Three Products. One Billion. The PM Lessons That Actually Matter.

Product management has a mythology problem. We collect the wins — the acquisitions, the billion-dollar exits, the hockey-stick charts — and strip out everything that makes those outcomes genuinely instructive: the failed bets, the pivots born from desperation, the moments where the right call required burning down the product everyone had worked to build.

The three stories I want to tell you are not myths. They are case studies in the actual craft of building — the grinding, ambiguous, occasionally humbling work of figuring out what a product should be and finding the nerve to make it so. Slack. Quibi. Instagram. One became a $27 billion enterprise communications platform. One became one of the most expensive cautionary tales in tech history. One became a billion-dollar acquisition eighteen months after its founders stripped everything interesting out of it.

All three were built by smart, experienced people. That is the point.

---

It started in 2009. Butterfield had left Yahoo with a payout and a hunger to build something stranger than photo sharing. He gathered a small team, raised a modest round, and set out to build a massively multiplayer online browser game called Glitch. Not a shooter. Not a strategy game. A collaborative, whimsical world where players planted crops in the imagination of eleven sleeping giants and cooked food for imaginary creatures and built towns with strangers from across the internet.

It was, by all accounts, genuinely beautiful. Beautiful in concept, beautiful in execution. The team cared deeply about every pixel, every mechanic, every piece of lore. They shipped. They iterated. They believed.

Nobody came.

Or more precisely: not enough people came, and the people who came did not stay in the numbers the model required. Glitch launched in beta in 2011. By November 2012, after three years of building and $17 million in investment, Butterfield shut it down. He wrote a public farewell post that remains one of the most gracious and honest post-mortems in the industry — not blaming the market, not blaming timing, just acknowledging that the game they built required a specific kind of imagination from its players, and not enough players showed up with that imagination intact.

The game had failed. But while Butterfield's team had been building it, they had built something else. Something small. Something internal. Something nobody had designed to be a product at all.

To coordinate across the game's distributed engineering team, they had built a persistent, channel-based messaging tool. It was not remarkable by external standards — just a better way to talk, to share context, to move work through a team that was spread across time zones and disciplines. They used it every day. They loved it. It solved a problem that nobody had asked them to solve, in a domain they had not set out to work in, as a byproduct of work they were shutting down.

Butterfield looked at the ruin of Glitch and asked the only productive question available: what did we build in the process of building this?

The answer was Slack. And the rest of that story involves a $26.5 billion acquisition by Salesforce in 2021 and a product used by hundreds of thousands of companies around the world. Built in the wreckage. Shipped in four months after they decided it was real. Scaled because it solved a problem so fundamental — human coordination at work — that once you had it you could not imagine having gone without it.

---

The idea behind Quibi was elegant in the way that wrong ideas often are. In 2018, Katzenberg looked at how people consumed media on their phones and identified what he believed was an underserved format: high-quality, narrative video designed specifically for mobile, in episodes no longer than ten minutes. "Quick bites" — hence the name. Premium content. A-list talent. A subscription model. Launched into a world where every commute, every queue, every lunch break was an opportunity.

The pitch was compelling enough to raise $1.75 billion before a single frame of content was shown to a real subscriber. Disney, NBCUniversal, Sony, Alibaba — serious money from serious people who believed in the thesis. They hired top directors, commissioned exclusive shows, built a technical infrastructure from scratch with a feature called Turnstyle that let you rotate your phone and the video would reframe itself between portrait and landscape. It was genuinely impressive engineering in service of a bet nobody had tested with actual humans.

Quibi launched in April 2020. By October 2020 — six months later — it was dead.

The autopsy has been written many times, and most versions blame one of three things: the timing (COVID-19 locked people home, so nobody was commuting; the "mobile snacking on-the-go" use case evaporated overnight), the product decisions (no ability to cast to a TV, no screenshotting, no sharing clips — a video platform with no virality), or the content itself (expensive but largely forgettable, without the breakout hit that anchors a new platform in culture).

All three analyses are correct. None of them are the root cause.

The root cause is that nobody outside of the building ever validated the core premise. Not seriously. The question — will people pay $7.99 a month for premium mobile-only vertical video on a standalone app? — was never answered with data before $1.75 billion was committed to answering it at scale. There were focus groups. There were internal convictions. There were investor theses. There was the authority of Katzenberg's Hollywood track record and Whitman's operational credibility.

What there was not, anywhere in the story, was a genuine signal from the market before the bet was placed.

This is the thing about conviction: it feels exactly like insight. The internal sensation of believing something strongly is indistinguishable from the internal sensation of being right. The only way to tell the difference is to take your conviction outside and let the market touch it before you scale it.

Katzenberg has said in hindsight that the COVID timing was the fatal blow — and maybe it accelerated the end, but it did not create the underlying problem. The underlying problem was that the product was built on what two brilliant, experienced people thought people wanted, not on what people had demonstrated they would actually do. A $1.75 billion bet on a belief. And when the belief collided with reality, there was nothing underneath it to hold.

The most painful detail in the whole story is this: YouTube already existed. Netflix already existed. TikTok was already exploding. Users had already demonstrated what they wanted from video on their phones — and none of those signals pointed toward a paid, isolated, non-shareable, TV-only-on-phone format. The market was speaking. The founders just were not listening, because the conviction was too strong and the capital too available to require it.

---

In 2009, Kevin Systrom had an idea for an app that combined the location check-in mechanics of Foursquare with the social aesthetics of whiskey bar culture — a place for people to share where they were, earn points for showing up, make plans with friends, and post photos. He called it Burbn, after his love of American bourbon.

He built it. He pitched it at a party to two venture capitalists from Baseline Ventures. They gave him $500,000 on the spot. He used it to hire one co-founder — Mike Krieger — and begin building the full product.

When they launched Burbn, people used it. Not many. But real users, trying the real product. And Systrom and Krieger did something that most PMs with a funded product and an eager small user base never do: they watched. Closely. Without the assumption that they already knew what was working.

What they saw was uncomfortable. The app was cluttered. The check-in feature was confusing. The points mechanic went largely ignored. The plans functionality was barely touched. But the photos — the ability to post pictures with a filter and share them — those were getting used. People were taking photos with Burbn in a way that felt different from the surrounding features. There was something real there, underneath all the complexity they had built.

Systrom and Krieger could have iterated. They could have simplified the onboarding. Cleaned up the UX. A/B tested the points mechanic. Run a few sprints of polish and called it a v2. Instead, they made a decision that most product teams are not brave enough to make: they decided to start again.

They stripped the product down to four things: take a photo, apply a filter, write a caption, share it. Everything else — the check-ins, the points, the plans, the Foursquare-like features they had spent months building — gone. Not paused. Not moved to a roadmap. Gone.

Instagram launched in October 2010. One million users in two months. Ten million by June 2011. Eighteen months after launch, Facebook acquired it for one billion dollars. At the time it had thirteen employees.

Think about what the strip required. They had to look at months of work and decide it was wrong — not because it was built badly, but because the user signal pointed somewhere simpler. They had to convince themselves and each other that the product they should be building was smaller than the one they had built. They had to kill features they had debated, designed, shipped, and watched people use. They had to bet that the thing people were actually doing — the photo sharing, quiet in the corner of a complicated app — was more real than the full vision they had launched with.

That bet required a specific kind of courage that product roadmaps do not develop in people. It required the discipline to follow the signal even when it meant going backwards. It required intellectual honesty about the gap between what you wanted users to love and what users were actually doing. And it required a definition of success that put market behaviour above internal consensus — the willingness to let reality correct your vision instead of forcing your vision onto reality.

---