Product-market fit has a ceiling
The fit you found is real. It just doesn't cover your whole market.
There's a story founders tell about product-market fit. You search for it, you suffer for it, and one day you cross it. Customers pull. Revenue compounds. The hard part is over, and now the job is pouring fuel on the fire. The story is half true. What you crossed was fit with one segment, reached through one channel, sold one way. And that segment is smaller than your market.
Rethinking the model
The people who study this at scale have stopped treating fit as a single event.
First Round Capital describes four levels of product-market fit, from nascent to extreme, and notes that most companies stall at the middle levels, where there's real repeatability but the marginal customer still takes effort to win.
Sequoia walks founders through three distinct archetypes of fit in its Arc program rather than a yes-or-no diagnosis. Roger Norton makes the point with Tesla's master plan: the roadster, the family sedan, and the mass-market car were three separate product-market fits, each with its own customer, pursued in sequence. Fit was something the company had to find again at every stage.
The revenue benchmarks tell the same story from the other direction. Bessemer's T2D3 model, the canonical growth path for venture-backed software, asks you to triple from one million to three, then triple again to nine. Nobody triples to nine by doing more of what got them to three. The math forbids it.
The cohort problem
Your first customers are a real cohort. They are also a finite one.
The customers who got you to your first few million are the ones whose problem your product fits perfectly, who discovered you through the channel you happened to build, and who responded to the exact story you happened to tell.
That's a real cohort. It's also finite, and it's almost always smaller than the founder believes, because the founder is usually a member of it. You built for people like you, found them, and converted them at a rate that felt like destiny.
Then the cohort starts to empty. Acquisition gets a little more expensive each quarter. The campaigns that printed money produce a little less. The instinct at this point is to spend harder, because spending harder is what worked. But the segment that was a perfect match for your message is running out of people in it, and money can't manufacture more of them.
The segment that was a perfect match for your message is running out of people in it, and money can't manufacture more of them.
A real example
A channel ceiling looks like a performance problem until you see it clearly.
I lived a version of this. Before going fractional, I founded and scaled Skreened, a custom-apparel company.
Skreened had a channel that worked, and for years the answer to growth was more of that channel. Somewhere around eight million in revenue, the numbers stopped cooperating. The honest read was that the direct channel had a ceiling, and the next phase of growth lived on platforms we had been avoiding: Amazon, Etsy, eBay, Google Shopping.
We avoided them for reasons that sounded responsible. Listing fees ate margin. Returns worked differently on every platform. Part of the customer relationship would be handed over. Every one of those concerns was legitimate, and the conversation about cannibalizing direct sales was a real conversation, held more than once.
What wasn't fully appreciated going in was that the channel decision was the easy half. The hard half came after, and it was a marketing problem. A buyer who finds you on Amazon discovered you differently, trusts differently, and decides differently than a buyer who lands on your own site. We had written product titles and descriptions for one context. They had to be rewritten for several, each with its own logic of discovery, because the words that worked on the Skreened site were close to invisible everywhere else.
The product didn't change. The customer, in the broadest sense, didn't change. But who we were talking to, where we found them, and what those buyers needed to hear changed completely. That's a new fit, found on the far side of a revenue threshold, and it cost real effort to find it.
Binary vs. gradient
Treating fit as a binary leads you to spend a year solving the wrong problem.
When fit is either present or absent, the early signs of segment exhaustion read as a performance problem. So you spend a year tuning campaigns aimed at a cohort that's already mostly converted.
Treating fit as a gradient changes the question. You can see the next fit coming before you're forced into it. The founders who navigate the three-million mark well aren't the ones with the best original playbook. They're the ones who notice when the playbook's audience is running out.
Reading the signal
The signal shows up in your numbers long before it shows up in your gut.
The dashboard will tell you something is wrong without telling you what. How you notice the ceiling approaching is its own subject, and that's where I go next.
Part one of four
Where this goes next.
This is the first piece in a short series on what happens after the first product-market fit runs its course. Next: why your dashboard can tell you something is wrong without telling you what.
It's also one part of a larger picture: how demand actually gets generated for a considered, high-ticket purchase.
Questions this raises
Not sure whether you're out of segment or out of message?
That's the conversation. A short call tells us both whether your next move is a positioning problem, a channel problem, or neither.
Start a conversation