The funnel is not a lie — but it is a ceiling. When you build your company around funnel thinking, you build a business that requires constant feeding. Every quarter you need more leads, more budget, more reps. The flywheel builds a business that feeds itself.
Why the Funnel Dominates (and Why That's a Problem)
Most companies default to funnel thinking because it is measurable, linear, and fits neatly into attribution models. You track leads, MQLs, SQLs, opportunities, and close rates. Marketing feeds sales. Sales feeds revenue. Everyone has a number. This feels like sophistication — and for a while, it works.
But here is what funnel thinking structurally misses: it treats customers as outcomes, not inputs. Once someone converts, the funnel's job is done. There is no mechanism for compounding. Every growth dollar you spend this quarter has zero effect on what your growth costs next quarter.
The data is damning. SaaS companies with product-led growth — the clearest proxy for flywheel mechanics — traded at a 1.5x revenue multiple premium over sales-led companies in OpenView's 2024 SaaS benchmarks. Dropbox's viral referral loop drove user acquisition at under $1 per user while Salesforce was spending $5,000–$7,000 to close a single enterprise seat. Same addressable market, completely different unit economics — and the difference was not product quality, it was architecture.
How Flywheels Actually Compound
A flywheel is a self-reinforcing loop where each rotation reduces the cost or increases the value of the next. The classic example is Amazon: more customers bring more third-party sellers, which deepens selection, which lowers prices, which attracts more customers. Each component actively strengthens the others. Critically, momentum accumulates — the flywheel gets cheaper to spin over time, not more expensive.
Slack is a tighter case study for B2B. Individual users inside a company invited colleagues. Colleagues invited their teams. Teams spread to adjacent departments. Each new seat inside an existing account reduced effective CAC to near zero because the distribution was happening inside the product itself. By the time Salesforce acquired Slack for $27.7B in 2021, the business had grown to $902M ARR with a sales efficiency ratio that made traditional SaaS look obsolete.
LinkedIn's flywheel tells the same story at a longer time horizon. In 2004 they had 1M users and weak monetization. The loop that broke them through was user-generated content: members posted career updates, which drove organic search traffic, which brought in new members, who posted more content. By 2016, they had 400M users and sold to Microsoft for $26.2B — a return of more than 85x on the total venture capital invested. That compound growth was not bought. It was architected.
Five Signs You Need a Flywheel, Not a Funnel
- •Your CAC is growing year over year while your product has not meaningfully changed — classic symptom of funnel economics deteriorating as the cheapest acquisition channels saturate
- •More than 40% of new customers come from referrals or organic search, yet you are still allocating 70%+ of marketing budget to paid acquisition — you already have a flywheel loop forming, you are just not investing in it
- •Retention is above 90% NRR but growth is still primarily top-of-funnel dependent — your product is sticky but not viral, which means you are sitting on untapped expansion and referral loops
- •Existing customers have low expansion rates and are not inviting colleagues or generating visible social proof — your product may solve a problem without creating a network, which is a strategic design question worth revisiting
- •Your most formidable competitors — Notion, Figma, Atlassian, Canva, HubSpot — all reached $100M+ ARR with essentially no traditional outbound sales function. That is not luck. It is flywheel architecture at the product level.
How to Diagnose Your Own Business
I have looked at hundreds of companies across 65+ investments. The single most useful diagnostic I use is this: draw a circle and try to write a loop where each step makes the previous step more valuable or cheaper. If you cannot close the loop without writing "and then we do more marketing", you have a funnel, not a flywheel.
The second diagnostic is cohort analysis. Pull your last 24 months of new customer cohorts and ask: did customers from earlier cohorts generate more organic referrals and less-assisted acquisition than newer cohorts? If yes, your flywheel is working. If referral rates are flat or declining as your customer base grows, you are scaling a funnel — and funnel economics erode under scale.
Building a flywheel is not free. It requires product investment, patience, and often a willingness to give value before you extract it. Dropbox gave away 500MB of storage for referrals. LinkedIn let anyone post content for free for years. The ROI is asymmetric — the early quarters look expensive, the later ones look like magic. In my experience, the founders who refuse to make that investment are the ones who spend their Series B trying to buy growth they should have built into the product at Series A.
The funnel tells you how many customers you acquired last quarter. The flywheel tells you whether each customer makes the next one cheaper to acquire. Only one of those models builds a business worth owning at scale.
Stay current with VC and startup trends at Value Add VC. Originally published in the Trace Cohen newsletter.