Sit through enough founder pitches and “flywheel” starts to sound less like a business mechanism and more like a word chosen for how it makes a slide feel. More customers mean more data, more data means a better product, a better product means more customers - the loop gets drawn with arrows curving back into itself, and the room nods, because the shape is satisfying even when nobody in it has asked whether each arrow is actually true.
A real flywheel is a specific, provable claim: this input measurably increases that output, which measurably increases the next one, and the cycle compounds faster than linear effort would. Most of what gets called a flywheel in an Indian founder deck is a hope dressed as a mechanism - a description of how growth could work in theory, presented with the confidence that usually accompanies something that’s been checked.
The tell is always the same: no number attached to any arrow
Ask a founder what percentage of their “flywheel” actually repeats - what share of month-one customers are still active in month six, what the referral rate really is, whether the “better product from more data” claim has ever been tested against a version of the product that didn’t have the extra data - and the diagram usually stalls. The word was describing an aspiration about the business, not a measured property of it. That’s not a crime; most early-stage companies don’t have the mechanism yet. The problem is presenting the hope and the measurement in the same confident tone, so the listener can’t tell which parts are real.
This matters more than it looks like it should, because the word changes what the team spends time on. A founder who’s told themselves the flywheel is already spinning stops asking the question that would actually tell them: what is our repeat rate, and is it getting better or worse each quarter. The metaphor becomes a reason not to check.
Retention is the unglamorous word that replaces it
The founders worth modelling skip the loop diagram and lead with the plain number - repeat purchase rate at ninety days, logo retention at twelve months, the referral percentage with a real denominator attached. It’s a less exciting sentence than “flywheel,” and it’s also the only version of the claim that can be wrong in a way anyone would notice, which is exactly why it’s trustworthy. A number that can fail is doing real work. A metaphor that can’t fail isn’t measuring anything - it’s just describing a shape.
What to do with the diagram instead of deleting it
The fix isn’t banning the word from every deck; the fix is making the diagram earn its place by attaching a real figure to each arrow before it ships to an investor or a customer. If an arrow can’t carry a number yet, label it as a hypothesis rather than a fact, and say so in the same sentence. A founder who can point to “we think this loop exists, and here’s the test running right now to confirm it” is further ahead, not behind, of one who’s drawn a confident circle and called it a growth strategy. The circle was never the asset. The retention number underneath it always was.