The Move That Looks Like a Mistake
In 2004, a small photo-sharing app called Flickr was spun out of a failed massively multiplayer online game called Game Neverending. The game flopped. The photo feature players used to share screenshots inside it did not. So the founders, Stewart Butterfield and Caterina Fake, stripped out the game and shipped the side feature as a standalone product. Flickr was eventually acquired by Yahoo for somewhere between $22 million and $35 million.
Then Butterfield did it again. He spent years building another online game, Glitch, which also failed. But the internal messaging tool his team had built to coordinate development? That became Slack. A company that sold for $27.7 billion to Salesforce in 2021.
Twice, a game failed. Twice, something that was never the product became the product. That is not luck. That is a pattern worth understanding.
The Adjacent Possible Is Not Random
When founders talk about pivoting, they usually frame it as desperation. The original idea didn’t work, so you find something that does. But that framing misses what’s actually happening in the best cases. The companies that successfully enter markets they weren’t built for are not flailing. They are following the logic of what Stuart Kauffman called the “adjacent possible,” the set of things that become achievable precisely because of where you currently stand.
Slack didn’t enter the enterprise communication market cold. It entered with a team that had already built and used the tool under genuine operational pressure, an intimate understanding of what remote coordination actually felt like from the inside, and a product that had been tested in one of the most chaotic environments imaginable: a startup trying to keep a game alive. The market entry looked accidental. The product quality it produced was not.
This is the hidden logic. When a company enters an adjacent market, it is often carrying non-obvious advantages that purely native competitors in that market don’t have. The outsider framing obscures what is actually a structural edge.
What You Know That the Industry Doesn’t Know It Needs
Amazon Web Services is the clearest large-scale example of this. Amazon built AWS because it needed it. The internal infrastructure required to run a massive e-commerce operation at scale was genuinely novel, and Amazon had spent years solving problems that cloud customers would eventually need solved. When AWS launched publicly in 2006, it wasn’t Amazon guessing that companies might want cheap computing. It was Amazon selling something it had already built, under fire, for its own survival.
The companies that were “supposed” to own that market, IBM, HP, the established enterprise vendors, had not been forced to solve the same problems. They sold hardware and services to companies that ran their own servers. They had no internal pressure to rethink the model from scratch. Amazon had that pressure because its own business demanded it.
This pattern repeats. Toyota’s production methodology, which became the gospel of lean manufacturing, came directly from Toyota solving its own cost and quality problems under post-war resource constraints. It didn’t come from consulting firms or business school theory. It came from a company that had no choice but to figure it out.
The implication for startups is direct: the best market entries are not strategic diversification exercises. They are the result of a company realizing that the solution it built for itself is more broadly applicable than it originally looked. The question is whether you can recognize that moment when it arrives.
Why Incumbents Can’t Follow
One reason this strategy works is structural, and it doesn’t get discussed enough. When a startup enters a new market carrying capability it built for a different purpose, incumbents in that market face a genuine response problem. They can see the new entrant. They cannot easily replicate the conditions that produced the entrant’s actual advantage.
An enterprise software company looking at Slack in 2015 could not simply decide to build a better internal communication tool. They would be building it for external customers from the start, under entirely different pressures, with entirely different feedback loops. The knowledge Slack had accumulated from building for itself, then refining through a specific team’s actual use, was not transferable via competitive analysis.
This is related to why, as I’ve written before, tech giants don’t expand randomly, and the order they enter markets is the strategy. Sequence matters. Each new market entry builds on capabilities that were forged somewhere else, and that compound advantage is hard to see from the outside and nearly impossible to replicate on demand.
The incumbent’s typical response, acquiring the startup or cloning the product, often fails to capture the underlying advantage. Microsoft Teams is a capable product. It has never felt like something built by people who desperately needed it to exist.
How to Read Your Own Capabilities Honestly
None of this means founders should deliberately build things that fail so they can find the hidden gem inside the wreckage. What it means is that the capabilities a team develops while solving one hard problem deserve serious examination as potential products in their own right.
The questions worth asking are specific. What have we built internally that we would miss if we didn’t have it? What do we know, from direct operational experience, that people in adjacent markets clearly don’t know yet? Where have we been forced to solve something that others have not been forced to solve?
Those questions will surface things that look modest from the inside, because familiarity flattens perceived value, but that look genuinely novel from the outside. The most disruptive startups solve problems nobody has officially admitted exist yet, and the companies with the best chance of doing that are often the ones sitting on hard-won internal solutions they haven’t thought to productize.
Butterfied didn’t set out to build an enterprise communication platform. He set out to build a game, built a communication tool to survive the process, and had the clarity to recognize which of those things was actually valuable. That clarity, the willingness to see your own byproducts clearly, is rarer than the strategy itself.
The myth is that great startups find a market and go conquer it. The reality is messier and more interesting: they solve a hard problem in one context, develop capabilities that transfer, and then notice that the world outside looks a lot like the problem they already solved.
