In August 2012, Stewart Butterfield sent an internal memo to his team at Tiny Speck. They were building a massively multiplayer online game called Glitch. The game had launched, struggled to find an audience, and was burning through investor money with no clear path to scale. The memo was called ‘We Don’t Sell Saddles Here,’ and it laid out a vision for the internal communication tool the team had built for themselves. Butterfield wasn’t pivoting away from failure. He was pivoting toward something he could see that almost nobody else could.
From the outside, this looked like a company folding. Tiny Speck had raised around $17 million to build a game. They were shutting the game down. The tech press covered it with the gentle sympathy reserved for startups that didn’t make it. A few outlets noted that the team was going to try to commercialize an internal chat tool, which sounded like the kind of face-saving pivot founders announce before quietly disappearing.
That internal chat tool became Slack.
The reason the Slack story still matters, years after it became startup canon, is not the happy ending. It’s the mechanism. Butterfield and his team had spent years building something people genuinely wanted to use. Glitch failed as a consumer game for a dozen reasons, many of them outside the team’s control. But in building it, they had also built something that solved a real, chronic problem: the chaos of coordinating a distributed creative team across email threads and IRC channels. They were the customer. The tool worked because they needed it to work.
This is the thing that outside observers almost always miss when a startup pivots: the best pivots are not random directional changes. They are transfers of hard-won, specific knowledge into a new container. The Tiny Speck team did not become Slack by deciding to enter the enterprise communication market. They became Slack because they had already solved the problem for themselves, repeatedly, under real pressure, and they understood the solution in a way that competitors couldn’t replicate from first principles.
Contrast this with the pivots that really are failures. A consumer app with no traction decides to ‘go B2B’ because enterprise sounds more serious and the checks are bigger. A hardware company with supply chain problems pivots to ‘consulting.’ These moves have the same surface structure as what Butterfield did, but they lack the underlying transfer of capability. They’re not pivots. They’re retreats with a rebrand.
The distinction is what you’re carrying with you. The startup that stops solving its original problem wins is a useful frame here, but only if you’re clear about what ‘the original problem’ actually was. For Tiny Speck, the original problem was never ‘build a successful game.’ It was ‘coordinate talented, distributed people working on complex creative problems.’ Glitch was one attempt at that. Slack was a better one.
YouTube is another version of this story. The company launched in 2005 as a video dating site. The founding team noticed that users were uploading all kinds of videos, not just dating profiles. They removed the dating mechanic and let it become a general video platform. From the outside, the pivot looked like abandoning the product. From the inside, it was recognizing what the product actually was. The capability they had built, reliable video hosting and streaming at a moment when that was genuinely hard, transferred perfectly. The use case that customers invented for them was better than the one the founders had planned.
The pattern shows up in company after company. Twitter grew out of a podcasting platform called Odeo after Apple announced that iTunes would support podcasts natively, essentially killing Odeo’s core business. The team ran internal hackathons to find something new, and a side project about short status updates became the entire company. Instagram started as Burbn, a check-in app with photo-sharing as a secondary feature. The founders stripped everything except the photos. Both of these looked, at various points, like a company failing to find product-market fit. They were, but the operative word is ‘finding.’
What made these pivots work was not the pivot itself. It was the founders’ willingness to look honestly at what was actually working, even when ‘what was working’ was a small, unsexy piece of a larger product that had failed. Butterfield’s memo is famous partly because of how clearly he describes this: they weren’t entering a market, they were solving a problem they understood better than anyone else alive.
Here is where I want to push back on the standard pivot narrative, because there’s a version of this story that startup culture tells incorrectly. The version I hear too often frames the pivot as a moment of inspiration, a founder who saw something nobody else saw and bravely changed direction. That’s not what happened at Tiny Speck. What happened was more like triage. They were running out of money. They had built something useful as a side effect. They made a rational decision to stop wasting resources on a product that wasn’t working and double down on one that was.
Bravery had nothing to do with it. Intellectual honesty did. Butterfield looked at the data, looked at his team’s actual capabilities, and made a call that looked bad from outside because the outside world was evaluating the wrong product.
The lesson for founders is not ‘be ready to pivot.’ Everybody says that. The lesson is more specific: when you’re evaluating what’s working in your company, audit the tools, processes, and side effects that you built for yourself. The communication tools. The internal dashboards. The workarounds your team invented because the available solutions were inadequate. Sometimes the real product is the infrastructure you built to build the wrong product.
The press will call it a failure. They’re covering what you shut down. You’re carrying what you learned.
