Keynote Briefing · Impact Summit
The orthodoxy that captured growth — and the architecture that replaces it.
Funnelism is not a model — it is an -ism. A belief system. The doctrine that linear, acquisition-centric pipelines are the natural architecture of growth. Its scripture was written in 1898 by E. St. Elmo Lewis, hardened into operational dogma by Aaron Ross in 2011, and enforced for two decades by the unholy trinity of Salesforce stage semantics, HubSpot inbound playbooks, and VC unit economics (LTV/CAC, CAC payback, the Rule of 40) that demanded clean stage accounting to fund growth.
The funnel won SaaS not because it described reality, but because it aligned with how SaaS was funded, measured, and managed. Pipeline review became the central management ritual of the modern revenue org. The SDR/AE/CSM split became the human architecture of Funnelism. The weekly forecast call became its weekly mass.
There has always been a heretic tradition. Frederick Reichheld proved in 1990 that a 5% defection reduction produced 25–125% profit increases — yet loyalty economics never displaced acquisition economics because it could not be turned into a weekly meeting. Ben Chestnut called the funnel "a meat grinder" in 2013. Wayne Peterson named the religion itself in 2015: "I am not a Funnelist. I don't worship at the altar of Funnelism." Brian Halligan publicly killed his own framework in 2018, pivoting HubSpot from funnel to flywheel. Andrew Chen and Brian Balfour at Reforge declared that growth loops, not funnels, describe how the fastest-growing products actually grow.
But every partial reform — PLG, Customer Success, Community-Led, ABM, RevOps, even the Flywheel — left funnel logic load-bearing underneath. They patched the model; they did not replace it.
The AI-Native era is where Funnelism finally breaks. Telemetry-native, loop-native, recurring-impact-native architecture is the successor. Bowtie, growth loops, and the impact graph compose into the same thing: a system where customer impact is the growth mechanism, not its consequence. The advantage compounds. The funnel does not.
The 1898 ur-text. E. St. Elmo Lewis (1872–1948), an American advertising advocate from Philadelphia, articulated the slogan "attract attention, maintain interest, create desire" in 1898. He added "get action" later. The four-stage formulation we know as AIDA was publicly cemented on January 6, 1910. HIGH
The attribution is partially disputed. Joseph Addison Richards (1893), Fred Macey (1899–1900), and Frank H. Dukesmith (1904) all published variants. Akinori Iwamoto in 2023 challenged the canonical Lewis-as-inventor story. The honest framing: Lewis didn't invent the funnel — he formalized a structure already drifting through 1890s advertising and gave it a teachable shape. MED
Arthur Sheldon (1911) extended the model by adding "Satisfaction" — meaning AIDAS — and explicitly framing it around repeat patronage. The retention idea was already inside the original literature. The market threw it away. HIGH
The migration from advertising to sales happened over six decades — through Strong (1925), Borden's marketing mix work, and the rise of the salesperson as the operational unit of B2B distribution. By the time the SaaS era began, the funnel was already so naturalized that it felt like physics.
The hardening moment: Aaron Ross, 2011. Predictable Revenue (subtitle: Turn Your Business Into a Sales Machine with the $100 Million Best Practices of Salesforce.com) operationalized the funnel into an enforced workflow. Ross's innovation — specialization of SDR (prospecting) and AE (closing) roles, "Cold Calling 2.0" via cold email referrals, and ~100 emails per SDR per week to executives asking for warm intros — required clean stage accounting to function. Salesforce stage semantics (Lead → Qualified → Opportunity → Closed-Won) became the spinal column of every SaaS revenue org built between 2011 and 2020. HIGH
This is the moment the funnel stopped being a metaphor and became operational dogma. It became the workflow. It became the comp plan. It became the org chart.
Four mutually reinforcing forces locked the funnel into place:
ARR, LTV/CAC, CAC payback, the Rule of 40, Burn Multiple, Magic Number — all four formulas that dominate VC diligence presuppose clean stage accounting. David Skok's 2010 SaaS Metrics 2.0 framework codified the LTV/CAC 3:1 ratio with CAC payback under 12 months. Bessemer's efficiency benchmarks reinforced it. You cannot prove unit economics to a VC partner without naming the stages, counting conversion, and projecting the funnel. HIGH
The weekly pipeline call became the central liturgy of the SaaS org. Forecast accuracy became the proxy for managerial competence. The architecture of the meeting — stages on the X axis, deal value on the Y, ownership in rows — determined how the company thought, hired, and paid people. MED
Funnel-friendly metrics outcompeted relationship-native metrics because they were countable, comparable, and forecastable. Brand affinity is real but not weeklyable. MQL-to-SQL conversion is. The funnel didn't win the contest of truth. It won the contest of measurability.
HubSpot, Marketo, Salesforce, and Outreach didn't merely support the funnel — they were the funnel, materialized in software. Every dropdown, every required field, every report template encoded the four-stage progression. Once 10,000 companies were running on the same shape of CRM, the funnel hardened into infrastructure. The shape of your CRM became the shape of your strategy.
The MQL-to-SQL handoff became the most fragile point in the modern revenue operation — two teams with different incentives, different tools, different vocabularies, and conflicting definitions of "qualified" attempting a precise transfer of context. The religion's central sacrament is also its weakest link. HIGH
What did each one solve, and where did it leave funnel primacy intact?
| Reform | Solved | Funnel logic still load-bearing? |
|---|---|---|
| PLG Wes Bush / OpenView, 2016 |
Self-serve trial/freemium replaces SDR-first acquisition; product delivers value before purchase. | Yes. Still measures Visitor → Free Signup → Activated → Paid. Compressed the stages; did not eliminate stage logic. |
| Customer Success Mehta, Murphy, Gainsight, 2016 |
Made retention and expansion a first-class function. Stages: onboard → adopt → expand → renew. | Yes — perhaps the most pernicious. Added a second funnel downstream. "Land-and-expand" is two linear pipelines, not a loop. |
| Community-Led David Spinks / CMX, 2019–2020 |
Surfaced that strong communities show better retention, lower CAC, higher NPS — and this is structural, not fluff. | Partial escape. Genuinely loop-shaped, but in practice most companies retrofitted "community" as a top-of-funnel awareness tactic. |
| Growth Hacking Sean Ellis / McClure AARRR, 2007–2010 |
Brought experimental rigor and product instrumentation to acquisition. Built canonical Dropbox referral loop. | Funnel-native by construction. AARRR is explicitly called the "Pirate Funnel." The vocabulary and linearity survived. |
| ABM Engagio/Demandbase, 2014–2015 |
Inverted the funnel — start with target accounts, work outward. Replaced spray-and-pray. | Yes. ABM is the funnel flipped, not the funnel replaced. The 13 average buying-committee stakeholders still progress through stages. |
| RevOps ~2018–2019 |
Unified Marketing/Sales/CS Ops into one function accountable to revenue. Created the "revenue funnel." | No escape at all. RevOps operationalized the funnel across the full lifecycle. The most sophisticated administrative apparatus, not a replacement. |
The pattern: every reform either (a) compressed or extended the funnel, (b) ran a parallel funnel downstream, or (c) inverted the funnel's direction. None replaced the primitive. The primitive is linear stage progression. Until that goes, Funnelism survives in its tooling.
The most developed post-funnel architectures share three properties: compounding (output reinvested as input), multi-loop composability (acquisition, engagement, monetization loops feed each other), and impact-as-mechanism (customer outcomes are the growth engine, not its consequence).
The structural advantage. AI-Native companies are born telemetry-native — every action, retention signal, and usage event flows back into the product in milliseconds. They are also loop-native by necessity: a coding tool that doesn't make developers more productive doesn't get adopted, and adoption is the growth engine.
This is what loop-native looks like. The funnel didn't grow Cursor. Cursor's loops did.
Most AI-native companies are structurally closer to loop architecture and culturally still running funnel tactics dressed in AI clothes:
The hypothesis to defend on stage: the AI-Natives who win the next decade are the ones who architect for recurring impact and compounding loops. The ones who ride the acquisition wave will compound less than the ones who compound outcomes, even at lower top-of-funnel volume.
Where is funnel logic still load-bearing? Honestly:
The honest synthesis: the funnel is a legitimate primitive for managing single deals through a procurement process. It is an illegitimate primitive for modeling how a business grows.
Funnelism is the category error of mistaking the first for the second. The right architecture composes both: loops at the business layer, funnels at the deal layer. Most current GTM stacks invert this.
| Funnel era | Impact era |
|---|---|
| MQL / SQL | Engaged user / qualified usage signal |
| Lead | Active user with measurable impact event |
| Pipeline coverage | Loop coverage (active loops × loop velocity) |
| CAC | CAC + advocacy-adjusted CAC (a-factor) |
| LTV | NRR-compounded LTV (open-ended, not capped) |
| LTV / CAC | NRR × loop efficiency |
| Stage conversion rate | Loop velocity & loop closure rate |
| Quota | Impact velocity (rate of outcomes produced) |
| Forecast | Compounding projection (NRR × new logo loops × time) |
| Pipeline review | Loop review |
| SDR / AE / CSM split | Impact pod (cross-functional, owns a loop end-to-end) |
| ARR | Recurring Impact Revenue (RIR) — ARR weighted by realized outcomes |
| Win / Loss | Loop closed / loop stalled |
| Funnel optimization | Friction removal from compounding loops |
| Marketing → Sales → CS handoff | Continuous lifecycle ownership |
The vocabulary shift is not cosmetic. Every renamed primitive is a different unit of management. A manager who runs a "loop review" instead of a "pipeline review" hires differently, comps differently, and reports to the board differently. The vocabulary is the lever.
| Year | Event | Significance |
|---|---|---|
| 1898 | E. St. Elmo Lewis articulates "attract attention, maintain interest, create desire" | The ur-text. The funnel's founding scripture. |
| 1910 | Lewis publicly fixes the four-stage formula (adds "get action") | AIDA enters the language. |
| 1911 | Arthur Sheldon adds "Satisfaction" (AIDAS) | Retention is already in the original literature. The market ignores it. |
| 1921 | C. P. Russell coins the acronym "AIDA" | The funnel gets a brand. |
| 1925 | Strong's Psychology of Selling attributes the model to Lewis | Canonization. |
| 1990 | Reichheld & Sasser publish "Zero Defections" in HBR | First hard data that retention dominates acquisition. Ignored. |
| 1996 | Reichheld publishes The Loyalty Effect | Book-length argument. Still ignored at the operating level. |
| 2003 | Reichheld publishes NPS in HBR | A retention-era metric travels — but as supplement, not replacement. |
| 2007 | Dave McClure presents AARRR at Startup Metrics for Pirates | The "Pirate Funnel" — growth hacking born funnel-native. |
| 2010 | Sean Ellis coins "growth hacking" / Skok publishes SaaS Metrics 2.0 with LTV/CAC 3:1 | The funnel becomes the language of capital. |
| 2011 | Aaron Ross publishes Predictable Revenue | The funnel hardens into operational dogma. SDR/AE split enforced. |
| 2013 | Ben Chestnut publishes "Why I Hate Funnels" | First widely-circulated verbal demolition. |
| 2015 | Wayne Peterson names "Funnelism" as an ideology (Mar 23) | The -ism is christened. |
| 2016 | OpenView coins "Product-Led Growth" / Reforge founded | Partial reform begins. PLG and growth loops formalized. |
| 2018 | Halligan's INBOUND keynote — funnel to flywheel | The largest funnel-tooling company publicly abandons the funnel. |
| 2019–20 | CMX consolidated by Bevy; Community-Led Growth crystallizes | Community as growth substrate, not marketing tactic. |
| 2023 | WbD releases updated Bowtie + Revenue Architecture textbook | Bowtie as operating design for recurring-revenue growth. |
| 2025–26 | Cursor / Anysphere hits $2B ARR in ~3 years | Loop-native architecture beats every Funnel-era growth record. |
The funnel never described how customers behave. It described how managers manage. Funnelism's grip on SaaS was never about reality — it was about the weekly meeting. The reason loops keep losing to funnels, despite being demonstrably better at modeling growth, is that loops cannot be turned into a recurring 9 a.m. Monday agenda. Funnels can. Until the post-funnel architecture produces its own ritual — a "Loop Review" with a defined cadence, a defined artifact, and a defined accountability seat — Funnelism survives by default, not by argument.
The historical evidence supports this. Reichheld won the economic argument in 1990. Chestnut won the moral argument in 2013. Halligan won the strategic argument in 2018. Chen and Balfour won the technical argument over the following decade. The funnel won every year anyway — because none of the alternatives produced an equally legible weekly management ritual.
The implication for the Impact Summit keynote: the winning move is not a better critique of the funnel. It is the design of the Loop Review — a Monday-morning management ritual whose primitive units are loops, whose metrics are loop velocity and impact compounding, whose seat-of-accountability is the impact pod, and whose forecast is a compounding projection rather than a pipeline coverage ratio.
Whoever ships that ritual — and the dashboard, the cadence, the comp plan, and the board-report template that go with it — replaces Funnelism. Not by arguing against it. By making it less convenient than the alternative.
This is the SPICED + Bowtie + Revenue Architecture endgame. Not a competing methodology. The operating system of the post-funnel weekly meeting. The semantic layer that makes loops manageable — and therefore inevitable.