Stewart Butterfield did not build Slack to sell to teams. He built it to save a dying game.
In 2012, his gaming company Glitch was shutting down. The game had failed to find an audience despite years of work and real investment. But Butterfield’s team had built an internal communication tool during that project, something to keep a distributed team coordinated. That tool, which they were using themselves every day, felt like it might be worth something. So they pivoted.
The first real test of Slack was with a single team: their own. Then they opened it up to a few others. Tiny Speck (the company behind Glitch) became Slack’s first genuine external pilot customer, and everything about what Slack became, its channel structure, its notification logic, its emoji reactions, was shaped by the needs of that original user: a small, scrappy software team trying to get work done.
Here is the thing nobody talks about. That first customer nearly locked them in forever.
The Trap of Premature Product-Market Fit
When you build for one user, you solve their problems. That feels like success. The product improves, the user is happy, you ship features, you iterate. It looks exactly like traction. What you’re actually doing, in many cases, is building a very polished solution to a very specific problem that belongs to one kind of person.
For Slack, that first user was a technical team. Engineers. People who were comfortable with keyboard shortcuts, who wanted integrations with GitHub and Jira, who thought nothing of learning a new tool. The early Slack reflected this. It was powerful, a little dense, and built on assumptions that made total sense if you were a developer and almost none if you were, say, a marketing team at a mid-size insurance company.
As Slack grew, they had to make an active, deliberate decision to expand beyond that original user’s gravitational pull. They had to add features that technical users didn’t need and sometimes actively disliked. They had to simplify onboarding. They had to rethink what “a workspace” meant when the people joining weren’t already comfortable with the idea of channels as a metaphor.
The first customer had given them a product. The market they actually wanted required a different one.
Airbnb’s version of this story is less polished, more painful. Brian Chesky and Joe Gebbia’s first paying guests stayed in their apartment in San Francisco in 2007, sleeping on air mattresses. These were conference attendees who couldn’t find a hotel room. The specific problem Airbnb solved for those first users was: I need a cheap place to sleep in a city during a sold-out event.
For a while, Airbnb was essentially a conference overflow service. They literally targeted events. They hustled their way through SXSW and the Democratic National Convention trying to sign up hosts and guests in cities where hotels were full. It worked, and it also defined what “Airbnb” meant to its earliest users: a budget option, a backup plan, something slightly weird that you tried because you had no other choice.
That perception nearly killed them. Investors passed repeatedly. The product felt too niche, too dependent on specific circumstances. The first customer, in this case, had trained the founders to think small. It took Y Combinator, Paul Graham’s famous “go to your users” push, and a lot of time before Airbnb stopped optimizing for the conference-overflow use case and started building for the traveler who chose them first.
Flickr’s story is different in a way that’s instructive. Stewart Butterfield, remarkably, is involved here too. Flickr grew out of a feature inside a massively multiplayer online game called Game Neverending. The photo-sharing piece was a side mechanic. Their first users were game players who wanted to share screenshots and images with each other.
Those first users cared about community. They cared about tagging, commenting, social connection around images. That was the culture Flickr inherited. It made the early product genuinely great in a specific way. It also meant Flickr was optimized for enthusiasts, hobbyists, people who thought of photography as a social act. When smartphones arrived and the market shifted toward casual users who just wanted somewhere to put their photos, Flickr wasn’t built for them. The first customer had shaped a product that the next ten million customers found slightly alien.
What the First Yes Actually Commits You To
The pattern across these three companies is the same. The first customer is not a market signal. They are a design constraint. Say yes to them, as you should, and you are now solving their problem. Everything you learn, you learn from them. Your intuitions about what “the user” wants come from watching them. Your feature roadmap reflects their friction points.
This is how it has to work. You cannot build a product without a user. The alternative, building for an imagined average user before you have real ones, is worse. But the founders who navigate early stages well are the ones who hold this awareness: the first customer is a starting point, not a destination.
The mistake, and it is a common one, is treating early traction as confirmation that you have found your market. Finding the first hundred customers is hard enough; it is tempting to conclude that whoever showed up must be who you’re for. Sometimes they are. Often they are a narrow proxy for something much larger, or much different.
Slack eventually served everyone from engineering teams to HR departments to nonprofits. But it required years of deliberate work to stop being a tool designed for the first users and become something broader. The same is true of Airbnb, which had to shed the conference-overflow identity before it could become a default travel choice. Flickr, which never quite made that transition cleanly, provides the cautionary version of the story.
The practical implication is not complicated, but it requires honesty most founders resist. As you build for your first customer, you should also be asking what kind of company you are not building. What assumptions are you encoding because this particular user has this particular problem? Which of those assumptions will help you scale, and which will become walls?
Say yes to the first customer. Learn everything they will teach you. Just don’t let them write your entire future. Listening to the wrong customer at the wrong moment is a real way to lose. And the first customer, by definition, is someone whose needs you understood before you fully understood your own company.
Butterfield understood this, eventually, twice. Flickr and Slack both came from pivots, both started with a single concrete user, and both required their founders to look beyond that user to find the larger thing. That he did it twice, from game mechanics to photo sharing, from game communication tools to enterprise messaging, is either impressive or statistically suspicious. Either way, the lesson holds.
