Flywheel

For a long time growth was understood as something you push. In a company where ad spend brings sign-ups and cutting the spend takes them away again, earning more next quarter means pushing harder than this quarter. Budget meetings opened with the same question every time: if we stop advertising this month, what percentage of sign-ups disappears? When the answer was close to all of them, that company's growth simply was the push.

The problem with that shape gets clearer as the numbers get larger. The money it took to go from ten thousand users to twenty thousand costs the same again to go from a hundred thousand to two hundred thousand, and when a competitor bids on the same channel the unit price rises. That's how a company ends up spending more every year and growing at exactly the same rate.

A flywheel inverts the question. It asks first whether the effort spent this time can be arranged to make the next time easier. A heavy wheel takes everything you have to start turning, but once it carries speed the same push moves it faster — the term asks whether a business can be given that shape. If choosing where to compete is what strategy does, the flywheel is what explains why a given choice gets stronger the longer it runs.

Three to five stages are chained into a loop, and the output of the last stage returns as the input of the first. In a document collaboration tool it runs like this: the more documents a team accumulates, the more useful search becomes; once search is useful, the next department puts its documents there too; more documents make search more useful still. Because the loop is closed, every turn starts from a higher point than the last.

Four checks tell you whether a loop actually turns.

1. Is it closed — does the last stage really connect back to the first? More ads bring more sign-ups, more sign-ups bring more revenue, more revenue buys more ads: that looks circular, but the added users don't bring the next users, so it's just a bigger budget.
2. Does each link have a real cause — every arrow should be explainable in one sentence. "More users make the product better" is missing its explanation. "More users mean more documents in the index, so a new hire finds an answer faster" has one.
3. How long is one turn — shorter cycles compound sooner. A product where someone invites a colleague within a week and one where it takes six months feel nothing alike, even with an identical diagram.
4. Where is the weakest link — the slowest stage sets the speed of the whole loop. If documents pile up but search results are poor, nothing reaches the next stage, and that's where the budget goes.

The judgment that comes out of those checks is what the concept is actually for: when deciding where to spend effort, it forces the weakest stage to be looked at first.

It appears most often in strategy documents and fundraising conversations. When an investor asks whether the company gets stronger as it gets bigger, this diagram is the answer, and without a loop a growth plan is just a budget plan. It also shapes quarterly goals. When several initiatives are on the table, ranking them by which one touches the weak stage turns a long list of strategic priorities into a short one.

A loop that has been turning for a while becomes a competitive advantage. A rival can rebuild a feature in a month, but three years of accumulated documents and a whole team's fluency with how that search behaves cannot be rebuilt on any schedule. The time spent turning the loop is itself the asset that's hard to catch.

There's a direction to watch. Loops run backward too. Fewer documents make search less useful, which sends people to another tool, which leaves fewer documents still. A single quarter of elevated churn pulls the next stage down with it, so a company that has drawn its loop reads the numbers for each stage in sequence rather than one at a time.

"Which stage of the flywheel does this feature turn?"

Asking whether the initiative sits inside the loop or is a one-off outside it.

"The third stage is blocked, so pushing the front of the loop does nothing right now."

Arguing to fix the weakest link first.

"Our flywheel takes about two months to complete one turn."

The loop exists, but the effect arrives slowly.

Grace, Pagedeck's chief executive, uses this diagram in investor meetings. A team accumulates documents, search becomes useful, search pulls in the next department, more departments trigger a company-wide paid conversion, and that revenue funds better search. Four stages.

One quarter the diagram changed an actual decision. Ravi, the growth marketer, proposed a campaign to raise new sign-ups, and Mia brought numbers showing how rarely a team actually invited a neighboring department. With the second stage blocked, pushing the first one wouldn't move the loop, so that quarter's budget went into fixing the invite flow instead of advertising. When the invite rate rose the next quarter, the same ad spend brought in more people than before.

The most common mistake is believing that drawing a circle creates a flywheel. Round diagrams joined by arrows show up in so many decks that the picture starts to feel like the thing. A loop requires causation between the stages; stages connected without it are just a sequence diagram. A quarter spent in that state goes into pushing the first stage over and over, with no way to find inside the diagram why the numbers aren't moving. Checking is easy: point at one arrow, ask why it happens, and if the answer comes back as a number it's a loop, while a plausible-sounding sentence means it isn't one yet.

The other misreading is that once it turns, it no longer needs pushing. The phrase "accelerates on its own" invites that. In practice the first turns take the most force, and during that stretch the only options are pushing from outside through marketing or sales. Skip the unglamorous work of recruiting early users by hand and you can burn a lot of time without ever finding out whether the loop holds at all.