Every marketing team I’ve ever encountered thinks in funnels. Awareness at the top, consideration in the middle, conversion at the bottom. It’s clean. It’s intuitive. It fits on a slide. And it’s almost entirely wrong.
Funnels are linear. Stuff goes in the top, some of it leaks out along the way, and whatever survives reaches the bottom. The fundamental assumption is loss. You accept that most of what enters will never convert, and you optimize by widening the top or plugging leaks in the middle. This is how most people think about marketing. It’s also how most people think about hiring, operations, and growth in general. Linear. Sequential. Lossy.
I think in flywheels. And flywheels are a fundamentally different machine.
The Mechanics of Compounding
A flywheel is a system where each component’s output becomes another component’s input. There’s no top and no bottom. There’s only motion, and the more motion there is, the harder the system is to stop. Amazon figured this out early: lower prices drive more customers, more customers drive more sellers, more sellers drive more selection, more selection drives lower prices. Each piece feeds the next. The system accelerates itself.
When I build a marketing system, I don’t think about stages. I think about loops. Here’s a concrete example, without getting too deep into the industry specifics. We run campaigns that generate leads. Those leads become clients. Happy clients leave reviews. Reviews improve our local search rankings and conversion rates. Better rankings generate more leads. The output of the system feeds back into the input of the system. That’s a flywheel.
But here’s where it gets interesting: most people stop at the obvious loop. They see the reviews-to-rankings connection and think they’ve built a flywheel. They haven’t. They’ve built a single loop. A real flywheel has multiple interlocking loops, and the art is in finding the connections that aren’t obvious.
Those same happy clients refer other families. Referrals convert at three to four times the rate of cold leads, which means lower cost per acquisition, which means more budget for the campaigns that started the whole cycle. That’s a second loop feeding the first. The data from those campaigns tells us which messages resonate, which we feed back into the campaign creative, which improves performance. Third loop. The operational improvements that make clients happy in the first place, shorter wait times, better communication, more consistent service, those get documented and systematized, which makes it easier to maintain quality as we scale, which keeps the review engine running. Fourth loop.
Now you have a system that compounds. Every dollar, every hour, every decision feeds multiple loops simultaneously. This is what funnels can never do. Funnels are one-directional. Flywheels are omnidirectional.
Why People Default to Funnels
Funnels are popular because they’re simple and they give people a sense of control. You can point to a stage and say “we need to fix this.” You can measure drop-off rates between stages and feel like you understand what’s happening. Funnels satisfy the human need for narrative, there’s a beginning, a middle, and an end.
Flywheels are harder to think about. They don’t have a clear starting point. The metrics are interconnected in ways that make attribution messy. When everything feeds everything, it’s harder to point at one thing and say “that’s what’s working.” This makes executives nervous. It makes slide decks complicated. It makes quarterly reviews ambiguous.
I don’t care. I’d rather have a system that’s hard to explain and actually works than one that’s easy to explain and slowly bleeds out.
Flywheels Beyond Marketing
The flywheel principle applies to everything I build, not just marketing systems.
Team building is a flywheel. You hire good people, they do good work, good work attracts other good people who want to work alongside them, more good people means more good work. The compounding effect of a strong team is nonlinear. One great hire doesn’t add linearly to your output, they multiply it, because they raise the standard for everyone around them and they attract the next great hire.
Knowledge management is a flywheel. Every problem you solve and document makes the next similar problem faster to solve. The faster you solve problems, the more time you have to document solutions. The more solutions you document, the more institutional knowledge you build. The more institutional knowledge you have, the less dependent you are on any single person. This is why I’m obsessed with systems and documentation. Not because I love process for its own sake, but because documented process compounds.
Even personal development works this way. The more you build, the better your judgment gets. Better judgment leads to better decisions about what to build next. Better projects attract better collaborators. Better collaborators teach you things that improve your judgment further. I’ve lived this loop since I was fourteen, and the speed of it keeps increasing.
The Discipline of the Flywheel
Here’s what most people miss: flywheels require patience and discipline that funnels don’t. A funnel gives you immediate, measurable results. Pour leads in the top, count conversions at the bottom, calculate ROI, repeat. A flywheel takes time to spin up. The first few rotations are slow. The connections between loops haven’t fully formed. The compounding hasn’t kicked in yet.
This is where most people give up. They try the flywheel approach for a quarter, don’t see hockey-stick results, and retreat to the funnel because at least the funnel gives them numbers they can put in a report. They trade long-term compounding for short-term legibility.
I’ve learned to resist this. When I design a system, I think about what it looks like at month one and what it looks like at month twelve. If the month-twelve version isn’t dramatically better than the month-one version, not because of more spend or more effort, but because of compounding, then I haven’t built a flywheel. I’ve built a funnel with extra steps.
The question I ask about every system I build: does this get better on its own over time, or does it only get better when I push harder? If it’s the latter, it’s a funnel. If it’s the former, it’s a flywheel. And flywheels are the only systems worth building.
Every company I’ve worked at, every project I’ve launched, every piece of infrastructure I’ve designed, I evaluate it against this standard. Build the thing that feeds itself. Build the thing that compounds. Build the flywheel, not the funnel.
