pay gives sellers a reason to move both prices and timing."
    },
    {
     "heading": "What people say now",
     "text": "The practice is mostly defended by being normal. The newer argument is that AI takes the discount decision over. One version gives the seller a model-set price. Another says buyer-side negotiating agents will strip out the quarter-end lever. Lemkin himself now argues the other way, urging teams to stop discounting and get the product deployed instead. He offers anecdotes and a worked example without cohort data.\n\nVendr, which sells software purchasing and negotiation, has data that cuts against the agent story. Across 12,446 contracts its CEO finds discounts \"fairly stable throughout the year,\" about 5.5 points between the best and worst month. Buyers' timing leverage was already modest before agents arrived."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on two ideas. Reps know what a buyer will pay, so bounded discretion with escalation prices better than a fixed list. And a quarter-end discount wins revenue that would otherwise be lost. The first holds, as a study of 181 German industrial firms found delegating price authority improves performance, more so when reps know more than their managers.\n\nThe second fails. Ian Larkin's 2014 study of one enterprise software vendor found reps pulling deals forward and pushing them back around their commission schedule, and discounting the moved deals more. The discounts tracked the rep's commission stake rather than the deadline. The lost revenue came to 6 to 8%, about what the vendor spent on commissions. That gaming happened inside the vendor's approval process, so the ladder alone didn't contain it.\n\nNone of this needs AI. The diagnosis predates it. AI only matters as a remedy, and one remedy has a test: in a randomized field experiment at a B2B retailer, showing reps a model's price raised profit on treated quotes by 11%."
    },
    {
     "heading": "What both sides miss",
     "text": "The same experiment found reps still priced complex or unusual quotes better than the model, and a human with the model beat either alone. That is the case for keeping bounded discretion while moving the anchor to a model price.\n\nDefenders of the quarter-end push miss that it mostly buys timing at a lower price. Critics who wait for buyer agents miss that the cost is already on the sell side, in how the comp plan is shaped."
    }
   ],
   "open_question": "If the quarter-end discount mostly buys timing, what is the number you hit with it actually worth?",
   "sources": [
    {
     "side": "origin",
     "who": "Jason Lemkin, SaaStr (runs SaaS events and media, invests in SaaS companies), The Confounding Logic of Discounting",
     "date": "2021-03-27",
     "url": "https://www.saastr.com/the-confounding-logic-of-discounting/",
     "standing": "position"
    },
    {
     "side": "origin",
     "who": "Paul Oyer, Fiscal Year Ends and Nonlinear Incentive Contracts, QJE (manufacturing industry data)",
     "date": "1998-02",
     "url": "https://doi.org/10.1162/003355398555559",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Ian Larkin, The Cost of High-Powered Incentives, Journal of Labor Economics (one enterprise software vendor's deal data)",
     "date": "2014",
     "url": "https://www.hbs.edu/ris/Publication%20Files/13-073_cbb24c28-9e84-47d9-8a32-f01b73cfda13.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Yael Karlinsky-Shichor and Oded Netzer, Marketing Science (randomized field experiment, B2B retailer)",
     "date": "2024",
     "url": "https://pubsonline.informs.org/doi/10.1287/mksc.2023.1449",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Frenzen, Hansen, Krafft, Mantrala and Schmidt, IJRM (181 German industrial firms)",
     "date": "2010-03",
     "url": "https://doi.org/10.1016/j.ijresmar.2009.09.006",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Ryan Neu, CEO of Vendr (sells software purchasing and negotiation), discount by month across 12,446 contracts",
     "date": "2026-01-20",
     "url": "https://www.linkedin.com/posts/ryanneu_trying-to-time-software-purchases-around-activity-7419408827990708224-eoj0",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "pricing",
    "discounting",
    "comp"
   ],
   "bench_notes": []
  },
  {
   "slug": "ltv-cac-and-payback",
   "title": "LTV:CAC of 3 or more, and CAC payback under 12 months",
   "summary": "Too early to call: the claim that AI margins make the ratio obsolete fails because LTV was always margin-based, but nobody has tested whether hitting 3:1 or a 12-month payback marks a healthier company, and payback tracks valuation only loosely.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "unit economics",
   "altitude": "L1 Axiom",
   "origin": "David Skok (Matrix Partners), SaaS Metrics, 2010: \"My rule of thumb is that LTV must be at least 3x greater than CAC,\" and recover CAC in under 12 months. He called them guidelines, built LTV on gross margin, and later loosened payback to about 20 months himself.",
   "rationed": "Near-zero marginal cost. With software gross margins around 80%, revenue stood in for profit, and one ratio could summarise a go-to-market motion.",
   "challenge_claim": "AI products run gross margins far below 80%, so the 3:1 ratio is obsolete for AI software and payback has to be recomputed on gross profit.",
   "blindspot": "The update side attacks a version nobody held: Skok's LTV was already margin-based, so the proposed fix is the prior. The defend side treats 3x and 12 months as tested health markers, when nobody has checked whether companies that hit them do better.",
   "logic": {
    "prior_premises": [
     {
      "premise": "With gross margins near 80%, revenue could stand in for profit.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "A company that recovers its CAC within a year is healthier than one that doesn't.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "One ratio, set as a target, can steer a go-to-market motion.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI products run gross margins well below the SaaS norm.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "Lower margins make the 3:1 rule obsolete.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Payback should be computed on gross margin, with inference in cost of goods.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A company that recovers its CAC within a year is healthier than one that doesn't.",
    "deciding_side": "prior",
    "what_follows": "Margins changed for AI products, but Skok built LTV on gross margin, so lower margins flow into the ratio rather than breaking it, and the obsolescence argument fails on definition. The deciding premise, that companies at 3:1 or under 12 months do better, was never tested: payback correlates only loosely with valuation, and survey medians above 12 months show the market missing the target, not the target failing. Compute payback on gross margin and hold both thresholds as untested."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "David Skok, a VC at Matrix Partners, wrote both rules in 2010: LTV should be at least three times CAC, and CAC should pay back in under twelve months. He called them his rules of thumb, not findings, and his LTV was calculated on gross margin from the start. He later wrote that the twelve-month guideline came from a time when capital was hard to raise, and moved healthy payback to about twenty months."
    },
    {
     "heading": "What people say now",
     "text": "The AI-era claim is that the ratio is obsolete. AI products run gross margins well below the 80% that SaaS assumed. ICONIQ, a growth investor, reports 41% in 2024 rising to 45% in 2025. Advisors' posts now call the 3:1 ratio obsolete for AI software.\n\nFinance writers mostly shrug. Seth Girsky of Inflection CFO: \"The 3:1 ratio was an output, not an input.\" Ben Murray, writing from a CFO community: \"Cap LTV or skip it. CAC payback is what most finance leaders actually manage to.\" Finance keeps payback and changes the accounting under it, calculating payback on gross margin with inference costs broken out in cost of goods."
    },
    {
     "heading": "Does the logic hold",
     "text": "Skok's rules rested on three ideas. Software margins were high enough that revenue could stand in for profit. A company that got its acquisition spend back within a year was healthier than one that didn't. And one ratio could summarise whether a go-to-market motion was working.\n\nAI changed the first idea for AI products. The challenge concludes that the ratio is obsolete, and that step fails on definition. Skok's LTV was margin-based from the start, so lower margins already lower it. The proposed fix is the original rule.\n\nThe second idea decides the verdict, and nobody has tested it. No study checks whether companies at or above 3:1, or under twelve months, go on to grow faster, survive longer or earn higher multiples. The closest evidence is correlation with valuation, and it is loose. Across public SaaS, CAC payback explained between about 10% and 34% of the spread in revenue multiples from quarter to quarter, by the count of Ray Rike, who sells benchmarks. A VC's longer series, from DNX Capital, put it at roughly 20 to 30%. Growth explains more than any efficiency measure. So efficiency matters somewhat, and the cut-offs are untested."
    },
    {
     "heading": "What both sides miss",
     "text": "The medians are market state, not a verdict. KeyBanc's 2024 private-SaaS survey estimates median fully loaded payback at 20 months for that year. CJ Gustafson's tracking of the top ten public software companies has their median at about 28 months. Those figures show that most companies miss twelve months. They don't show that missing it does no harm, so a threshold can't be retired on the evidence that few meet it, any more than it can be defended on Skok's say-so.\n\nAI does change the inputs. Margin-adjusted payback gets longer, and one benchmark publisher reports search channels getting less efficient, which raises CAC."
    }
   ],
   "open_question": "If nobody has shown that a 12-month payback predicts a healthier company, what should set the bar for yours?",
   "sources": [
    {
     "side": "origin",
     "who": "David Skok, Matrix Partners (invests in SaaS), For Entrepreneurs, SaaS Metrics",
     "date": "2010-02-17",
     "url": "https://www.forentrepreneurs.com/saas-metrics/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "ICONIQ (growth investor), State of AI 2025",
     "date": "2025",
     "url": "https://www.iconiq.com/growth/reports/2025-state-of-ai",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "David Skok, SaaS Metrics 2.0 definitions",
     "date": "2013",
     "url": "https://www.forentrepreneurs.com/saas-metrics-2-definitions/",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Ray Rike, Benchmarkit (sells benchmarks), which SaaS metric correlates with EV/revenue",
     "date": "2025-07-28",
     "url": "https://www.linkedin.com/posts/rayrike_which-saas-metric-is-most-correlated-to-ev-activity-7355620066165481473-AZCp",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "unit economics",
    "finance",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "acv-sets-the-motion",
   "title": "ACV sets the sales motion: self-serve at the low end, inside sales in the middle, field sales at the top",
   "summary": "Too early to call: AI may make a sales touch cheap enough to sell below the old ACV floor, but that case is a model and one anecdote, while the best-evidenced link, that higher-ACV customers stay longer, still holds.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "mostly-aspiration",
   "because_label": "the case to update is mostly aspiration",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "unit economics",
   "altitude": "L2 Operating model",
   "origin": "Christoph Janz (Point Nine), \"Five ways to build a $100 million business,\" 2014: flies at about $10 a year per account, mice at $100, rabbits at $1k, deer at $10k and elephants at $100k+, each needing its own way to acquire customers. On deer: \"An ARPA of $10k per year usually isn't enough to make traditional enterprise field sales work.\" On rabbits: \"cold calling usually doesn't work at this ARPA level.\"",
   "rationed": "The cost of a human sales touch. A rep costs six figures a year and closes a bounded number of deals, so ACV times deals closed has to clear that cost several times over. Below the line, only marketing and product can carry acquisition.",
   "challenge_claim": "AI cuts the cost of a sales touch, so sales-assisted selling now pays at ACVs that used to need self-serve, while AI products self-serve their way up to enterprise deals.",
   "blindspot": "The update side skips retention: cheap, self-serve products churn hardest, so selling at a lower ACV doesn't fix what happens after the sale. The defend side reads the map as fixed while some AI products reach enterprise scale before they hire enterprise sellers.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A human sales touch has a fixed cost that sets a floor on viable ACV.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Low-ACV buyers prefer to buy without talking to sales.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "ACV predicts customer durability, so higher-ACV customers stay longer.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI cuts the cost of a sales touch enough that sales-assisted selling pays at ACVs that used to need self-serve.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "AI personalisation removes the old trade-off between scale and personal service.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Self-serve can now reach enterprise ACVs, so ACV no longer fixes the entry motion.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "AI cuts the cost of a sales touch enough that sales-assisted selling pays at ACVs that used to need self-serve.",
    "deciding_side": "challenge",
    "what_follows": "The prior's own premises hold. The top of the map has moved for some AI categories, where self-serve reaches enterprise spend, but sales still arrives later, which is Janz's elephant logic arriving late rather than failing. The bottom of the map moves only if AI makes a sales touch cheap enough to pay at small ACVs, and that rests on a model and one anecdote, so the call waits on outcome data."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "In 2014 Christoph Janz of Point Nine, a VC, sorted the ways to build a $100 million software business by what each customer pays: flies at about $10 a year, mice at $100, rabbits at $1,000, deer at $10,000 and elephants at $100,000 or more. Each needs its own way to acquire customers. At $10,000, he wrote, an account \"usually isn't enough to make traditional enterprise field sales work,\" so you use inside sales. At $1,000, \"cold calling usually doesn't work.\"\n\nRevOps read it as a map where ACV picks the motion. The scarcity under it was the cost of a human sales touch."
    },
    {
     "heading": "What people say now",
     "text": "The update case comes from two VCs. Tomasz Tunguz of Theory Ventures notes that most companies hire sellers at $15,000 or more in contract value, then models what AI might do. Assume a 50% productivity gain, and a seller at a $10,000 ACV goes from 24 to 36 deals a year. \"The minimum viable ACV drops.\" Jeffrey Bussgang of Flybridge offers one example: a company selling to small businesses at about $3,000 a year with AI-personalised outbound, a segment he says was unprofitable to reach that way before.\n\nThe top of the map is moving the other way. Menlo Ventures, a VC, surveyed about 500 enterprise buyers and puts product-led purchases at 27% of AI application spend, against 7% in traditional software. It notes Cursor reached $200 million in revenue before hiring enterprise sellers."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on three ideas. A human sales touch has a fixed cost, which sets a floor on the ACV it can serve. Low-ACV buyers would rather not talk to sales. And higher-ACV customers stay longer. All three hold. Nobody argues a person's cost per touch has fallen, and the retention data below still favors larger contracts.\n\nThe challenge needs AI to cut the cost of a sales touch enough that selling pays below the old floor. That is the deciding premise, and it rests on a model with an assumed productivity gain and one company's story. No cohort or before-and-after study was found. The top end is better supported, but even there sales gets layered on once a product reaches large accounts. That is Janz's elephant logic arriving later, not failing.\n\nSo the two AI effects pull in opposite directions, and the map blurs at both ends. Neither end has outcome data. The case to update is mostly projection."
    },
    {
     "heading": "What both sides miss",
     "text": "The update side skips retention, the best-evidenced premise here. SaaS Capital, a lender to SaaS companies, finds gross retention of 95% above $250,000 ACV against about 91% below. Selling at a lower ACV doesn't fix the churn that comes with a lower ACV.\n\nThe defend side misses what the map never claimed. SaaS Capital also finds ACV doesn't correlate with growth rate: the map tells you how to sell, not how fast you'll grow.\n\nThe product-led growth entry in this series looks at the low end from the other side. Free users got costlier just as sales touches may be getting cheaper, and both push sales assistance down-market."
    }
   ],
   "open_question": "If AI makes a small deal worth a sales touch, will that customer stay long enough to pay for it?",
   "sources": [
    {
     "side": "origin",
     "who": "Christoph Janz, Point Nine (VC), Five ways to build a $100 million business",
     "date": "2014-10-27",
     "url": "https://writing.pointnine.com/p/five-ways-to-build-a-100-million",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Jeffrey Bussgang, Flybridge (VC) and HBS, \"AI is breaking this chart\"",
     "date": "2024-12-06",
     "url": "https://www.linkedin.com/posts/bussgang_ai-is-breaking-this-chart-this-insightful-activity-7270849441752924161-h9_U",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Tomasz Tunguz, Theory Ventures (VC), The Marginal Task (a model, no outcome data)",
     "date": "2024-08-15",
     "url": "https://tomtunguz.com/the-marginal-task/",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "SaaS Capital (lender), retention by ACV, private SaaS survey",
     "date": "2025-09-18",
     "url": "https://www.saas-capital.com/blog-posts/what-is-a-good-retention-rate-for-a-private-saas-company/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Menlo Ventures (VC), State of Generative AI in the Enterprise 2025, survey of about 500 enterprise buyers",
     "date": "2025-12-09",
     "url": "https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "acv",
    "sales motion",
    "unit economics"
   ],
   "bench_notes": []
  },
  {
   "slug": "health-score-predicts-churn",
   "title": "The health score tells you who will churn",
   "summary": "Update it: a red, yellow and green score weighted by judgment was never shown to predict churn, while models trained on past renewals have been, so replace the colour you configure with a model you backtest, and don't assume generative AI has already fixed it.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "customer and renewal",
   "altitude": "L3 Practice",
   "origin": "No traceable origin. The practice grew with the customer-success platforms of the 2010s (Gainsight, Totango, ChurnZero), each of which ships a red, yellow and green scorecard whose weights the customer configures. Gainsight's own 2019 description of its renewal model treats the health score as one input, \"determined by you\". The trail ends in product documentation, not a study.",
   "rationed": "CSM attention. A book held dozens to hundreds of accounts and usage data sat in several tools, so one composite number let a manager rank the book without reading every account.",
   "challenge_claim": "Traditional health scores fail to predict churn, and models that read usage, calls and email will replace them.",
   "blindspot": "The update side often credits generative AI with a fix nobody has measured; the published evidence is for models trained on past renewals, and it predates the current wave. The defend side points to a vendor's cohorts where the right inputs separate retention, which shows a well-built score can work, not that the score you configured does.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A composite of usage, survey scores, support load and CSM sentiment, weighted by judgment, separates future churners from renewers months ahead.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The inputs you can count, such as logins and survey scores, are the inputs that matter.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A team has enough past churn events to set the weights.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Models trained on past renewal outcomes predict which B2B accounts will churn, well enough to target saves.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Generative AI that reads calls and email makes the health score accurate.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Teams will backtest a score instead of trusting its colour.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "Models trained on past renewal outcomes predict which B2B accounts will churn, well enough to target saves.",
    "deciding_side": "challenge",
    "what_follows": "The hand-weighted score was never validated, and its assumption that countable inputs are the ones that matter fails on the evidence available. A model trained on past renewals has been shown to predict churn and to make saves pay, so the update follows: score from outcomes, not judgment. It does not follow that generative AI is the fix; that premise is unproven, and resting the update on it would make this too early to call."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The health score has no author. It grew with the customer-success platforms of the 2010s, each of which ships a red, yellow and green scorecard: usage, survey scores, support tickets and the CSM's own read, weighted however the customer chooses. Gainsight describes the score as one input to renewal, \"determined by you.\" The rule it carried was simple. Watch the colour, and the colour tells you who will leave.\n\nIt was built for scarce attention. A CSM's book ran to dozens or hundreds of accounts, and the data sat in several tools, so one number let a manager rank the book without opening every account."
    },
    {
     "heading": "What people say now",
     "text": "Darlene Kelly, who directs customer-success research at TSIA, says \"traditional customer health scores are failing to predict churn\" and has launched research to measure it, with no data published yet. Gartner analysts, in a client note whose public abstract shows no data, recommend adding generative AI to health scores \"to improve account churn and growth candidate identification.\" Practitioners are more careful. Campbell Robertson, who sells CS and AI advisory work, puts AI summarization in the \"works today\" column and \"predictive churn scores\" among the things that are \"still aspirational\" and \"often mislead more than they help.\"\n\nGainsight, which sells customer-success software, defends the idea with its own cohorts. Accounts with a verified outcome, a customer's stated return, three months before renewal had a 30% higher gross retention rate, and low adoption with a verified outcome beat high adoption without one."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rested on three ideas. A judgment-weighted composite separates churners from renewers months ahead. The inputs you can count are the ones that matter. And a team has enough past churn to set the weights. The first was never tested in public, in either direction. The second fails on the evidence there is: Gainsight's own data makes usage necessary but not sufficient, and a study of 3,959 software subscriptions found usage improves churn prediction only once it is structured by timing and granularity, which is weighting a model learns, not one a person sets.\n\nThe challenge's strongest premise holds. In a study of about 5,000 wholesale B2B customers, a model trained on past outcomes ranked churn risk well, and in a randomized field test, calling the customers it flagged cut churn against random targeting and lifted revenue. So the update follows: score from outcomes, not judgment. What does not follow is that generative AI is the fix. No outcome data yet shows that reading calls and email makes the score more accurate."
    },
    {
     "heading": "What both sides miss",
     "text": "Critics say \"health scores are dead\" when the evidence says something narrower. The configured colour was never validated, and a backtested model was. Defenders cite cohorts where the right inputs separate retention, which defends a well-built score, not the one on the dashboard.\n\nBoth skip the size problem. B2B customer counts are typically small, which makes a model harder to train. A team with few churn events may not have the history to backtest anything, and the honest score for that team says so."
    }
   ],
   "open_question": "Has anyone backtested the health score you report to the board against the renewals it was supposed to predict?",
   "sources": [
    {
     "side": "update",
     "who": "Darlene Kelly, TSIA (analyst position; research under way, no data yet)",
     "date": "2025-10-20",
     "url": "https://www.tsia.com/blog/metrics-to-machine-learning-reinventing-customer-health-models",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Gattermann-Itschert and Thonemann, Industrial Marketing Management (about 5,000 wholesale B2B customers, field experiment)",
     "date": "2022-10",
     "url": "https://doi.org/10.1016/j.indmarman.2022.09.023",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Sanchez Ramirez, Coussement, De Caigny, Benoit and Guliyev, Industrial Marketing Management (3,959 software subscriptions)",
     "date": "2024-06",
     "url": "https://doi.org/10.1016/j.indmarman.2024.05.008",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Gainsight (sells customer-success software), Helminski and McNally, on its own renewal cohorts",
     "date": "2023-12-06",
     "url": "https://communities.gainsight.com/for-csms-by-csms-190/success-for-all-verified-outcomes-and-success-plans-18268",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Schmidt, Maziarka and Marino, Gartner (abstract only)",
     "date": "2025-04-24",
     "url": "https://www.gartner.com/en/documents/6388643",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Campbell Robertson (CS consultant, sells CS and AI advisory)",
     "date": "2025-07-21",
     "url": "https://campbellrobertson.substack.com/p/the-ai-powered-customer-success-revolution",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "customer success",
    "churn",
    "health score"
   ],
   "bench_notes": []
  },
  {
   "slug": "nps-predicts-growth",
   "title": "NPS is the one number that predicts growth",
   "summary": "Update it: peer-reviewed tests found NPS no better than plain satisfaction at predicting retention or growth well before AI arrived, and AI changes how you would measure NPS, not whether it predicts growth.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "evidence-against",
   "because_label": "evidence against it",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "customer and renewal",
   "altitude": "L2 Operating model",
   "origin": "Frederick F. Reichheld, \"The One Number You Need to Grow\", Harvard Business Review, December 2003: \"The best predictor of top-line growth can usually be captured in a single survey question: Would you recommend this company to a friend?\" The share of promoters, he wrote, correlated with growth rates among competitors in most industries studied.",
   "rationed": "Survey cost and executive attention. Satisfaction surveys were long, costly and hard to act on, and leaders wanted one number frontline teams could move that tied to growth.",
   "challenge_claim": "NPS is fading as a measure, and AI that reads customer interactions can infer sentiment without a survey.",
   "blindspot": "The update side treats AI as the reason to drop NPS, when the case against its growth claim was made between 2007 and 2015 on survey data. The defend side points to a conditional result, NPS changes on a broad brand sample predicting short-term sales, which is not the one-number claim most teams run on.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Net Promoter Score predicts a firm's future growth better than other customer metrics.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Survey answers are an honest read of customer loyalty.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "A survey is the only way to read sentiment across the customer base.",
      "now": "changed",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI-inferred sentiment from customer interactions predicts retention and growth better than surveyed NPS.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Organizations are abandoning NPS as a measure.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "Net Promoter Score predicts a firm's future growth better than other customer metrics.",
    "deciding_side": "prior",
    "what_follows": "The premise that decides is the growth claim, and it failed on outcome tests before AI: NPS predicts about as well as satisfaction, not better, and only under conditions. AI changed the measurement premise, since promoter status can now be inferred without a survey, but predicting the NPS answer is not predicting growth. The update follows from the old evidence, not from AI."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "In December 2003 Fred Reichheld of Bain, the consultancy that co-owns the Net Promoter trademark and sells NPS work, published \"The One Number You Need to Grow\" in Harvard Business Review. One survey question, would you recommend this company to a friend, was \"the best predictor of top-line growth,\" and the share of promoters tracked growth among competitors in most industries studied. It answered a real problem. Satisfaction surveys were long and hard to act on, and leaders wanted one number the front line could move."
    },
    {
     "heading": "What people say now",
     "text": "The originator has moved. In \"Net Promoter 3.0\" (2021), Reichheld and his Bain co-authors concede that self-reported scores have been gamed, with \"pleading,\" \"bribery\" and \"manipulation,\" and introduce Earned Growth Rate, built on audited revenue and \"far more resistant to gaming.\" The growth link moved off the survey.\n\nBain has also shown the survey can be skipped. Its Predictive NPS work inferred promoter and detractor status from transactions, call metrics and transcripts, with over 80% accuracy and four times the lift on detractors in one financial-services build. Gartner predicted in 2021 that by 2025 more than three quarters of organizations would abandon NPS as a measure of success for customer service and support. That prediction is limited to customer service, and it rested on a survey of 42 heads of customer service."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on three ideas. NPS predicts growth better than other customer metrics. Survey answers are an honest read of loyalty. And a survey is the only way to read sentiment across the base.\n\nThe first was tested and failed long before AI. Keiningham and colleagues (*Journal of Marketing*, 2007), on 21 firms and over 15,500 interviews, failed to replicate NPS's \"clear superiority.\" Keiningham was at Ipsos Loyalty, a competing measurement firm, but de Haan, Verhoef and Wiesel (2015) reached a similar answer on 93 firms. Top-two-box satisfaction predicted retention best, NPS nearly tied, the best metric varied by industry, and combining metrics beat any one. Jeff Sauro, re-running Reichheld's own data against future rather than past growth, found NPS explained on average 38% of growth variance at two years and 30% at four. Some signal, not the one number.\n\nAI changed the other two premises. Gaming broke the honest-answer premise, and inference broke the survey-only one. Neither touches the growth claim. A model that predicts the NPS answer has predicted the answer, not growth, and no outcome evidence shows AI-inferred sentiment predicting growth better."
    },
    {
     "heading": "What both sides miss",
     "text": "The strongest defence is narrower than the prior. Baehre and colleagues (2022) found NPS predicted short-term sales growth only when measured as changes, on a sample of all potential customers. NPS levels among current customers, the number most teams track, did not predict growth. Critics who say \"AI killed NPS\" miss that the case was already made. Defenders who cite the 2022 result are defending a brand-health tracker, not the score on the CS dashboard."
    }
   ],
   "open_question": "If a model can now infer every customer's NPS answer, what was the answer ever telling you about growth?",
   "sources": [
    {
     "side": "origin",
     "who": "Frederick F. Reichheld, Bain (co-owns the Net Promoter trademark, sells NPS consulting), Harvard Business Review",
     "date": "2003-12",
     "url": "https://hbr.org/2003/12/the-one-number-you-need-to-grow",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Keiningham, Cooil, Andreassen and Aksoy, Journal of Marketing (21 firms, 15,500+ interviews; Keiningham was at Ipsos Loyalty, a competing loyalty-measurement firm)",
     "date": "2007-07",
     "url": "https://journals.sagepub.com/doi/10.1509/jmkg.71.3.039",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "de Haan, Verhoef and Wiesel, International Journal of Research in Marketing (93 firms, 18 industries)",
     "date": "2015-06",
     "url": "https://doi.org/10.1016/j.ijresmar.2015.02.004",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Baehre, O'Dwyer, O'Malley and Lee, Journal of the Academy of Marketing Science (7 US sportswear brands, 5 years)",
     "date": "2022-01",
     "url": "https://link.springer.com/article/10.1007/s11747-021-00790-2",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Reichheld, Darnell and Burns, Bain (sells NPS consulting), Net Promoter 3.0 (HBR, reposted by Bain)",
     "date": "2021-10-18",
     "url": "https://www.bain.com/insights/net-promoter-3-0/",
     "standing": "authority"
    },
    {
     "side": "nuance",
     "who": "Ilker Carikcioglu, Bain (sells NPS consulting), Predictive NPS (one client model)",
     "date": "2021-01-28",
     "url": "https://www.bain.com/insights/sidestepping-survey-fatigue-predictive-nps-gets-to-the-heart-of-customers-attitudes/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "nps",
    "customer success",
    "metrics"
   ],
   "bench_notes": []
  },
  {
   "slug": "nrr-120-is-best",
   "title": "Aim for 120% net revenue retention, with gross retention around 90%",
   "summary": "Update it: NRR still goes with growth, but usage ramps and committed credits can lift it without durable demand, and 120% is now a top-decile result, so read committed and usage expansion apart, with gross retention beside them.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "customer and renewal",
   "altitude": "L2 Operating model",
   "origin": "Bessemer Venture Partners, State of the Cloud 2023, \"Good, Better, Best\" benchmarks for Series B and C enterprise software: net revenue retention of 100% good, 110% better, 120%+ best. These were fundability targets, not measured medians; Bessemer's 2019 version set bands by segment instead. The gross retention floor comes from SaaS Capital, a lender, 2025: GRR \"must be at least 90%\" to keep pace with peers.",
   "rationed": "Attention and capital in a seat-and-license world. Expansion was a discrete, contracted event, so one number could stand for the installed base's demand, and investors could price a base that grew by itself.",
   "challenge_claim": "Usage and AI pricing let NRR rise without durable demand, through ramps and committed credits, so one NRR figure, and the 120% bar, no longer mean what they did.",
   "blindspot": "Defenders lean on cross-sectional correlations, partly mechanical, whose link to valuation swings with the market. Critics have the mechanism but no study that follows usage-inflated NRR forward to growth or churn. Both treat gross retention's 90% as a growth signal, when it barely correlates with growth.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Higher NRR produces faster growth and a higher valuation.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "An expansion counted in NRR reflects durable customer demand, because seat and license expansions are contractual and spread evenly across the base.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "120% is an attainable bar that separates the best.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "Gross retention of about 90% is the floor for B2B software.",
      "now": "holds",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Usage and AI pricing let NRR rise without durable demand, through ramps, committed credits and breakage.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "NRR no longer predicts growth or value in the AI era.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "An expansion counted in NRR reflects durable customer demand, because seat and license expansions are contractual and spread evenly across the base.",
    "deciding_side": "prior",
    "what_follows": "What changed is what a point of NRR means: under usage pricing and committed credits it can come from a ramp rather than a contracted expansion, and that mechanism holds. The claim that NRR stopped predicting growth is unproven, and the challenge doesn't need it. So NRR stays, read with committed and usage expansion apart, and 120% goes as a universal bar because it is now a top-decile result."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Bessemer Venture Partners put the number in its State of the Cloud 2023 benchmarks for Series B and C enterprise software: net revenue retention of 100% is good, 110% better, 120% or more best. These were targets for raising money, not measured medians, and Bessemer added that they shift by stage and by segment. Its 2019 version had used segment bands instead, with 80 to 100% for small business, 90 to 120% for mid-market, and above 100% for enterprise. The flat 120% came later.\n\nThe gross retention floor comes from SaaS Capital, which lends to SaaS companies and says that to keep pace with peers, it \"must be at least 90%.\""
    },
    {
     "heading": "What people say now",
     "text": "The defense is strong on one point, which is that NRR goes with growth. SaaS Capital's survey of more than 1,000 private B2B companies finds \"a strong and exponential correlation\" between the two.\n\nThe challenge is about what a point of NRR now means. L.E.K., a consultancy that advises software companies and their investors, writes that under usage pricing \"NRR is often inflated in the early months of a new customer ramp.\" It suggests tracking contracted and usage revenue apart.\n\nThe bar itself moved. Tomasz Tunguz of Theory Ventures, a VC, found median net dollar retention across 25 public software companies fell from 125% in 2022 to 112% in 2025. The private median sits near 101%."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rested on three ideas. Higher NRR produces growth and value, an expansion reflects durable demand because seat adds are contractual, and 120% is an attainable bar for the best.\n\nThe first still holds, with caveats. Part of the growth link is mechanical, since expansion revenue is part of growth. The valuation link swings with the market. Ray Rike, who sells benchmarks, reports NRR explaining 40% of the variance in public multiples in mid-2021 and 7% in early 2023. The third has moved. 120% was the 2021 to 2022 public norm and is now a top-decile result.\n\nThe second is the one that changed. A seat expansion is a discrete, contracted event; a usage expansion can be a ramp that runs high early, and committed credits can hold it up by contract. The challenge doesn't need NRR to stop predicting growth, and nobody has shown that it has. It needs only that a point of usage NRR is not a point of seat NRR, and that mechanism holds. The price-per-seat entry in this series covers why usage-priced firms post higher NRR; this one asks whether that NRR means the same thing.\n\nWhat follows is a revision, not a retirement. Keep NRR, drop 120% as a universal bar, and read committed expansion apart from usage expansion."
    },
    {
     "heading": "What both sides miss",
     "text": "Gross retention says less about growth than the 90% floor implies. SaaS Capital finds a median of 91%, 95% above $250K ACV, and little correlation between gross retention and growth. The floor guards the base; it doesn't predict the rate.\n\nFor low-priced AI apps the floor isn't close. ChartMogul, a billing vendor, with Kyle Poyar, found AI-native median gross retention of 40% and NRR of 48%, with gross retention at 23% below $50 a month. Poyar calls the AI buckets, about 50 companies each, \"directional.\"\n\nAnd no one has followed usage-inflated NRR forward to see what happened to growth or churn."
    }
   ],
   "open_question": "When a customer's usage ramps, how much of that NRR will still be there in a year?",
   "sources": [
    {
     "side": "origin",
     "who": "Bessemer Venture Partners (VC), State of the Cloud 2023",
     "date": "2023-04-11",
     "url": "https://www.bvp.com/atlas/state-of-the-cloud-2023",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "SaaS Capital (lender), Research Brief 32: Retention, survey of 1,000+ private B2B SaaS companies",
     "date": "2025",
     "url": "https://www.saas-capital.com/wp-content/uploads/2025/09/RB32WS1-2025-B2B-SaaS-Retention-Benchmarks.pdf",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "L.E.K. Consulting (advises software companies and investors), How AI is redefining SaaS metrics and forecasting",
     "date": "2025-07-24",
     "url": "https://www.lek.com/insights/technology/how-ai-redefining-saas-metrics-and-forecasting",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Tomasz Tunguz, Theory Ventures (VC), NDR decline across 25 public software companies",
     "date": "2026-03-07",
     "url": "https://tomtunguz.com/software-ndr-decline-2026/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Kyle Poyar for ChartMogul (billing vendor), SaaS retention: the AI churn wave",
     "date": "2025-12-10",
     "url": "https://chartmogul.com/reports/saas-retention-the-ai-churn-wave/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "retention",
    "expansion",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "retention-five-times-cheaper",
   "title": "Keeping a customer costs five times less than winning one",
   "summary": "Update it: the five-times ratio traces to an \"experts agree\" line, the 25 to 95% profit band was never measured, and loyal customers were not found to be cheaper to serve, so keep the case for retention but drop the numbers.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "never-sourced",
   "because_label": "never sourced",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "customer and renewal",
   "altitude": "L1 Axiom",
   "origin": "Two claims, usually glued together. The five-times ratio traces, per Keiningham and colleagues (Loyalty Myths, 2005), to Technical Assistance Research Project work in the late 1980s, given further credibility by Hart, Heskett and Sasser in HBR in 1990. The profit band comes from Reichheld and Sasser, \"Zero Defections\", HBR 1990, on client examples. Amy Gallo's 2014 HBR article gave it the modern wording: \"increasing customer retention rates by 5% increases profits by 25% to 95%.\"",
   "rationed": "Customer-level profit data. Budgets went overwhelmingly to acquisition, defection went unmeasured, and the numbers were a wake-up argument for funding retention when nobody could compute the economics customer by customer.",
   "challenge_claim": "Retention is the cheapest growth there is: keeping a customer costs a fifth of winning one, and five more points of retention add 25 to 95% to profit.",
   "blindspot": "Critics who stop at \"the numbers are fake\" miss that retention still goes with higher value in B2B software. Defenders who need the numbers miss that the argument does not need them, and that the source HBR links says something much smaller.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Acquiring a customer costs about five times as much as keeping one.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Long-tenured customers cost less to serve and pay more, so retention compounds profit.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Raising retention by 5 points raises profit by 25 to 95%.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "In B2B software, net revenue retention is a dominant driver of company value.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "AI lowers acquisition cost, so retention's cost advantage narrows.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "AI makes switching easier, so churn rises and retention matters more.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Raising retention by 5 points raises profit by 25 to 95%.",
    "deciding_side": "prior",
    "what_follows": "The numbers fail on their own origin: the profit band was never measured and the cost ratio ends in an experts-agree line, and the mechanism under both, that loyal customers are cheaper to serve, failed in a multi-company panel. Retention mattering still holds, on separate evidence. AI did not change any of this, and neither AI-era argument has outcome evidence."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Two numbers travel together. The first, that winning a customer costs five times as much as keeping one, traces through the loyalty researchers Keiningham, Vavra, Aksoy and Wallard to work by the Technical Assistance Research Project in Washington in the late 1980s. An HBR article by Hart, Heskett and Sasser in 1990 lent it credibility, and the trail ends in an \"experts agree\" line with no measurement behind it.\n\nThe second comes from Reichheld and Sasser's \"Zero Defections\" (HBR, 1990): companies could boost profits by almost 100% by keeping 5% more customers, drawn from client examples. Amy Gallo's 2014 HBR article gave it the wording everyone quotes, \"25% to 95%,\" and added that acquisition is \"five to 25 times more expensive… depending on which study you believe,\" naming no study."
    },
    {
     "heading": "What people say now",
     "text": "The source Gallo links is a 2001 Bain brief by Reichheld. It says that \"in financial services, for example, a 5% increase in customer retention produces more than a 25% increase in profit.\" One industry, more than 25%. No 95% appears.\n\nKeiningham and colleagues, whose firm sells loyalty research, called the five-times ratio a myth in 2005: acquisition and retention costs get misallocated, and the answer shifts with product lifecycle and customer mix. Jon Lombardo of LinkedIn's B2B Institute, which has an interest as a seller of acquisition advertising, adds that the \"5%\" was a cut in defection from 10% to 5%, half of all defections, and assumed it cost nothing to achieve."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior needs three things. Acquisition costs about five times retention. Long-tenured customers cost less to serve and pay more, so retention compounds. And five points of retention buy 25 to 95% more profit. The first was never sourced. The third was never measured as stated. The second is the mechanism under both, and it failed on outcome data. Reinartz and Kumar (2002) tracked over 16,000 customers at four companies for four years and found that \"in none of the four companies\" were long-standing customers consistently cheaper to manage. At the B2B service provider, loyal customers cost more to serve.\n\nAI changed none of these premises. The AI-era arguments, that cheaper acquisition narrows retention's edge or that easier switching raises churn, have no outcome evidence either way."
    },
    {
     "heading": "What both sides miss",
     "text": "The case for retention survives without the numbers. McKinsey's study of over 100 B2B software companies found the top quartile by valuation ran 113% net revenue retention against 98% for the bottom quartile, and top-quartile retention companies held higher valuations across cycles. That is correlational and tracks value rather than profit, though it is real. Critics who stop at \"the numbers are fake\" lose that. Defenders who need the numbers to make the case are leaning on a line its own source does not support."
    }
   ],
   "open_question": "If you took the five-times line out of your retention business case, would the case still stand?",
   "sources": [
    {
     "side": "origin",
     "who": "Reichheld and Sasser, Zero Defections, HBR (Bain summary; client examples)",
     "date": "1990-09",
     "url": "https://www.bain.com/insights/zero-defections-quality-comes-to-services-harvard-business-review-hbr/",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "Fred Reichheld, Bain brief, Prescription for cutting costs",
     "date": "2001-10-24",
     "url": "https://media.bain.com/Images/BB_Prescription_cutting_costs.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Reinartz and Kumar, HBR (16,000+ customers, four companies, four years)",
     "date": "2002-07",
     "url": "https://foster.uw.edu/wp-content/uploads/2016/07/11_Reinartz_Kumar_2002.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Keiningham, Vavra, Aksoy and Wallard, Loyalty Myths excerpt (authors' firm, Ipsos, sells loyalty research)",
     "date": "2005",
     "url": "https://www.ipsos.com/sites/default/files/publication/2003-08/Ipsos_Loyalty_Myth_8_Excerpt.pdf",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Pulido, Erdogan and Berliner, McKinsey (100+ B2B SaaS companies)",
     "date": "2025-11-19",
     "url": "https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-net-revenue-retention-advantage-driving-success-in-b2b-tech",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "retention",
    "unit economics",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "csm-per-two-million-arr",
   "title": "Hire one CSM for every $2M of ARR",
   "summary": "Too early to call: AI is taking admin out of the CSM job and leaders want bigger books, but nobody has measured whether retention holds when books grow, and the $2M ratio itself was never shown to protect retention in the first place.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "mostly-aspiration",
   "because_label": "the case to update is mostly aspiration",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "customer and renewal",
   "altitude": "L2 Operating model",
   "origin": "Jason Lemkin, SaaStr, \"Hire One Customer Success Manager for every $2m in ARR\", dating to about 2013 (the post is now marked updated 2024). He called it \"where most of us end up at, at scale\": a description of practice, not a tested optimum.",
   "rationed": "CSM hours. Proactive coverage needed a person per relationship, and the ratio turned ARR into headcount so the CS budget could be planned like quota capacity.",
   "challenge_claim": "AI takes the admin out of the job, so a CSM can carry far more ARR, and coverage should be designed as a system of roles rather than a ratio.",
   "blindspot": "The update side treats bigger books as proof that bigger books work; the data shows the books grew, not what happened to retention. The defend side treats $2M as a retention threshold, when it began as a description of where companies landed and one survey finds no link between the ratio and retention.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Holding coverage near $2M of ARR per CSM protects retention, so stretching the book loses revenue.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "A CSM's capacity is fixed by manual work: prep, data pulls and notes.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "One number, about $2M, fits every segment.",
      "now": "fails",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI lets a CSM cover more ARR without hurting retention.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Coverage designed as a system of roles retains better than coverage set by a ratio.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Holding coverage near $2M of ARR per CSM protects retention, so stretching the book loses revenue.",
    "deciding_side": "prior",
    "what_follows": "What changed is the work a CSM spends hours on, and the books have already grown. Neither side's causal premise has outcome evidence: nobody has shown the ratio protects retention, and nobody has shown the stretched books kept it. Until one of those is measured, the honest call is that it is too early, not that the ratio is safe or safe to drop."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Jason Lemkin of SaaStr, who runs its events and media business and invests in SaaS companies, wrote the rule around 2013: hire one customer success manager for every $2M of ARR. He framed it as \"where most of us end up at, at scale,\" an observation of what companies did, not a tested optimum. The KeyBanc SaaS survey he cites put the overall figure near $1.7M, and by segment about $1.3M for small business, $1.5M mid-market and $2.6M enterprise.\n\nThe ratio rationed CSM hours. Proactive coverage needed a person on each relationship, and a ratio turned ARR into headcount the way capacity planning turns quota into reps."
    },
    {
     "heading": "What people say now",
     "text": "The pressure is to stretch the book. ChurnZero, which sells customer-success software, finds call summarization is the top AI use in CS, and 83% of leaders name reducing CSM admin or growing the managed book as a top scaling priority.\n\nJeff Breunsbach, a CS practitioner, goes further, arguing that the ratio \"anchors you to a linear growth model\" and coverage should be designed as a system of roles. That is an argument with no result behind it."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior needs three things. Coverage near $2M per CSM protects retention. A CSM's capacity is fixed by manual work. And one number fits every segment. The third already fails on the rule's own sources, which give a range from about $1.3M to $2.6M by segment. The second is the one AI changed, as summarization and data pulls take less time and leaders plan to spend that time on bigger books.\n\nThe first was never measured, and it decides the question. A survey of 132 tech companies by CJ Gustafson found \"no meaningful correlation between staffing ratio and revenue retention or renewal rate,\" which is self-reported and does not settle it. ChurnZero finds companies with CSMs report higher net retention than those without, 98% against 90%, which says having CSMs matters while leaving the ratio open. The challenge's own causal premise is equally open, because nobody has published retention for books that grew."
    },
    {
     "heading": "What both sides miss",
     "text": "What has not been measured is the whole point. No published study links ARR per CSM to gross or net retention, and nobody has published whether bigger books kept their customers. Defenders hold a number that was never tied to retention. Critics read a bigger book as success before anyone has checked what it did to retention."
    }
   ],
   "open_question": "If your CSMs now carry bigger books, did retention hold, and would anyone know if it hadn't?",
   "sources": [
    {
     "side": "origin",
     "who": "Jason Lemkin, SaaStr (runs SaaS events and media, invests in SaaS companies; cites the KeyBanc SaaS survey)",
     "date": "2024-02-02",
     "url": "https://www.saastr.com/the-2-million-dollar-man-woman-how-to-think-about-scaling-your-customer-success-team/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "CJ Gustafson, Mostly Metrics (survey of 132 tech companies, self-reported)",
     "date": "2026-01-25",
     "url": "https://www.mostlymetrics.com/p/the-customer-success-benchmarks-you-ve-been-waiting-for",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Jeff Breunsbach, CCO newsletter (CS practitioner)",
     "date": "2025-10-05",
     "url": "https://cco.beehiiv.com/p/your-csm-customer-ratio-is-a-trap",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "ChurnZero (sells customer-success software), 2025 CS Revenue Leadership Study, n=793",
     "date": "2025",
     "url": "https://info.churnzero.com/hubfs/2025%20CS%20Revenue%20Leadership%20Study.pdf",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "customer success",
    "capacity",
    "benchmarks"
   ],
   "bench_notes": []
  },
  {
   "slug": "quarterly-business-reviews",
   "title": "Run a business review with every customer every quarter",
   "summary": "Too early to call: AI takes away the data-assembly job that filled most reviews, but nobody has shown that the quarterly cadence itself, as distinct from a value conversation at the right moment, raises renewal.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "customer and renewal",
   "altitude": "L3 Practice",
   "origin": "No traceable origin. The earliest dated canonical treatment is Gainsight's in 2014: Lincoln Murphy called the QBR \"one of the most important activities your CSMs will perform\", and Dan Steinman asked whether it was \"Value or Fluff?\" and proposed renaming it the executive business review. Murphy also reported the inside joke that QBRs rarely happen quarterly.",
   "rationed": "Visibility and executive attention. Customers couldn't easily see their own usage and return, executives were hard to get in a room, and the CSM needed a forcing function to get that time on a schedule before renewal.",
   "challenge_claim": "Customers can now ask AI for their own usage and trends, so a quarterly ritual of narrating data adds nothing and should go.",
   "blindspot": "The update side pins the decline on AI, when the sharpest critiques of the format, that it was backward-looking, vendor-centric and costly to prepare, came years before it. The defend side cites a renewal lift from having a review at all, which says nothing about holding one every quarter.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Customers can't see their own usage and value data, so the vendor must package it for them.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "A fixed quarterly cadence of reviews raises renewal and expansion.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Buyers value the review and show up for it.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI removes the data-assembly work, so the quarterly ritual adds no value.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "The format was failing for reasons that predate AI: backward-looking, vendor-centric and costly to prepare.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A fixed quarterly cadence of reviews raises renewal and expansion.",
    "deciding_side": "prior",
    "what_follows": "The premise that changed is the data job: customers can see more of their own usage without a vendor deck. That removes one reason for the meeting, not the case for or against it, because whether a quarterly cadence raises renewal was never tested. The challenge that holds is older than AI, so AI is not what changed the verdict, and the call stays open."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody owns the QBR. The earliest dated treatment that reads as canonical is Gainsight's, in 2014. Lincoln Murphy called it \"one of the most important activities your CSMs will perform.\" Dan Steinman asked whether it was \"Value or Fluff?\" and suggested calling it an executive business review, to get the sponsor in the room. Murphy also passed on the inside joke that QBRs rarely happen quarterly, so the cadence was aspirational from the start.\n\nThe format rationed two things. Customers couldn't easily see their own usage and return, and executives were hard to get in a room. A standing quarterly slot fixed both before renewal came due."
    },
    {
     "heading": "What people say now",
     "text": "Chloe Siegmann, a CS practitioner, argues customers can now ask AI for their own usage trends on demand, so the data-narration part of the meeting is gone. Jason Lemkin, who runs SaaStr's events and media business and invests in SaaS companies, writes as a buyer and calls QBRs \"simply upsell exercises\" and says he had already read the data being shown.\n\nThe defence is vendor data. A 2023 buyer survey commissioned by Clientshare, a vendor, found 68% of 200 buyers at large firms called regular reviews the most important moment for a supplier to show value. That is stated preference. A case study from Matik, also a vendor, reports that its customer's accounts that got a business review renewed 32.8% higher than those that didn't."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on three ideas. Customers can't see their own data, so the vendor packages it. A quarterly cadence raises renewal. And buyers value the meeting enough to attend. Practitioners report that AI changed the first because the data job shrinks when customers can query their own usage. Buyer value is split, with a commissioned survey on one side and buyers like Lemkin on the other.\n\nThe cadence premise is the one that decides, and it has never been tested. The lone cohort figure measures having any review at all rather than a quarterly one, and selection skews who gets reviewed because teams may pick accounts already likely to renew. The challenge has the same gap. \"AI made the ritual pointless\" remains an unverified argument. Its stronger premise, that the format was failing anyway, holds, but it predates AI, so AI is not what moved the verdict."
    },
    {
     "heading": "What both sides miss",
     "text": "Both argue about the meeting when the open question is the trigger. Donna Weber, who sells onboarding and CS consulting, proposes holding a value review whenever there is value to show instead of following the calendar. No test yet separates a quarterly cadence from a value conversation at the right moment, and that is the comparison that would settle this."
    }
   ],
   "open_question": "If the quarterly review stopped tomorrow, which renewals would you lose, and could you name them?",
   "sources": [
    {
     "side": "origin",
     "who": "Lincoln Murphy, Gainsight (sells customer-success software)",
     "date": "2014-09-23",
     "url": "https://www.gainsight.com/blog/quarterly-business-review-qbr/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Jason Lemkin, SaaStr (runs SaaS events and media, invests in SaaS companies; writing as a buyer)",
     "date": "2024-11-21",
     "url": "https://www.saastr.com/the-death-of-the-qbr/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Chloe Siegmann (CS practitioner)",
     "date": "2026-08-13",
     "url": "https://www.linkedin.com/posts/chloe-siegmann_customersuccess-accountmanagement-ai-activity-7493765001007017984-8tht",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Donna Weber (onboarding and CS consultant, sells that advisory work)",
     "date": "2026-07-02",
     "url": "https://donnaweber.com/the-qbr-is-dead-long-live-the-value-review/",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Clientshare (vendor) and Insight Avenue, commissioned buyer survey, n=200",
     "date": "2023-11",
     "url": "https://www.myclientshare.com/hubfs/whitepaper/QBR-research.pdf",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Matik (vendor), case study of its own customer",
     "date": "2025-05-27",
     "url": "https://www.matik.ai/blog/how-bazaarvoice-used-matik-to-increase-renewal-rates",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "customer success",
    "renewal",
    "cadence"
   ],
   "bench_notes": []
  },
  {
   "slug": "crm-is-the-record",
   "title": "The CRM is the system of record: if it's not in Salesforce, it didn't happen",
   "summary": "Update it: the CRM is losing its job as the interface but not as the record, and the record never held the buying committee, so treat it as the record of amount, stage and owner, not of the deal.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L2 Operating model",
   "origin": "No originator. The mantra shows up in sales-ops writing by 2016, already on memes, T-shirts and mugs, and SaaStr reported it in 2023 as the golden rule of Salesforce's internal forecast process. The architecture is older: marketing-automation syncs are built with the CRM as the winner for accounts and opportunities.",
   "rationed": "Capture cost. Activity existed only if a rep typed it in, so the CRM was the one place a manager could see a deal.",
   "challenge_claim": "Capture has moved to machines and the work has moved into chat and email, so the CRM becomes one input to the record rather than the record itself.",
   "blindspot": "Updaters attack the interface and the typing, then concede the CRM still holds the data; a16z's own GTM survey found CRM usage rising since AI tools arrived. Defenders miss that the record never held the buying committee or most meetings, so \"didn't happen\" was always false for the parts of a complex deal that decide it.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Activity exists only if a rep types it in, so the CRM is the one place a manager can see a deal.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "The CRM is where the work happens, so it is the natural home for the record.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "What gets logged covers what matters about the deal.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Machines now capture calls and email, so evidence of the work exists outside the CRM.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "The work has moved into chat and email, away from the CRM's interface.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "When the interface and the typing move, the record moves with them.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "What gets logged covers what matters about the deal.",
    "deciding_side": "prior",
    "what_follows": "Capture and the work both moved, which ends the CRM's job as the interface. The challenge overreaches when it concludes the record moved too: usage is rising, and both sides treat the CRM as where the data lives. The premise that really fails is older, that the record covered what matters, so the update is narrower: the CRM is the record of amount, stage and owner, not of the deal."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody wrote it down first. By 2016 one sales-ops team called \"if it's not in there, it didn't happen\" its mantra, in a piece noting the line was already on memes, T-shirts and mugs. By 2023 SaaStr was reporting it as the golden rule of Salesforce's own forecast process. Under the slogan sits an architecture: syncs that let the CRM win for accounts and opportunities, and a weekly forecast call that only counts what's logged."
    },
    {
     "heading": "What people say now",
     "text": "The update side says the CRM is losing its job. a16z, which invests in the layer above the CRM, says it is \"turning into just an input; one of many inputs, into the systems of intelligence which we use to get work done.\" Dominik Dudy, a revenue leader across EMEA SaaS, adds a regional point: \"The German AE who tells the buyer 'let me log this in Salesforce' is the AE who never gets the next call.\"\n\nThe defenders say agents raise the CRM's value. Pratik Mehta, a fractional RevOps consultant who builds HubSpot systems: \"When agents start doing the work, the CRM does not become less important. It becomes the brain.\" Benn Stancil, from the data side: \"it seems like we'll always have systems of record that are full of important information.\"\n\nThe awkward fact sits in the same a16z piece. Its own GTM survey found that \"CRM usage has actually risen\" since AI tools arrived, because agents writing call notes back give reps fresh reason to open it."
    },
    {
     "heading": "Does the logic hold",
     "text": "The mantra rested on three ideas. Activity existed only if a rep typed it in. The CRM was where the work happened, so it was the natural home for the record. And what got logged covered what mattered about the deal. The first two have changed. Machines capture calls and email, and the work has moved into chat and email.\n\nThe update side is right about those changes and goes too far with them. Its conclusion, that the CRM becomes one input among many, needs the record to leave along with the interface and the typing. The usage data says it hasn't, and both sides still treat the CRM as where the data lives. The premise that fails is the third, and it failed long before AI, because the record never held the buying committee or most meetings. So the update that follows is narrower. The CRM stays the record of amount, stage and owner, and stops being read as the record of the deal."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides underrate how much of the deal was never in the record. In my own CRM, almost no opportunities carried contact roles, and the few that did mostly left the role itself blank. When I needed \"who was on this deal\", I had to rebuild it from call participation. That's harder than it sounds. An invite list isn't an attendee list, and a meeting stored as scheduled may have happened or not.\n\nWhat changed is who reads the gap. A manager knew the committee wasn't in the CRM. An agent reading the opportunity doesn't, and a wrong value in the record now reaches every agent that reads it."
    }
   ],
   "open_question": "Should the record move to where the work happens, or should the evidence of the work be written back to the CRM?",
   "sources": [
    {
     "side": "origin",
     "who": "Mark Emmons, Medium (quoting Alex Miller, Zinc)",
     "date": "2016-06-23",
     "url": "https://mark-62118.medium.com/if-its-not-in-salesforce-it-didn-t-happen-21595dde9379",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Dominik Dudy, revenue leader across EMEA SaaS (no product)",
     "date": "2026-06-12",
     "url": "https://www.linkedin.com/posts/dominik-dudy_if-its-not-in-salesforce-it-didnt-happen-activity-7471098216302747648-XXd6",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Pratik Mehta, fractional RevOps consultant (builds HubSpot systems for clients)",
     "date": "2026-04-22",
     "url": "https://www.linkedin.com/posts/pratik-mehta1_revops-hubspot-buildinpublic-activity-7452551507964678144-sXcH",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Benn Stancil, \"The context layer\"",
     "date": "2025-08-29",
     "url": "https://benn.substack.com/p/the-context-layer",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Ahern, Zhang and Immerman, a16z (venture investor in AI software; its own GTM survey)",
     "date": "2026-05-14",
     "url": "https://www.a16z.news/p/from-system-of-record-to-system-of",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "crm",
    "system of record",
    "buying committee"
   ],
   "bench_notes": [
    {
     "slug": "invited-attended-mentioned",
     "title": "Being invited or mentioned does not mean someone attended a meeting"
    },
    {
     "slug": "meeting-happened-is-evidence",
     "title": "Whether a meeting happened depends on which source saw it"
    }
   ]
  },
  {
   "slug": "data-decays-30-percent",
   "title": "B2B data decays 22.5 to 30% a year",
   "summary": "Update it: no study stands behind the 22.5 to 30% figure, decay varies widely by role and industry, and corrections can do more harm than the decay, so measure your own rate instead of quoting one.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "never-sourced",
   "because_label": "never sourced",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L1 Axiom",
   "origin": "No traceable study. HubSpot's Database Decay Simulation page footnotes MarketingSherpa for 2.1% a month, which annualizes to 22.5%. No MarketingSherpa study, year, sample or method is linked, and a 2026 search of MarketingSherpa's site didn't find one. The page uses decay to sell inbound lead flow. The 30% version is a later rounding. The 70% figure is older and measured something else: a self-reported business-card survey from 2002, updated in 2009.",
   "rationed": "Verification cost. Re-checking a contact took a person's time, so one blended refresh rate was the only planning number anyone could afford.",
   "challenge_claim": "The number is roughly right for job changes, but over two years, not one. Lusha measured about 12% over 12 months and about 26% over 24 months in its own database.",
   "blindspot": "Debunkers replace one average with another, though federal turnover data shows people leaving jobs more than twice as fast in some industries as in others. Both sides treat decay as the world changing. In my builds, a large share of wrong records came from the correction machinery: a vendor field misread, or a single source trusted.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Someone measured how fast B2B contact data goes stale.",
      "now": "fails",
      "kind": "prevalence"
     },
     {
      "premise": "One blended annual rate is close enough to plan refreshes for every role and industry.",
      "now": "fails",
      "kind": "prevalence"
     },
     {
      "premise": "Re-checking a contact takes a person's time, so a single planning number is the affordable choice.",
      "now": "changed",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "One vendor's measurement of its own database stands for the wider market.",
      "now": "unproven",
      "kind": "prevalence"
     },
     {
      "premise": "Job changes are the decay that matters.",
      "now": "fails",
      "kind": "prevalence"
     },
     {
      "premise": "A corrected average, quoted over the right period, can stand in for your own rate.",
      "now": "fails",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "Someone measured how fast B2B contact data goes stale.",
    "deciding_side": "prior",
    "what_follows": "The prior fails on its own terms: no study stands behind it, and one rate can't cover a fourfold spread by industry. The leading correction keeps the same logic with a better-sourced number, so it inherits the flaw and rests on one vendor's cohort. What changed is the cost of checking, and it points to measuring your own rate by role class, false movers included."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "From a footnote. HubSpot's database-decay calculator credits MarketingSherpa with 2.1% a month, which compounds to 22.5% a year. No study, year, sample or method is linked, and a 2026 search of MarketingSherpa's site came up empty. HubSpot used the figure to sell inbound lead flow. Later retellings rounded it to 30%, and a 70% figure circulates too."
    },
    {
     "heading": "What people say now",
     "text": "2026 brought a tracing wave, mostly from data vendors with an interest in the answer. Lusha, which sells contact data, looked for the original and found none. In its own database, about 12% of senior US sales leaders changed roles over 12 months and about 26% over 24. Its verdict: \"The number is roughly right. It is being quoted over the wrong period.\"\n\nThe 70% figure measured something else. It comes from John Coe's business-card survey, first run in 2002 and updated in 2009: about 1,200 people reported what had changed on their card in the past year, including title changes inside the same company. Federal turnover data splits the average another way. In 2025, people left employers in accommodation and food services at an average of 5.5% a month, against 2.1% in finance and insurance. A figure means little without what it measured, its sample and its year."
    },
    {
     "heading": "Does the logic hold",
     "text": "The number rested on three ideas. Someone had measured it. One blended rate was close enough for every role and industry. And re-checking a contact took a person's time, so a single planning number was the affordable choice. The first never held. The second fails against the federal turnover data, where one industry loses people about two and a half times as fast as another. The third is the one that changed, because tools now re-check contacts on their own and make measuring your own rate cheap.\n\nThe leading correction keeps the second idea and swaps the number. Lusha's figure is better sourced than a footnote, but it is one vendor's database and one cohort, senior US sales leaders, and it counts only job changes. Quoting it over two years instead of one is still quoting an average. What follows from the premise that changed is not a better average. It is your own rate, measured by role class."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides treat decay as people moving. In my own builds, some of the worst damage came from the corrections. A vendor field that looked like \"latest employer\" was really the most recent start date, so a new board seat read as a job change. When I re-checked the people it had flagged as movers, nearly half were still at the company.\n\nDecay also runs on different clocks. I once confirmed a senior executive as still in role, and three days later found a replacement announced two months earlier. A company's posture changes over months. One annual rate hides both.\n\nAnd a job change doesn't decay a record so much as split it. The person now has two employments, and a model with one \"current employer\" field has to lose one of them."
    }
   ],
   "open_question": "If agents re-verify every contact at the moment of use, does a decay rate still matter, or does the cost just move?",
   "sources": [
    {
     "side": "origin",
     "who": "John Coe, Biznology (source of the 70% figure: his own business-card survey, 2002, updated 2009)",
     "date": "2015-02-13",
     "url": "https://biznology.com/2015/02/b2b-data-decay-list-rental-buyer-beware/",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Lusha research (sells contact data)",
     "date": "2026-09-12",
     "url": "https://www.lusha.com/blog/b2b-data-decay-rate-measured/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "US Bureau of Labor Statistics, JOLTS Table 20, annual average total separations rates by industry",
     "date": "2026-03-13",
     "url": "https://www.bls.gov/news.release/jolts.t20.htm",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "data quality",
    "enrichment",
    "benchmarks"
   ],
   "bench_notes": [
    {
     "slug": "design-for-decay",
     "title": "Compute a fact's age when it's shown instead of in a refresh job"
    },
    {
     "slug": "job-change-splits-one-person",
     "title": "A job change leaves two records for one person or keeps their old job"
    },
    {
     "slug": "employment-is-a-time-window",
     "title": "People who leave a company lose the link to their past work"
    }
   ]
  },
  {
   "slug": "firmographics-good-enough-to-route",
   "title": "Employee count and revenue are good enough to route and segment on",
   "summary": "Update it: data vendors now concede in print that employee count and revenue are modeled estimates, and a hand check flips many segment verdicts, so the fields can't be trusted near a routing threshold.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "evidence-against",
   "because_label": "evidence against it",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L3 Practice",
   "origin": "No traceable origin. It's encoded in practice rather than stated: segment and territory rules keyed on employee bands and revenue, filled by a data vendor and rarely re-checked.",
   "rationed": "Research cost. Checking a company's real size took a person's time per account, so a vendor's number was the affordable stand-in.",
   "challenge_claim": "Revenue and employee counts in B2B data platforms are modeled estimates, not verified facts. One vendor puts its own headcount ranges at plus or minus 20 to 40%, and headcount can sweep in contractors, resellers and a parent company's staff.",
   "blindspot": "Vendors concede the fields are estimates and sell a better estimate; none publishes how often a segment verdict flips. Practitioners who route on the fields are silent about them. Agents make the gap worse, because they apply a threshold literally where a person would have noticed the account was obviously mid-market.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Vendor employee counts and revenue are close enough to the real figures to sort accounts into segments.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Checking a company's real size takes a person's time per account, so the vendor number is the affordable stand-in.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "A person looking at a routed account notices when it is obviously in the wrong segment.",
      "now": "changed",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Vendor size fields are modeled estimates, not verified facts.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "The estimation error is large enough to change routing decisions near a threshold.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Vendor employee counts and revenue are close enough to the real figures to sort accounts into segments.",
    "deciding_side": "prior",
    "what_follows": "The deciding premise fails: vendors call the fields modeled, and the errors are large and sit near the thresholds. What AI changed is the reviewer, because an agent applies the threshold literally, so the error now reaches routing unchecked. Checking still takes per-account work, which is why the check belongs at the thresholds, not on every account."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody wrote this one down. It lives in configuration: a routing rule that sends accounts above some employee count to enterprise, a territory plan cut by revenue band, both filled by whichever data vendor the company bought. The fields fill reliably, and fill rate is easy to mistake for accuracy."
    },
    {
     "heading": "What people say now",
     "text": "The clearest statement against the prior comes from a data vendor. Apollo, which sells these fields: \"Revenue and employee count fields in B2B data platforms are modeled estimates, not verified facts.\" Firecrawl, which resells several of the providers it tested, scored six of them against filed headcounts for 200 US companies, 150 of them public, in October 2026. A count within 1.2 times the filing earned full credit and one within twice it earned half. The best provider scored 77% on employee count; the rest scored between 46% and 57%. Anders Krohn of Kernel, another data vendor, puts the headcount problem in one line: \"that's like counting every Uber driver as an Uber employee.\"\n\nAll three sell something here. One of them is conceding a weakness in its own product, which makes the concession worth more than the pitches.\n\nPractitioners are quiet. I found almost no RevOps posts defending the fields and almost none attacking them, while routing keeps running on them."
    },
    {
     "heading": "Does the logic hold",
     "text": "Routing on these fields rested on three ideas. The vendor's numbers were close enough to sort accounts into segments. Checking a company's real size took a person's time per account, so the vendor number was the affordable stand-in. And a person looking at a routed account would notice when it was obviously in the wrong segment.\n\nThe first idea fails, and the vendors say so. The fields are modeled, Apollo puts its own headcount ranges at plus or minus 20 to 40%, and headcount sweeps in contractors and a parent company's staff. The errors also sit where routing decides, near the thresholds. AI changed the third idea. An agent applies the threshold literally, with nobody noticing that an obviously mid-market account has landed in enterprise. So the update follows from one premise that failed and one that changed. The second idea still holds because checking takes work, which puts the check at the thresholds instead of on every account."
    },
    {
     "heading": "What both sides miss",
     "text": "Nobody publishes how often a segment verdict changes when you check, which is the number that matters. I checked mine. On a sample of accounts verified by hand on the web, about 4 in 10 segment verdicts flipped against the vendor fields, even though both size fields were nearly always filled.\n\nThe errors had shapes:\n\n- After an acquisition, the count reads the parent, leaving a small company showing tens of thousands of employees.\n- A size-band picklist straddles the routing threshold, so the band can't answer the question.\n- A cluster of \"large\" accounts turned out to be acquired or defunct shells.\n\nWhat worked was dull. Look the company up on its own LinkedIn page or site, keyed on its web domain. Name plus country plus industry fails on common names."
    }
   ],
   "open_question": "Should agents route on a checked company size with the vendor field as a hint, or is a size threshold the wrong tool for routing?",
   "sources": [
    {
     "side": "update",
     "who": "Apollo insights (sells B2B data; conceding its own fields)",
     "date": "2026-05-12",
     "url": "https://www.apollo.io/insights/how-accurate-are-company-revenue-and-employee-counts-in-a-b2b-data-platform",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Firecrawl (resells several of the providers it scored), company-enrichment benchmark against filed headcounts",
     "date": "2026-10-02",
     "url": "https://github.com/firecrawl/benchmark-company-enrichment",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Anders Krohn, founder and CEO, Kernel (sells company data)",
     "date": "2025-08-21",
     "url": "https://www.linkedin.com/posts/anderskrohn_enterprise-revops-teams-know-that-headcount-activity-7364206596505198593-b18L",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "data quality",
    "routing",
    "segmentation"
   ],
   "bench_notes": []
  },
  {
   "slug": "human-approves-every-write",
   "title": "A human approves every write to the system of record",
   "summary": "Update it: approving every agent write recreates the manual work, and research from other fields shows blanket review lets errors through while risk-based review catches as much for less effort, so tier review by consequence; how a field earns its way out is unsettled.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L3 Practice",
   "origin": "No traceable origin. It extends the admin canon of change control (build in a sandbox, approve, then deploy) to agent writes, and it's the default in governance advice: \"At least one human review point in every AI-assisted workflow.\"",
   "rationed": "Write volume. People and a few integrations wrote slowly enough that a reviewer could look at each change.",
   "challenge_claim": "Approving every write recreates the manual work the agent was meant to remove. Review belongs at the write boundary, tiered by consequence: rules graduate out of review one at a time, while owner changes, merges, deletes, stage and opt-out fields always stay in review.",
   "blindspot": "Updaters rarely say how a field graduates out of review; the proposed metric is how often reviewers edit the agent's proposals, and a low edit rate can mean rubber-stamping as easily as readiness. Defenders miss that blanket review decays, which other fields have measured. And an approval is only as good as its source, because an agent can write one.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Writes come slowly enough that a reviewer can look at each one.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "A person who approves a change has checked it.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A wrong write to the system of record costs enough to justify a check.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Approving every agent write recreates the manual work the agent was meant to remove.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Consequence can be judged by field, so low-stakes writes can leave review while high-stakes ones stay.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "How often reviewers edit a field's proposals shows when it is ready to leave review.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A person who approves a change has checked it.",
    "deciding_side": "prior",
    "what_follows": "Write volume changed, and under it the premise that an approval means a check has failed: in other fields, reviewers pass a measurable share of wrong automated output, and review quality falls as volume rises. The challenge's core premises hold, with a randomized analog showing risk-based review catching as much as blanket review at under half the effort, so review by consequence follows. The graduation test rests on an unproven premise, so the update stops at tiers."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Admin canon for change control is to build in a sandbox, get it approved, and then deploy. When agents started writing to the CRM, the rule moved down from configuration to data. One consultant's governance advice for Salesforce admins still states it flatly: \"At least one human review point in every AI-assisted workflow.\""
    },
    {
     "heading": "What people say now",
     "text": "The playbooks are moving away from \"every write\". The emerging line is a tier by consequence. One playbook spells out what never leaves review: \"owner changes, merges, deletes, lifecycle or stage changes, and opt-out fields.\" Other rules graduate out of review one at a time.\n\nThe graduation test is the part nobody has settled. Chris Fitkin of metacto, a vendor, proposes one: \"If humans constantly rewrite the agent's updates, the workflow is not ready for more autonomy.\""
    },
    {
     "heading": "Does the logic hold",
     "text": "The rule rested on three ideas. Writes came slowly enough that someone could look at each one. A person who approved a change had checked it. And a wrong write to the system of record cost enough to justify the check. Agents changed the first by writing at a volume no reviewer can read. That breaks the second, which was already weak: a reviewer facing more than they can read starts passing things through.\n\nNobody has measured this in CRM approval queues, but other fields have measured it. In clinical decision support, a 2026 meta-analysis found that in real-world use, 26.6% of erroneous computerized ECG interpretations were not corrected or recognized by the clinicians reviewing them. In primary care, the more reminders clinicians got in a visit, the fewer they accepted. The likelihood of accepting one fell 30% for each additional reminder per visit. Review gets worse as volume rises, which is exactly what agents bring.\n\nThe third idea still holds, and it is what the tiered model keeps. The challenge needs approving every write to recreate the manual work, which it plainly does, and consequence to be judged by field, so low-stakes writes can leave review. Clinical-trial monitoring tested that second move. In the ADAMON trial, 213 sites were randomised between extensive on-site monitoring and monitoring adapted to risk. Risk-based review was non-inferior on major errors and took less than half the effort. So moving from every write to tiers follows. The last need, that an edit rate shows when a field can leave review, is one vendor's proposal, and the automation-bias research suggests a low edit rate can mean reviewers stopped looking."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides assume an approval comes from a person. An approval is only as good as its source. One of my agents routes requests to background workers that need my sign-off before any send or CRM write. In testing, it wrote an approval line into a worker's instructions on its own, for a message it had drafted itself. To the worker, that line read as my go-ahead. Now only my own words count as approval, and the line saying a review happened is written by the system that ran it, never by the agent.\n\nI don't approve every write any more. Proposals are grouped, and a group applies only when a hand-audited sample from it passes. A run that proposes too many changes at once, or pushes most records the same way, is held for me. A value nobody could verify counts as wrong."
    }
   ],
   "open_question": "Once a field has left review, what pulls it back when its error rate drifts?",
   "sources": [
    {
     "side": "defend",
     "who": "Jeremy Carmona, Salesforce Ben (consultant doing AI governance for Salesforce orgs)",
     "date": "2026-02-04",
     "url": "https://www.salesforceben.com/how-salesforce-admins-can-apply-data-governance-to-einstein/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Chris Fitkin, metacto (builds AI write-back systems)",
     "date": "2026-07-16",
     "url": "https://www.metacto.com/blogs/ai-write-backs-when-an-agent-should-update-crm-erp-or-ticketing-systems",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Zeng, Di, Lei, Li et al., npj Digital Medicine, automation bias meta-analysis (clinical decision support)",
     "date": "2026-10-23",
     "url": "https://www.nature.com/articles/s41746-026-03408-5",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Ancker et al., BMC Medical Informatics and Decision Making, reminder fatigue in primary care",
     "date": "2017",
     "url": "https://ncbi.nlm.nih.gov/pmc/articles/PMC5387195/pdf/12911_2017_Article_430.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Brosteanu et al., Clinical Trials, ADAMON cluster-randomised trial of risk-adapted monitoring",
     "date": "2017",
     "url": "https://doi.org/10.1177/1740774517724165",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "agents",
    "governance",
    "crm"
   ],
   "bench_notes": [
    {
     "slug": "review-floor-fresh-eyes",
     "title": "Every agent result gets one fresh-eyes review"
    }
   ]
  },
  {
   "slug": "reps-enter-the-data",
   "title": "Data entry is the rep's job, enforced in pipeline review",
   "summary": "Update it: machines now capture the activity, so reps should stop typing it, but the reasons behind a deal still live only with the rep, so the judgment fields stay theirs to confirm.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L2 Operating model",
   "origin": "No originator; a management norm. Every opportunity and interaction is logged by the rep and checked in the weekly forecast call. SaaStr described Salesforce's internal version of it in 2023, and a 2016 sales-ops piece already pictures ops \"nagging reps\" while automating what it could.",
   "rationed": "Capture cost. Activity existed only if someone typed it in, and the rep was the only person who saw it happen.",
   "challenge_claim": "Calls, email and meetings are captured automatically now, so reps should stop entering data and AI should create the record instead.",
   "blindspot": "The update side treats captured activity as the whole record, and treats a model's extraction as free of the bias a rep brought. The defend side has almost no voices left; the managers who still enforce the rule don't post about it, and the rule was failing before AI.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Only the rep sees the activity, so only the rep can record it.",
      "now": "changed",
      "kind": "prevalence"
     },
     {
      "premise": "Making reps enter data, and checking it in pipeline review, produces an accurate record.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "The rep holds the judgment behind a deal, why it moved and who matters, and it lives nowhere else.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Machines capture calls, email and meetings without anyone typing.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "What a machine extracts from that activity is accurate enough to stand as the record.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Everything that matters about a deal shows up in the captured activity.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Only the rep sees the activity, so only the rep can record it.",
    "deciding_side": "prior",
    "what_follows": "Capture is the premise that changed: the activity no longer needs a rep to type it, and rep entry was failing anyway. The challenge goes further and treats the extracted record as complete and correct, but the reasons behind a deal were never in the activity, and extraction accuracy is unproven. So the writer of the activity record changes; the judgment fields stay with the rep, who now confirms rather than types."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody wrote it down as a rule; it was how sales was managed. The rep logs every opportunity and interaction, and the weekly forecast call checks it. SaaStr described Salesforce's internal version in 2023. Seven years earlier, a sales-ops writer was already describing ops \"nagging reps\" to log while automating whatever could be automated."
    },
    {
     "heading": "What people say now",
     "text": "Nobody defends the rule anymore, at least not in public. \"CRM data hygiene is a losing battle,\" writes John Hulina, who founded an AI CRM service, and \"AI works best when it creates the data instead of waiting for humans to log it.\" Going further back, a RevOps Co-op panel notes that pipeline reviews \"become performance rituals rather than honest assessment.\" That critique predates AI.\n\nHulina makes the capture vendors' case that the machine can now write the record. Kevin Stout of Revductive, a RevOps consultancy, agrees with automating the transcript and still adds: \"One important point is to still make the fields required,\" so a person checks what the machine filled."
    },
    {
     "heading": "Does the logic hold",
     "text": "The rule rested on three ideas. Only the rep saw the activity, so only the rep could record it. Making reps record it, and checking in pipeline review, produced an accurate record. And the rep held the judgment behind a deal, knowing why it moved and who mattered.\n\nAI changed the first idea completely. The second had already failed. The third still holds, and it is the one the challenge needs to fail. Its conclusion, that AI should create the record, assumes both that extraction is accurate and that the activity contains what matters. The first is unproven. Ahoy, which sells an AI-native CRM, works the arithmetic: 95% extraction accuracy across 200 deals and six fields leaves 60 wrong values, and nothing marks which ones. The second fails, because a reason is not an activity. What follows is a change of writer for the activity, not for the judgment."
    },
    {
     "heading": "What both sides miss",
     "text": "Taking the rep out of the record doesn't take the bias out. A model reading a call has its own errors, and nobody owns them in pipeline review.\n\nThe deeper gap is what never reaches any system. In my own work, research disproved the written reason for an exclusion decision. The real reason, an existing relationship, was in no field, no note and no call that any research pass could reach. Verification can kill a stated reason. It can't prove there isn't another."
    }
   ],
   "open_question": "If no rep types the record, who answers for it in pipeline review?",
   "sources": [
    {
     "side": "update",
     "who": "John Hulina, founder, AI CRM Assist (sells AI CRM services)",
     "date": "2026-01-02",
     "url": "https://www.linkedin.com/posts/jhulina_revops-salesops-crm-activity-7412983441924927488-U08V",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Kevin Stout, Revductive (RevOps consultancy)",
     "date": "2026-09-10",
     "url": "https://revductive.com/blog/stop-adding-required-fields-automate-the-transcript",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Kevin Bluett, co-founder, Ahoy (sells an AI-native CRM)",
     "date": "2026-09-07",
     "url": "https://ahoy.ai/guides/trust-ai-written-crm-records/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "crm",
    "data capture",
    "pipeline review"
   ],
   "bench_notes": []
  },
  {
   "slug": "required-fields-enforce-quality",
   "title": "Required fields and validation rules enforce data quality",
   "summary": "Update it: a filled field was never a true field, and AI fills repeat that failure with more plausible values, so require a field only at the stage gate where a named report or routing rule reads it.",
   "verdict": "update",
   "verdict_label": "Update it",
   "because": "shifting",
   "because_label": "the mechanism has shifted",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L3 Practice",
   "origin": "A platform feature that became a practice: required fields and validation rules on save. Marketo's Salesforce sync respects Salesforce validation rules and logs a conflict when one rejects an update. Salesforce's own help pages document gaps, such as duplicate rules that don't run on updates from Workflow Rules or Process Builder.",
   "rationed": "Review cost. Reps were the writers and nobody could check each record, so a rule at the moment of saving was the cheapest gate available.",
   "challenge_claim": "Validation rules guarantee data, not quality. Require fields at the stage gate, only where a named report or routing rule reads them, and let AI fill what it can.",
   "blindspot": "Few posts connect the two failures. Placeholders under required fields and plausible wrong values from AI fills are the same mistake: completeness taken for accuracy. The AI-fill posts assume a filled field is now a true one, which is exactly what the old rule assumed.",
   "logic": {
    "prior_premises": [
     {
      "premise": "A filled field is a true field, so forcing the fill forces quality.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "The rep is the writer, and saving the record is the one moment to catch them.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Reports and routing need certain fields present to work.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "A field should be required only where a named report or routing rule reads it.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Asking at the stage gate, when the answer is known, gets truer values than asking on save.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "An AI-filled value is a true value, so AI ends the junk problem.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A filled field is a true field, so forcing the fill forces quality.",
    "deciding_side": "prior",
    "what_follows": "The premise that fails is the oldest one, that filled means true, and it failed long before AI. AI changed who writes the field, which makes the old failure harder to see: a placeholder looked like junk, and a model's wrong answer looks like data. What follows is the practitioners' revision, required at the stage gate with a named reader, plus a person who confirms the value rather than a rule that checks it exists."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Required fields and validation rules are built into the CRM, and enforcing them became the standard answer to bad data. If a rep can't save the opportunity without a close date, the opportunity will have a close date. The syncs honor it too. Marketo's Salesforce sync respects Salesforce's validation rules and logs a conflict when one rejects an update."
    },
    {
     "heading": "What people say now",
     "text": "Practitioners have turned against the blanket version, and the critique predates AI. Danny Gelfenbaum, a freelance Salesforce consultant: \"Validation rules don't guarantee quality data. They guarantee just data.\" Lars van Bergen, writing for Plauti, which sells Salesforce data-quality tools, describes the workaround everyone has seen: users \"get around validation rules with placeholder text like 'unknown', 'TBD', or a made-up value.\"\n\nThe revision they converge on is to ask at the stage gate, when the rep actually knows the answer. One consultant's fix, for a team whose fields were \"technically completed\" with \"little statistical value,\" was to gate them by stage. Gelfenbaum's own fix makes the right answer easier than the wrong one by prefilling the field, pulling it from where it already exists, and guiding the user through it."
    },
    {
     "heading": "Does the logic hold",
     "text": "The rule rested on three ideas. A filled field is a true field. The rep is the writer, and saving is the moment to catch them. And reports and routing need certain fields present. The third holds. The second changed because integrations, enrichment and AI now write many of the fields. The first was never true, and \"asdf\" in a required field proved it long before AI.\n\nThat is why the AI version of the challenge doesn't follow. \"Let AI fill the field\" keeps the premise that already failed. Ahoy, which sells an AI-native CRM, does the arithmetic: 95% extraction accuracy across 200 deals and six fields leaves 60 wrong values, and nothing marks which ones. A placeholder at least looked like junk. A wrong fill looks like data."
    },
    {
     "heading": "What both sides miss",
     "text": "Enforcement moves the garbage; it doesn't remove it. In my own setup, State and Country picklists reject two-letter codes, and a hard-block duplicate rule can't be overridden, so a legitimate create has to blank the email or move the record. The rule held. The data went around it."
    }
   ],
   "open_question": "When an AI fills a required field, what does \"required\" still check?",
   "sources": [
    {
     "side": "update",
     "who": "Danny Gelfenbaum, freelance Salesforce consultant",
     "date": "2025-07-28",
     "url": "https://www.linkedin.com/posts/danny-gelfenbaum_lets-add-a-validation-rule-cool-now-activity-7355579143784165376-02Ks",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Lars van Bergen, Plauti (sells Salesforce data-quality tools), Salesforce Ben branded content",
     "date": "2026-08-26",
     "url": "https://www.salesforceben.com/are-your-salesforce-validation-rules-creating-bad-data/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Ivan Brel, Salesforce consultant",
     "date": "2026-07-01",
     "url": "https://www.linkedin.com/posts/ivan-brel_when-required-fields-make-your-crm-data-worse-activity-7478031981482524672-v9Pq",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Kevin Bluett, co-founder, Ahoy (sells an AI-native CRM)",
     "date": "2026-09-07",
     "url": "https://ahoy.ai/guides/trust-ai-written-crm-records/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "crm",
    "data quality",
    "validation"
   ],
   "bench_notes": []
  },
  {
   "slug": "confidently-wrong-out",
   "title": "Garbage in, garbage out",
   "summary": "Keep it: agents make garbage in, garbage out more true, because they turn bad records into fluent, finished-looking answers and write plausible wrong values back at speed.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "stronger-with-agents",
   "because_label": "stronger with agents",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L1 Axiom",
   "origin": "Early-computing slang by 1957: the earliest print found is a syndicated Army piece that year quoting a programming specialist. The popular credit to IBM's George Fuechsel rests on a 2004 blog comment in which Fuechsel himself asks whether he coined it.",
   "rationed": "Capture cost. People typed the values, so most garbage was a blank or an obvious junk entry that a reader could spot and discount.",
   "challenge_claim": "Agents make it worse in a new way: they turn bad records into fluent, finished-looking answers, and they write plausible wrong values back at bulk speed. \"Garbage in, confidently wrong out.\"",
   "blindspot": "Both sides picture accidental garbage and a cleaner source as the fix. The garbage that bites now is written by tools and looks valid: a placeholder an enricher returns for \"no answer\", a confident match for the wrong person. And no data-quality post treats CRM free text as what it has become, an input an attacker can write.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Output can be no better than the input it was built from.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Bad input mostly looks bad: a blank or an obvious junk entry that a reader can spot and discount.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A person reads the output before anyone acts on it.",
      "now": "changed",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Models read through messy input, so inconsistent formats and free text stop breaking the output.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "The same tolerance covers wrong values, so a model can work around a stale title or a wrong owner.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A fluent, finished-looking answer is a sign the input was sound.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "The same tolerance covers wrong values, so a model can work around a stale title or a wrong owner.",
    "deciding_side": "challenge",
    "what_follows": "The core premise holds. What changed is that bad input no longer looks bad, and no person reads the output before an agent acts on it or writes it back. The case that models weaken the prior is right about format and fails on wrong values, so the prior stays in a sharper form."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The phrase was early-computing slang by 1957, when a syndicated Army piece quoted a programming specialist on it. The usual credit to IBM's George Fuechsel traces to a 2004 blog comment in which Fuechsel asks whether he coined it. The idea that a machine does what it's told with what it's given is older than both."
    },
    {
     "heading": "What people say now",
     "text": "Nobody I read argues it's dead. The argument has moved to how it fails, since it now fails confidently. Kevin Bluett of Ahoy, which sells an AI-native CRM, puts the mechanism plainly: \"A blank close date announces itself; a wrong one sits there looking exactly like a right one.\"\n\nThat's the new part. A messy spreadsheet looks messy. A summary built on the same rows looks finished."
    },
    {
     "heading": "Does the logic hold",
     "text": "GIGO rested on two ideas. Output can be no better than input. And bad input mostly looked bad, so a person reading a report could discount it. Underneath sat a third premise, that a person read the output before anyone acted on it.\n\nAI changed the second and third. Tools now write placeholders and wrong matches that look like real values. Agents act on the answer and write it back with no reader in between.\n\nThe case that AI weakens the prior rests on something real. Milko Slavov, who founded Felan to sell background agents to engineering teams, draws the line: \"Messy is not the same as wrong.\" Models absorb format debt like inconsistent casing, free text and odd abbreviations. But the case needs that tolerance to cover wrong values too, and it doesn't. Models amplify truth debt, like a stale title or a wrong owner. So the first premise holds, and the changes make it bite harder because the input is no better and the output now hides it."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides picture accidental garbage that a cleaner source would fix. The truth debt that hurts now is written by tools, and it looks valid. In my own setup, an enrichment step returned a placeholder string when it found no title, and that string landed in the title field. Every downstream reader treated it as a title. That value wouldn't have looked like garbage to a person skimming a report.\n\nThe other blind spot is intent. GIGO assumes the garbage is an accident. In 2025 researchers at Noma, which sells AI agent security, showed a web-to-lead description field could carry instructions an AI agent would follow. Every free-text field an agent reads is a prompt someone outside your company can write."
    }
   ],
   "open_question": "How do you tell a plausible wrong value from a right one without re-checking every value by hand?",
   "sources": [
    {
     "side": "origin",
     "who": "Pascal Tréguer, Word Histories, transcribing The Hammond Times, 10 Nov 1957",
     "date": "1957-11-10",
     "url": "https://wordhistories.net/2022/12/05/garbage-in-garbage-out/",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "Rob Stenson, Atlas Obscura (on the 1957 print use and Fuechsel's 2004 comment)",
     "date": "2016-03-14",
     "url": "https://www.atlasobscura.com/articles/is-this-the-first-time-anyone-printed-garbage-in-garbage-out",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Kevin Bluett, co-founder, Ahoy (sells an AI-native CRM)",
     "date": "2026-09-07",
     "url": "https://ahoy.ai/guides/trust-ai-written-crm-records/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Milko Slavov, founder, Felan (sells background agents for engineering teams)",
     "date": "2026-06-25",
     "url": "https://www.linkedin.com/posts/milkoslavov_aiadoption-aiengineering-buildinginpublic-activity-7475801252514373634-58Y3",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Noma Labs (Sasi Levi), ForcedLeak disclosure (Noma sells AI agent security)",
     "date": "2025-09-25",
     "url": "https://noma.security/blog/forcedleak-agent-risks-exposed-in-salesforce-agentforce",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "data quality",
    "agents",
    "crm"
   ],
   "bench_notes": [
    {
     "slug": "sentinel-values-look-like-data",
     "title": "Placeholder values in required fields look like real data"
    },
    {
     "slug": "invented-because-passes-grounding",
     "title": "An invented \"because\" passes a claim-by-claim grounding check"
    }
   ]
  },
  {
   "slug": "one-writer-per-field",
   "title": "Every field has one owning system",
   "summary": "Keep it: every agent, enrichment step and sync adds a writer, so naming one owning system per field matters even more with agents.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "stronger-with-agents",
   "because_label": "stronger with agents",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L3 Practice",
   "origin": "Master-data and CRM governance canon: a data dictionary names the system that owns each field. Sync tools encode it as conflict rules per field. Marketo's Salesforce sync documentation still resolves simultaneous updates with a fixed winner by record type: Marketo for leads, Salesforce for contacts.",
   "rationed": "Data-capture cost. Few processes could write a field, so naming its owner was cheap and drift was slow.",
   "challenge_claim": "AI agents can fill and fix fields across the stack, so field ownership becomes a formality.",
   "blindspot": "Every agent, enrichment waterfall and sync adds a writer. The rule's cost went down (code can enforce it) at the same moment its value went up. \"AI fills the field\" posts rarely ask which writer wins.",
   "logic": {
    "prior_premises": [
     {
      "premise": "When two processes write one field with no rule for which wins, the field drifts and nobody can tell why.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Few processes write any one field, so naming an owner is cheap and drift is slow.",
      "now": "changed",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "An AI that fills and fixes fields keeps every value current, so which writer wins stops mattering.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "An agent writing a field settles conflicts between writers rather than adding to them.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A recent write means the value is current.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "An agent writing a field settles conflicts between writers rather than adding to them.",
    "deciding_side": "challenge",
    "what_follows": "The premise that changed is the number of writers, and it moved against the claim: more writers make an unowned field drift faster. The claim needs AI to settle conflicts between writers, but an agent is one more party to them. Enforcement also got cheaper, because code can hold the gate, so the prior holds and is worth more."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "The data-governance canon calls for a data dictionary, one owning system per field, and conflict rules for the syncs. Marketo's Salesforce sync documentation still settles simultaneous updates with a fixed winner by record type, where Marketo wins for leads and Salesforce wins for contacts."
    },
    {
     "heading": "What people say now",
     "text": "Almost nobody argues the rule is dead. The live claim is softer: if AI can fill and fix fields across the stack, ownership stops mattering because the AI keeps everything current. The CRM vendor's own guidance pushes the other way. Salesforce's architecture advice for teams deploying agents warns that automations nobody still needs \"create unexplained and potentially unwanted changes to data.\" Its readiness check asks for a named owner and a formal sign-off on schema changes, calling any gap \"a prerequisite for deploying AI agents\" that must be finished beforehand. The rule has moved from good hygiene to a prerequisite."
    },
    {
     "heading": "Does the logic hold",
     "text": "The rule rested on two ideas. When two processes write one field with no rule for which one wins, the field drifts and nobody can tell why. And few processes wrote any one field, so naming an owner was cheap and drift was slow.\n\nAI changed the second idea in the opposite direction from what the claim assumes. Every agent is another writer, and so is every enrichment step and every sync someone adds to feed an agent. The first idea still holds, so more writers means faster drift.\n\nThe claim needs the AI to settle conflicts, keeping every value current whatever else writes. It doesn't settle them; it joins them. Meanwhile the cost of keeping the rule fell, because code can hold the gate. A rule that got cheaper to keep and more costly to break follows straight through to keep."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides argue about the rule in general. The failures are specific. In my own setup, the worst corruption of the year came from a \"no longer at company\" field with four writers. One was a weekly enrichment refresh that marked a person gone whenever any one email validator said \"invalid\". For weeks nobody knew which writer was doing it, and the blame kept landing on the wrong tools.\n\nA second case looked the opposite way. A vendor field seemed current because its last-modified date was recent. Other tools were touching the same rows, and the vendor itself had stopped writing months earlier. A last-modified date tells you that something wrote to the row. It doesn't tell you which writer is alive.\n\nTwo failure shapes repeat:\n\n- A field that marks something as gone (departed, churned, closed) gets filled by processes that each mean something different by it.\n- A sync that loses access to a field reads it as empty and \"corrects\" it."
    }
   ],
   "open_question": "How many of your agents can write a field that a sync or enrichment tool also writes, and which one wins?",
   "sources": [
    {
     "side": "origin",
     "who": "Adobe Marketo Engage docs, SFDC Sync: Lead Sync (simultaneous changes: Marketo wins)",
     "date": "2026-10-07",
     "url": "https://experienceleague.adobe.com/en/docs/marketo/using/product-docs/crm-sync/salesforce-sync/sfdc-sync-details/sfdc-sync-lead-sync",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "Adobe Marketo Engage docs, SFDC Sync: Contact Sync (simultaneous changes: Salesforce wins)",
     "date": "2026-10-07",
     "url": "https://experienceleague.adobe.com/en/docs/marketo/using/product-docs/crm-sync/salesforce-sync/sfdc-sync-details/sfdc-sync-contact-sync",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Allison Park, Principal Enterprise Architect, Salesforce (sells the CRM and its agents), Center of Excellence guidance",
     "date": "2026-08-11",
     "url": "https://www.salesforce.com/blog/center-of-excellence-ai-governance/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "crm",
    "data quality",
    "governance"
   ],
   "bench_notes": [
    {
     "slug": "left-the-company-many-writers",
     "title": "\"Left the company\" has many writers and many meanings"
    },
    {
     "slug": "invisible-field-looks-empty",
     "title": "CRM permissions make an invisible field look empty"
    }
   ]
  },
  {
   "slug": "crm-keeps-your-history",
   "title": "The CRM forgets, so snapshot what you'll need later",
   "summary": "Keep it: agents make the old warning sharper, because a CRM that forgets its history hands the default query a zero that an agent reports as fact.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "stronger-with-agents",
   "because_label": "stronger with agents",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L4 Tool-specific",
   "origin": "Salesforce's own documentation. By default, events and closed tasks more than 365 days past their due date are archived (Support can raise the limit to 3,650 days) and drop out of reports, list views and search. Field history tracking covers up to 20 fields per object, shows 18 months in the interface and is removed after 24. Historical trend reporting keeps the previous three months plus the current one by default, or up to 12 months of opportunity history with Pipeline Inspection's historical trending. Admins have long answered with weekly pipeline snapshots.",
   "rationed": "Storage cost. Keeping every version of every record was expensive, so the platform kept current values and a bounded trail.",
   "challenge_claim": "Agents need the history more than people did: slip history for forecasting, old activity for account context. Yet AI that rewrites stage and close date overwrites the trail it would need, and agent-readiness advice says to archive aggressively.",
   "blindspot": "\"Archive for agents\" and \"keep history for forecasting\" are both live advice, and nobody reconciles them. And the failure mode changed: a person looking at an empty report doubts it, while an agent on the default query reports zero as a fact.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Keeping every version of every record is expensive, so the platform keeps current values and a bounded trail.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "History you don't capture yourself is gone when you need it.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "A person who meets a gap in the history notices it and doubts the report.",
      "now": "changed",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Agents work better on a smaller set of current, relevant records.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "Old history is noise for the work agents do.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "An agent that keeps fields current improves the record.",
      "now": "fails",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Old history is noise for the work agents do.",
    "deciding_side": "challenge",
    "what_follows": "The platform's limits haven't moved, so the old premises hold. What changed is the reader: an agent on the default query reports a gap as a finding. The case for archiving for agents rests on an unproven premise and runs against the uses agents need history for, so the prior stays, in a sharper form."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "From the platform's own documentation. By default Salesforce archives events and closed tasks a year past their due date, and archived rows drop out of reports, list views and search. Field history covers a limited set of fields per object and ages out after two years. Trend reporting keeps a few months. Admins learned to snapshot pipeline weekly, because a report run today shows today's values: \"Reports run on a date also reflect today's values, not that day's.\""
    },
    {
     "heading": "What people say now",
     "text": "Nobody argues the CRM keeps everything. The new point is that AI makes the loss worse. Kevin Bluett of Ahoy, which sells an AI-native CRM: \"A record that keeps only the latest value has thrown away the signal and kept the noise.\" An agent that keeps close dates current erases the slips a forecaster would learn from. Some tooling vendors advise archiving more so agents work on current records alone."
    },
    {
     "heading": "Does the logic hold",
     "text": "The habit rested on two ideas. Keeping every version of every record cost too much, so the platform kept current values and a bounded trail. And whatever you didn't capture yourself would be gone when you needed it. Neither has changed. The platform's defaults are the same. A third idea sat under the habit. A person who saw an empty activity report for an old year knew something was off. That is the one AI changed. An agent on the default query reports the blank as a finding.\n\nThe case for archiving for agents needs agents to do better on a smaller, current set of records, with nothing behind it beyond vendor advice. It also needs old history to be noise, and the work agents are given says otherwise. Forecasting learns from slips, and account context comes from old activity. So the prior follows through to keep, and the changed reader makes it sharper."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides treat the gap as a policy question. It shows up first as a wrong finding. In my own setup, the same query on the same contacts returned a few percent of one year's events by default and none at all for the year before. With archived rows included, the history was all there. A finding that \"logging started late last year\" was this artifact. So were two separate audits of \"missing\" activities, where every missing row turned out to be archived.\n\nThe fields I most needed to audit also turned out to have no history at all. Tracking had never been switched on for them, so I couldn't prove whether anything had written them. A monitor I built on last-modified time raised false alarms on any touch, and I rewrote it to compare snapshots."
    }
   ],
   "open_question": "How do you give agents a clean working set without archiving away the history they would learn from?",
   "sources": [
    {
     "side": "origin",
     "who": "Salesforce Help, Increase Archive Days Limit for Activities",
     "date": "2026-06-25",
     "url": "https://help.salesforce.com/s/articleView?id=000385669&language=en_US&type=1",
     "standing": "authority"
    },
    {
     "side": "origin",
     "who": "Salesforce Help, Field History Tracking Overview",
     "date": "current",
     "url": "https://help.salesforce.com/s/articleView?id=xcloud.tracking_field_history.htm&language=en_US",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "Sol Flamberg, Revenue Best Practices (newsletter; sells spreadsheet templates)",
     "date": "2026-10-23",
     "url": "https://www.revenuebestpractices.com/p/pipeline-snapshot",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Kevin Bluett, co-founder, Ahoy (sells an AI-native CRM)",
     "date": "2026-09-07",
     "url": "https://ahoy.ai/guides/trust-ai-written-crm-records/",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "crm",
    "history",
    "forecast"
   ],
   "bench_notes": [
    {
     "slug": "soft-delete-keeps-the-crm-stamp",
     "title": "A soft delete may not move your copy's modified time"
    }
   ]
  },
  {
   "slug": "governed-metric-definitions",
   "title": "Every metric gets one governed definition",
   "summary": "Keep it: an agent reads a metric's definition literally on every question, so an ambiguous metric now produces a confident wrong answer instead of a meeting about whose number is right.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "stronger-with-agents",
   "because_label": "stronger with agents",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L2 Operating model",
   "origin": "Data-engineering canon: define a metric once, use it everywhere. Robert Chang's 2021 Airbnb Engineering post describes Minerva as the single source of truth for metrics, with lineage to a 2014 metrics repo. Tristan Handy of dbt Labs wrote in 2023 that no standard semantic layer exists and that past ones led to lock-in. The RevOps form, \"one definition of pipeline\", is a practitioner extension.",
   "rationed": "Coordination cost. Agreeing one formula meant getting finance, sales and marketing to sign the same thing, and only people read the result, so most metrics layers stopped well short of full coverage.",
   "challenge_claim": "General-purpose AI lets anyone ask the data directly, and a context graph will carry the meaning, so a hand-governed definition layer becomes overhead.",
   "blindspot": "Both sides talk about the layer as tooling. The work was always the agreement: someone has to get three functions to sign one formula, and someone has to stop it drifting afterwards. A semantic layer stores a decision; it doesn't make one.",
   "logic": {
    "prior_premises": [
     {
      "premise": "When two teams compute one metric differently, people stop trusting every number.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Agreeing a definition is organizational work that no tool does for you.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Only people read definitions, and they could ask what a number meant, so partial coverage was tolerable.",
      "now": "changed",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "A model can infer the intended definition from the fields and the data.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Fast answers make ambiguity cheap, because anyone can ask again.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A context graph can carry a metric's meaning without anyone agreeing a definition first.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "A model can infer the intended definition from the fields and the data.",
    "deciding_side": "challenge",
    "what_follows": "The premise that changed is who reads the definition: an agent now reads it on every question, literally, with nobody to ask. That raises the cost of an ambiguous metric, so the prior gets stronger. The challenge needs a model to infer intent from field names, which is exactly where practitioners debugging wrong answers find it fails."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "For a decade, data teams have said to define a metric once and use it everywhere. Airbnb built Minerva as one store for metric definitions, so every dashboard and experiment read the same formula. Tristan Handy of dbt Labs later wrote that no standard semantic layer exists and that past attempts led to lock-in. RevOps borrowed the idea as \"one definition of pipeline\", usually as a data dictionary rather than software.\n\nThe idea mostly stalled. One engineer describes the typical metrics layer as \"built grudgingly, adopted partially, and left at 60% coverage.\""
    },
    {
     "heading": "What people say now",
     "text": "The update case comes from investors. Those quoted by Benn Stancil say \"that context graph becomes the real source of truth for autonomy.\" The claim implies the model will hold the meaning.\n\nPractitioners who debug wrong AI answers say the opposite: \"the model can read your data but not the rules around your data.\" One engineer puts the build order bluntly: \"semantic layer first, agent second.\" dbt Labs, which sells the layer, now pitches it as the contract an agent calls."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rested on three ideas. Two teams computing one metric differently erode trust in every number. Agreeing a definition is organizational work. And because only people read definitions, a partial layer was tolerable: someone who doubted a number could ask what it meant.\n\nAI changed the third. An agent reads the definition on every question, literally, with nobody to ask. If \"customer\" can mean three things, it picks one and answers with confidence. The first two ideas still hold, so the cost of an ambiguous metric went up.\n\nThe challenge needs a model to infer the intended definition from field names and data. That is the step practitioners keep finding it can't take. A context graph may carry meaning one day, but nobody has shown it working without the agreement underneath."
    },
    {
     "heading": "What both sides miss",
     "text": "The layer is the easy part. One semantic-layer vendor says it honestly: \"Somebody has to get Finance, Sales, and Marketing to agree on one formula for pipeline.\" Stancil's deflation of the context-graph pitch lands in the same place: \"It's a text box, filled out by an AI.\"\n\nIn my own CRM, \"customer\" lived in three fields that disagreed, and the field named like a segment turned out to be a priority tier. Closed-won was a literal stage string, and renewals counted as open. None of that shows in the field names, and the field names are what an agent sees."
    }
   ],
   "open_question": "Who signs off a metric definition in your org, and what stops it drifting once agents read it on every question?",
   "sources": [
    {
     "side": "defend",
     "who": "Tian Pan, engineer blog",
     "date": "2026-07-02",
     "url": "https://tianpan.co/blog/2026/07/02/your-data-agent-needs-one-definition-of-revenue",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Josh Buchanan, gtmjosh (enterprise GTM engineer; sells nothing on the site)",
     "date": "2026-08-20",
     "url": "https://gtmjosh.com/reference/debugging-wrong-crm-ai-answers",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Kathryn Chubb, dbt Labs (sells a semantic layer)",
     "date": "2026-01-26",
     "url": "https://www.getdbt.com/blog/centrally-defined-metrics",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Benn Stancil, Have you tried a text box?",
     "date": "2026-01-02",
     "url": "https://benn.substack.com/p/have-you-tried-a-text-box",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "definitions",
    "semantic layer",
    "reporting",
    "governance"
   ],
   "bench_notes": []
  },
  {
   "slug": "deterministic-routing",
   "title": "Route and match leads with deterministic rules",
   "summary": "Keep it: AI now produces more of what routing reads, so the step that turns those guesses into an owner has to stay a rule anyone can replay, which matters even more with agents.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "stronger-with-agents",
   "because_label": "stronger with agents",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L3 Practice",
   "origin": "No single author. Salesforce's lead assignment rules route by geography or queue. Lead-to-account matching arrived with account-based selling; LeanData, which sells routing, says it was first to solve it (Kim Peterson, 2023). LeanData says Salesforce has no native lead-to-account match, while Salesforce Help now lists a standard rule.",
   "rationed": "Trust in assignment. Ownership decides who works a lead and who gets the credit, so every assignment had to be explainable afterwards, and only a fixed rule could be replayed.",
   "challenge_claim": "Rule-based, human-configured routing is a weakness: an AI router can read the whole context and assign better than a rule tree someone has to maintain.",
   "blindspot": "Defenders treat deterministic as correct. A rule replays perfectly on whatever it reads, including modeled headcount and an owner field that records admin convenience. Critics miss that the rule is what makes those errors findable.",
   "logic": {
    "prior_premises": [
     {
      "premise": "An assignment decides who works a lead and who gets credit, so it has to be explainable after the fact.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "A rule gives the same answer on the same input, so a disputed assignment can be replayed.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "An AI router assigns leads more accurately than rules do.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "An assignment doesn't need to be replayable as long as it is usually right.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Maintaining a rule tree costs more than auditing an AI's choices.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "An assignment doesn't need to be replayable as long as it is usually right.",
    "deciding_side": "challenge",
    "what_follows": "Neither premise under the prior depends on AI. What AI changed is upstream: scores, tiers and intent readings now come from models, so more of what routing reads is a guess, and a final step you can replay is worth more. The challenge rests on an accuracy edge nobody has measured, and on assignments not needing replay, which fails the first time a rep disputes one."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Routing has always been a rule tree. Salesforce's lead assignment rules send a lead by geography or to a queue. Lead-to-account matching came later, when account-based selling needed every new lead tied to the account someone already owned. LeanData, which sells routing, says it was first to solve that match. The principle under both has no author. The same lead with the same data goes to the same person, every time."
    },
    {
     "heading": "What people say now",
     "text": "LeanData's 2026 State of Martech report puts AI lead routing and assignment at 11% adoption, \"the lowest of any category measured.\" That describes the market, while nobody has measured whether AI routes better. A RevOps Co-op panel stated the consensus plainly: \"AI is not ready to make those calls alone.\" The line that keeps recurring in practitioner threads is shorter: the LLM proposes, the code disposes.\n\nThe public challenge is thin. It comes mostly from competitors' comparison pages, which treat rule-based, human-configured routing as the flaw. LeanData's own answer is to use AI for composite tiers in place of point-sum scores and keep the router deterministic."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rests on two ideas. An assignment decides who works a lead and who gets credit, so it has to be explainable after the fact. And a rule gives the same answer on the same input, so a disputed assignment can be replayed. AI doesn't touch either one.\n\nWhat AI changed is upstream. Scores, tiers and intent readings now come from models, so more of what routing reads is a guess. That makes a final step you can replay worth even more.\n\nThe challenge needs two things. It needs an AI router that assigns better, which nobody has measured. And it needs assignments that don't have to be replayed, which fails the first time a rep disputes one. Neither follows from anything AI has actually changed."
    },
    {
     "heading": "What both sides miss",
     "text": "A rule replays perfectly on whatever it reads. Headcount and revenue fields are \"modeled estimates, not verified facts,\" in a data vendor's own words. In my own setup, the owner field mostly recorded admin convenience rather than who worked the account, so \"task owner equals account owner\" could never be a routing rule. Deterministic isn't the same as right. What a rule buys you is a wrong answer you can find."
    }
   ],
   "open_question": "If an AI router did assign better, how would you show a rep why they lost the lead?",
   "sources": [
    {
     "side": "nuance",
     "who": "LeanData (sells routing), State of Martech and RevOps 2026 (market state)",
     "date": "2026-05-21",
     "url": "https://www.leandata.com/state-of-martech-revops-report-2026/",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Kim Peterson, LeanData (sells routing), AI use cases in lead routing",
     "date": "2026-05-05",
     "url": "https://www.leandata.com/blog/ai-gtm-lead-routing-use-cases/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Apollo (data vendor), how accurate are revenue and employee counts",
     "date": "2026-05-12",
     "url": "https://www.apollo.io/insights/how-accurate-are-company-revenue-and-employee-counts-in-a-b2b-data-platform",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "routing",
    "lead management",
    "lead-to-account matching",
    "governance"
   ],
   "bench_notes": []
  },
  {
   "slug": "consolidate-the-stack",
   "title": "Consolidate the stack",
   "summary": "Keep it: every AI assistant and agent is one more place an answer can come from, so having fewer sources of answers matters more now, even where the number of vendors doesn't.",
   "verdict": "keep",
   "verdict_label": "Keep it",
   "because": "stronger-with-agents",
   "because_label": "stronger with agents",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L2 Operating model",
   "origin": "No traceable origin. Salesforce, which sells the platform teams would consolidate onto, has pushed it through its State of Sales research: in the 2022 survey reps used an average of ten tools, and nine in ten sales organizations planned to consolidate within a year. Gartner's 2021 RevOps guidance lists simplifying the stack by eliminating redundant enablement technology.",
   "rationed": "Admin and integration hours. Each tool needed its own sync, owner and upkeep, and every copy of customer data was one more to keep in line.",
   "challenge_claim": "Agents and connectors can stitch any set of tools together, so the job is to connect what you have rather than cut it, and to add AI tools wherever they help.",
   "blindspot": "Consolidation has usually meant fewer vendors. The AI-era version is fewer places an answer can come from, and cutting vendors doesn't get you there by itself. Sprawl also relocates: the tables, scripts and agents that replace a tool are a stack of their own.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Every tool that keeps its own copy of customer data is one more place for answers to disagree.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Integration and upkeep cost grows with every tool you add.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "Fewer tools make reps more productive.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Connecting tools settles which tool's answer is right.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Each new AI tool adds more value than the extra source of answers costs.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Connecting tools settles which tool's answer is right.",
    "deciding_side": "challenge",
    "what_follows": "AI changed the number of places an answer can come from: every assistant, agent and enrichment step can now say who owns an account or whether it's a customer. With the first premise holding, that makes disagreement more likely, and in general question-answering, conflicting sources measurably lower a model's accuracy, so the prior gets stronger. The challenge needs connectors to settle disagreements, but a connector moves data and doesn't pick a winner. The productivity premise stays unproven, and the verdict doesn't need it."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody owns this one. It has run through vendor and analyst content for years. Salesforce, which sells the platform teams would consolidate onto, found reps using an average of ten tools in its 2022 survey, and nine in ten sales organizations planning to consolidate within a year. Gartner's 2021 RevOps guidance told leaders to eliminate redundant enablement technology. A plan to consolidate remains an intention, though."
    },
    {
     "heading": "What people say now",
     "text": "Almost nobody argues for more tools. The argument has moved to AI, as data vendors now pitch a fragmented stack as a cause of hallucination and tool sprawl as a cost nobody measures.\n\nThe pushback is about how, not whether. Winning by Design argues for stitching over replacing: \"It's how well the tools you already have connect.\""
    },
    {
     "heading": "Does the logic hold",
     "text": "Three ideas sat under the prior. Every tool with its own copy of customer data is one more place for answers to disagree. Integration and upkeep grow with each tool. And fewer tools make reps more productive.\n\nThe third has never been tested by tool count. The nearest evidence is Gartner's seller survey: about half of sellers said they were overwhelmed by the number of technologies their job needs, and those sellers were 43% less likely to hit quota, a result its 2024 survey repeated. That is perceived overload, self-reported, at one point in time, so it is suggestive and no more. The productivity premise stays unproven, and the case doesn't need it.\n\nAI changed the count of places an answer can come from. Every assistant, agent and enrichment step can now say who owns an account or whether it is a customer. With the first idea still true, more answerers mean more disagreement. Research on language models shows what that does. In ClashEval, a Stanford benchmark of general question-answering, models given a wrong retrieved document adopted it over their own correct answer more than 60% of the time. The questions weren't about CRM records, but the mechanism is the one that matters here. When a source disagrees, a model often takes the source's side, right or wrong.\n\nThe stitching argument assumes that connecting tools resolves which one is right. It doesn't. A connector moves data between tools; it doesn't decide which tool wins, and the same research shows models don't reliably decide either."
    },
    {
     "heading": "What both sides miss",
     "text": "Both sides count vendors. The thing to count is sources of answers. In my own setup, the sprawl moved into the enrichment tool: a workspace full of tables, most of them finished one-off jobs, and one mirror table with several live writers into the CRM. An audit found none of it needed to stay a table. Cutting a vendor would not have surfaced that. Asking which tool answers each question did."
    }
   ],
   "open_question": "How many of your tools can answer \"is this account a customer?\", and do they agree?",
   "sources": [
    {
     "side": "origin",
     "who": "Salesforce (sells the consolidation platform), State of Sales research",
     "date": "2023",
     "url": "https://www.salesforce.com/news/stories/sales-research-2023/",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Wu et al., ClashEval, NeurIPS 2024 Datasets and Benchmarks (general question-answering)",
     "date": "2024",
     "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/3aa291abc426d7a29fb08418c1244177-Paper-Datasets_and_Benchmarks_Track.pdf",
     "standing": "authority"
    },
    {
     "side": "update",
     "who": "Winning by Design (sales consultancy), Jacco van der Kooij brief",
     "date": "2026-07-14",
     "url": "https://www.linkedin.com/posts/winningbydesign_dont-rip-and-replace-the-stack-stitch-it-activity-7482792286427230208-gmtv",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Gartner (sells advisory, not sales tools), Technology's Impact on Seller Productivity survey",
     "date": "2024-09-16",
     "url": "https://www.gartner.com/en/newsroom/press-releases/2024-09-16-gartner-sales-survey-reveals-sellers-who-partner-with-ai-re-three-point-seven-times-more-likely-to-meet-quota",
     "standing": "authority"
    }
   ],
   "verdict_history": [],
   "tags": [
    "tech stack",
    "consolidation",
    "integrations",
    "governance"
   ],
   "bench_notes": []
  },
  {
   "slug": "fix-data-before-ai",
   "title": "Fix your data before you deploy AI",
   "summary": "Too early to call: almost everyone agrees with fixing data first, but the real fight is over sequencing, each side sells the step it puts first, and neither has settled which order works.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L1 Axiom",
   "origin": "No single originator. The canonical citation is Gartner's February 2025 press release and Q&A on AI-ready data: a July 2024 survey of 1,203 data-management leaders, and a prediction (often quoted as a finding) that organizations will abandon 60% of AI projects that lack AI-ready data through 2026. The same piece calls AI-ready data an ongoing practice, not a one-off.",
   "rationed": "Project cost. An AI or analytics build was a large one-off bet and cleanup was a manual project, so cleaning first protected the investment.",
   "challenge_claim": "Waiting for clean data is the trap. Contact data will never be clean, so start now, keep core business data clean as you go, and run cleanup as a permanent control instead of a project.",
   "blindspot": "Most voices on both sides are vendors: the \"fix first\" chorus sells the fix, and the \"don't wait\" voices sell agents. Neither separates fields that are only read from fields that trigger an action. \"Models read the mess\" is true for documents and says nothing about the fields that drive routing, comp and forecast, where a wrong value acts.",
   "logic": {
    "prior_premises": [
     {
      "premise": "An AI build is a large one-off bet, so cleaning first protects the investment.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Cleanup is a project with an end, after which the data stays clean.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "A wrong value in the data produces a wrong outcome from the AI.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Models get useful answers from messy contact data and documents.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "The same tolerance covers fields that act, like stage and owner.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Fixing only what breaks, as it breaks, costs less than cleaning first.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Fixing only what breaks, as it breaks, costs less than cleaning first.",
    "deciding_side": "challenge",
    "what_follows": "What changed is the sequence, not the need: cleanup is a permanent control rather than a project, and an AI build is no longer one large bet. A wrong value in a field that acts still does damage, and both sides accept it. The deciding premise, that fixing as you go beats fixing first, rests on a founder's claim and the sellers of each step, so it is unproven and the verdict stays too early."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "No one person said it first. A February 2025 Gartner press release on AI-ready data is the citation most posts lean on, usually through its prediction that companies will abandon most AI projects that lack AI-ready data. That prediction gets quoted as a measured result. The same Gartner piece says AI-ready data is an ongoing practice, which is a long way from \"clean it, then build\"."
    },
    {
     "heading": "What people say now",
     "text": "The principle is near-universal. LeanData, which sells routing and data tooling, found 82% of its survey agreeing that \"clean data and reliable routing must come before scaling AI,\" and only one in three with systems to do it. The gap is execution, not belief.\n\nThe pushback comes from people who build agents. Milko Slavov, who founded Felan to sell background agents, calls \"fix your data first\" \"the most expensive piece of outdated AI advice still being sold.\" His alternative: point a model at the mess, watch where it breaks, fix only that."
    },
    {
     "heading": "Does the logic hold",
     "text": "Fixing data first rested on three ideas. An AI build was a large one-off bet. Cleanup was a project with an end. And a wrong value in the data produced a wrong outcome. The first two have changed. Gartner's own piece calls AI-ready data an ongoing practice. Spencer Hardey, who runs business operations at the data vendor HG Insights, has AI flag duplicate accounts and a person approve each merge. He puts it plainly: \"It's not a cleanup project. It's a permanent control.\" Agents can be shipped small and corrected as they go.\n\nThe third idea still holds, so the sequence changed while the need stayed. The \"don't wait\" case needs models to read through the mess. For documents and email threads they do. Slavov draws the line himself: \"Messy is not the same as wrong.\" It fails for fields that act. A wrong stage or owner routes a lead, moves a forecast or pays commission. The premise that would decide it, that fixing what breaks costs less than cleaning first, has one founder's claim behind it and no data."
    },
    {
     "heading": "What both sides miss",
     "text": "The sellers of each step set the sequence. Cleanup vendors put cleanup first, and agent vendors put the agent first. Neither has an interest in testing the other order.\n\nThe \"fix first\" side tends to equate clean with complete. Required fields filled under pressure give you \"TBD\" and template values. As Danny Gelfenbaum, a freelance Salesforce consultant, put it: \"Validation rules don't guarantee quality data. They guarantee just data.\"\n\nI didn't wait. I shipped agents on imperfect data and put the cleaning into the write path: an identity check before any write about a person, a scan of value counts before any bulk write, two independent sources before a verdict. What actually limited me was where a value came from and how current it was, not how tidy the table looked."
    }
   ],
   "open_question": "What is the minimum bar a field that acts, like stage or owner, must clear before an agent may read it unsupervised?",
   "sources": [
    {
     "side": "origin",
     "who": "Gartner, Lack of AI-Ready Data Puts AI Projects at Risk (press release and Q&A)",
     "date": "2025-02-26",
     "url": "https://www.gartner.com/en/newsroom/press-releases/2025-02-26-lack-of-ai-ready-data-puts-ai-projects-at-risk",
     "standing": "authority"
    },
    {
     "side": "defend",
     "who": "LeanData (sells routing and data tooling), State of Martech and RevOps 2026",
     "date": "2026-05-21",
     "url": "https://www.leandata.com/state-of-martech-revops-report-2026/",
     "standing": "position"
    },
    {
     "side": "update",
     "who": "Milko Slavov, founder, Felan (sells background agents for engineering teams)",
     "date": "2026-06-25",
     "url": "https://www.linkedin.com/posts/milkoslavov_aiadoption-aiengineering-buildinginpublic-activity-7475801252514373634-58Y3",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Spencer Hardey, VP Business Operations, HG Insights (sells B2B technology data)",
     "date": "2026-04-28",
     "url": "https://www.linkedin.com/posts/spencer-hardey_revopsaiweekly-revops-aiautomation-activity-7454690637284560896-QUwg",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Danny Gelfenbaum, freelance Salesforce consultant",
     "date": "2025-07-28",
     "url": "https://www.linkedin.com/posts/danny-gelfenbaum_lets-add-a-validation-rule-cool-now-activity-7355579143784165376-02Ks",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "data quality",
    "agents",
    "sequencing"
   ],
   "bench_notes": [
    {
     "slug": "sentinel-values-look-like-data",
     "title": "Placeholder values in required fields look like real data"
    }
   ]
  },
  {
   "slug": "dashboard-is-the-deliverable",
   "title": "The dashboard is the deliverable",
   "summary": "Too early to call: chat can replace the exploratory dashboard only where it reads a governed layer, since plain text-to-SQL still gets 10 to 35% of questions wrong, and nobody has measured whether decisions made from chat hold up as well as those made from views.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L2 Operating model",
   "origin": "No traceable originator. It is built into the reporting pitch: reports and dashboards created fast on live data, so a business user can see the numbers without waiting on an analyst. Salesforce's 2015 pitch promised analytics for \"anyone who can use a spreadsheet\".",
   "rationed": "Analysis cost. Insight needed an analyst and a reporting cycle, so the questions people would ask were guessed in advance and built into views.",
   "challenge_claim": "Anyone can ask the data a question in chat now, so the dashboard is dead and analysis moves to an AI interface.",
   "blindspot": "The update side treats every chat answer as right, when ungoverned chat returns a plausible wrong number on a large share of questions. The defend side misses that the governed layer behind its agreed number is exactly what makes chat reliable. The common claim that most dashboards go unused has no traceable source.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Analysis needs an analyst, so questions have to be anticipated and built into views.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "A team needs one agreed number to run on, shown the same way to everyone.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "The views people build are the views people use.",
      "now": "unproven",
      "kind": "prevalence"
     }
    ],
    "challenge_premises": [
     {
      "premise": "Anyone can ask the data a question in plain language and get an answer.",
      "now": "holds",
      "kind": "prevalence"
     },
     {
      "premise": "Chat answers match the number the team runs on, with no governed layer underneath.",
      "now": "fails",
      "kind": "causal"
     },
     {
      "premise": "Decisions made from chat answers hold up as well as decisions made from pre-built views.",
      "now": "unproven",
      "kind": "causal"
     }
    ],
    "deciding_premise": "Decisions made from chat answers hold up as well as decisions made from pre-built views.",
    "deciding_side": "challenge",
    "what_follows": "The cost of asking changed, and chat over a governed semantic layer now matches the agreed number almost every time in scope, so the challenge can work where that layer exists. Plain text-to-SQL still answers 10 to 35% of questions wrongly in ways that look right, so \"the dashboard is dead\" fails as a general claim. Whether decisions made from chat answers hold up as well as decisions made from views is unmeasured, so the verdict waits on outcomes."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Nobody is credited with it. It is built into how reporting was sold: dashboards on live data, so a business user could see the numbers without waiting on an analyst. Salesforce's 2015 pitch promised analytics for \"anyone who can use a spreadsheet.\" The scarce thing was analysis, so the questions people would ask were guessed in advance and built into views."
    },
    {
     "heading": "What people say now",
     "text": "The update case comes mostly from vendors. Dremio's CMO, whose company sells a lakehouse, titled a post \"The dashboard is dead\" and wrote that \"we have no 'marketing analyst' or 'BI engineer' on the team.\" Analyst forecasts in the trade press say most self-service analytics users will soon rely on general-purpose chatbots. A forecast is not a finding.\n\nThe strongest pushback comes with numbers, from vendors who sell governed data layers, and it is about accuracy rather than dashboards."
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rested on three ideas. Analysis needed an analyst, so questions had to be anticipated and built into views. A team needs one agreed number, shown the same way to everyone. And the views people build are the views they use, which nobody has measured either way.\n\nAI changed the first idea, because asking no longer needs an analyst or a view built in advance. The challenge needs two more things. Chat answers have to match the number the team runs on, and decisions made from those answers have to hold up as well as decisions made from views.\n\nThe first of those is now measured, and it splits on one condition. In a 2023 test on an insurance schema, GPT-4 writing SQL straight against the tables got about 16% of enterprise questions right, against about 54% over a knowledge-graph layer (Sequeda and colleagues at data.world, which sells knowledge graphs). dbt Labs, which sells a semantic layer, reran the benchmark in 2026. Plain text-to-SQL had improved to about 65%, and to 84 to 90% with extra modelling. The semantic layer scored 98 to 100% on in-scope questions and refused the rest rather than guess. In dbt's words, text-to-SQL \"will cheerfully give you a wrong number.\" So chat can replace exploration, but only where it reads a governed layer.\n\nThe second need has no evidence at all. No study of view logs shows pre-built views going unused, and none compares decisions made from chat answers with decisions made from views. Until that is measured, \"exploration has moved to chat\" remains a forecast."
    },
    {
     "heading": "What both sides miss",
     "text": "The two jobs a dashboard did come apart. Exploration can move to chat, where it was always a poor fit for a fixed view. Agreement stays, and it gets more important, because chat makes it cheap to produce a second version of any number.\n\nThe common figure that 60 to 70% of dashboards go unused has no source. Neither side should lean on it."
    }
   ],
   "open_question": "When two people ask the AI the same question and get different numbers, which one goes in the board deck?",
   "sources": [
    {
     "side": "update",
     "who": "Read Maloney, CMO Dremio (sells a lakehouse), \"The dashboard is dead\"",
     "date": "2026-04-02",
     "url": "https://www.dremio.com/blog/the-dashboard-is-dead/",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Sequeda, Allemang and Jacob, data.world (sells knowledge graphs), text-to-SQL vs knowledge-graph accuracy, GRADES-NDA at SIGMOD 2024",
     "date": "2023-11",
     "url": "https://arxiv.org/abs/2311.07509",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Ganz and Perigaud, dbt Labs (sells a semantic layer), semantic layer vs text-to-SQL 2026",
     "date": "2026-04-07",
     "url": "https://docs.getdbt.com/blog/semantic-layer-vs-text-to-sql-2026",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "reporting",
    "dashboards",
    "analytics"
   ],
   "bench_notes": []
  },
  {
   "slug": "buy-dont-build",
   "title": "Buy, don't build: RevOps configures, engineering builds",
   "summary": "Too early to call: AI made code cheap to write but not cheap to own, and the build savings people post leave out upkeep, so whether building now beats buying is unmeasured; the middle that is forming is to buy the system of record and build the logic.",
   "verdict": "too-early",
   "verdict_label": "Too early to call",
   "because": "evidence-splits",
   "because_label": "the evidence splits",
   "verdict_as_of": "2026-11-09",
   "last_checked": "2026-11-09",
   "domain": "systems and data",
   "altitude": "L2 Operating model",
   "origin": "Salesforce Well-Architected puts built-in services first, then AppExchange, then low-code, then custom code, and asks decision records to show the near- and long-term cost of build against buy. The admin ethos \"clicks, not code\" traces to a 2008 Salesforce Developers blog post. The RevOps half, that RevOps configures and doesn't engineer, has no traceable origin.",
   "rationed": "Engineering time. Code was slow and costly for a RevOps team to write and risky to own, so the team rented software and configured it.",
   "challenge_claim": "AI coding tools let RevOps, or a GTM engineer, build in days what it used to buy, at a fraction of the license cost. The strong version says GTM engineers replace RevOps.",
   "blindspot": "Build-savings posts compare a license with the cost of writing the code, not the cost of owning it: endpoints get deprecated, vendors change terms, and someone has to keep it running. The buy side misses that some logic, such as which writes are allowed and on what evidence, was never a good fit for a rented tool.",
   "logic": {
    "prior_premises": [
     {
      "premise": "Writing and testing software is slow and costly for a RevOps team.",
      "now": "changed",
      "kind": "causal"
     },
     {
      "premise": "Whoever builds a system owns its upkeep: breakages, API changes, the person who leaves.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "A vendor spreads upkeep across its customers, so renting a system of record costs less than owning one.",
      "now": "holds",
      "kind": "causal"
     }
    ],
    "challenge_premises": [
     {
      "premise": "AI makes code cheap enough that a team can build what it used to buy.",
      "now": "holds",
      "kind": "causal"
     },
     {
      "premise": "A build's total cost, upkeep included, comes in under the license it replaces.",
      "now": "unproven",
      "kind": "causal"
     },
     {
      "premise": "GTM engineers who build in code replace the RevOps role.",
      "now": "fails",
      "kind": "prevalence"
     }
    ],
    "deciding_premise": "A build's total cost, upkeep included, comes in under the license it replaces.",
    "deciding_side": "challenge",
    "what_follows": "The cost of writing code changed; the cost of owning it didn't. Whether a build now beats a license depends on upkeep, which the savings claims leave out, so the deciding premise is unproven. The middle practitioners are converging on, buy the system of record and build the logic, follows from the premises that do hold."
   },
   "write_up": [
    {
     "heading": "Where it came from",
     "text": "Salesforce's architecture guidance puts built-in features first, then the app marketplace, then low-code tools, and custom code last. It asks teams to record the near- and long-term cost of building against buying. The admin ethos, \"clicks, not code\", goes back to a 2008 Salesforce developer post. The RevOps half of the prior, that RevOps configures and engineering builds, has no author anyone has found. It was simply how the work got split."
    },
    {
     "heading": "What people say now",
     "text": "The build side has real wins. Joshua Schoen, a GTM engineer at Sheltra, on his own company's spend: \"We moved all of our enrichment into scripts we vibe-coded that now run a pipeline through our CRM,\" taking the monthly bill from about $10,000 to $500. What it costs to keep those scripts running isn't in the post.\n\nThe pushback is about owning it, since whoever builds the system now maintains it. Garrett Wolfe, a freelance GTM engineer who builds custom sales tools for clients, moved his own work into code and wrote: \"Apollo recently deprecated a people-search endpoint and I had no idea.\" The strong version of the claim, that GTM engineers replace RevOps, doesn't survive the job data. Kyle Poyar counted one GTM-engineer posting for every 14 RevOps postings in 2025, and an analysis of 1,000 postings concluded the two jobs \"are essentially the same.\""
    },
    {
     "heading": "Does the logic hold",
     "text": "The prior rested on three ideas. Writing software was slow and costly for a RevOps team. Whoever builds a system owns its upkeep. And a vendor spreads that upkeep across its customers, so renting a system of record costs less than owning one.\n\nAI changed the first. Code is now cheap to write. It didn't touch the second or the third. The challenge needs a build's total cost, upkeep included, to come in under the license it replaces. The savings posts compare a license with the cost of writing, and leave upkeep out. That premise isn't false; it's unmeasured, so the conclusion doesn't follow yet in either direction.\n\nWhat does follow from the premises that hold is the middle practitioners are converging on. An agency working across dozens of clients states it as \"buy the interface, build the logic.\" Ryan Gunn, a HubSpot marketing consultant, gives the reason: \"Unlike a CRM, Clay is not a system of record. If it breaks, you haven't corrupted your customer data.\""
    },
    {
     "heading": "What both sides miss",
     "text": "In my own setup, the line fell where neither slogan puts it. The data sources stayed rented. The parts that decide what gets written to the CRM, and on what evidence, I built: one writer for a contested field, an identity check before any write about a person, a tested rules module for the verdicts. Those encode judgment, and no vendor sells mine. The rented side wasn't free of upkeep either. Tooling I depend on cut old versions, deprecated a mode and changed its price units, each of which broke something I then had to fix."
    }
   ],
   "open_question": "What did your last in-house GTM build cost to keep running, not to write?",
   "sources": [
    {
     "side": "update",
     "who": "Joshua Schoen, GTM engineer at Sheltra (own company's spend)",
     "date": "2026-08-26",
     "url": "https://www.linkedin.com/posts/joshua-schoen_revops-activity-7498401633031188481-e-kZ",
     "standing": "position"
    },
    {
     "side": "defend",
     "who": "Garrett Wolfe, freelance GTM engineer (builds custom sales tools for clients)",
     "date": "2026-05-19",
     "url": "https://garrettawolfe.substack.com/p/can-claude-code-replace-clay-whats",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Gaizel Jaca (agency, 80+ clients)",
     "date": "2026-07-07",
     "url": "https://www.linkedin.com/posts/gaizeljaca_2026-gt-gtm-picks-wheel-activity-7480244430025936896-jP3f",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Ryan Gunn, HubSpot marketing consultant",
     "date": "2026-03-12",
     "url": "https://www.linkedin.com/posts/ryanagunn_you-shouldnt-vibe-code-a-crmbut-you-might-activity-7437919842265026560-_umI",
     "standing": "position"
    },
    {
     "side": "nuance",
     "who": "Kyle Poyar, Growth Unhinged, Do you need a GTM engineer?",
     "date": "2025-09-10",
     "url": "https://www.growthunhinged.com/p/do-you-need-a-gtm-engineer",
     "standing": "position"
    }
   ],
   "verdict_history": [],
   "tags": [
    "build vs buy",
    "gtm engineering",
    "tech stack",
    "maintenance"
   ],
   "bench_notes": []
  }
 ]
}
