#347 2012 · Tiny Speck (Stewart Butterfield) · Software / enterprise communication
A founder shut down his second failed video game and looked at the tool his own engineers had built to survive it
the problem
Years of investment in a product had failed to find a market, leaving a team, some cash, and nothing else to show for it
background
Stewart Butterfield's Tiny Speck spent nearly four years and substantial funding building Glitch, a deliberately non-violent, collaborative browser game with no combat and no win condition, launched publicly in September 2011. Despite a devoted niche following, the game's unconventional design and Flash-based, non-mobile architecture meant it never attracted a mass audience large enough to cover its operating costs, and the company shut it down on 9 December 2012.
This was not Butterfield's first time watching a game fail: his earlier venture, Game Neverending, had also failed to find players a decade before — but a minor photo-sharing feature buried inside it had become popular enough on its own that Butterfield spun it out as Flickr, sold to Yahoo in 2005. Facing a second dead game, Butterfield's team had another asset most startups would have overlooked entirely: an internal chat tool they had built themselves during Glitch's development because existing options like IRC weren't good enough, one that was searchable, logged everything, and worked reliably across devices.
what everyone would do
The standard startup-failure playbook is to wind the company down and return remaining capital to investors, treating anything built along the way as sunk cost — the internal chat tool existed to serve the game, not as a product in its own right, and most teams in the same position would have shut it down along with everything else rather than examine it separately.
what they saw
Butterfield's team saw that the chat tool wasn't a byproduct of building Glitch, it was itself the more valuable thing they had built — good enough that their own team couldn't have shipped a game without it, which meant the same unmet need almost certainly existed in every other distributed team coordinating complex work with inadequate tools like IRC.
the move
Rather than wind the company down and return remaining funds, Butterfield proposed using the leftover capital and the same four-person founding team to turn the internal tool into a standalone product. Co-founder Cal Henderson reportedly summed up the pivot in one line: the tool is the product. Development on what became Slack began within a month of Glitch's shutdown, with almost no new hiring — the team that had built a communication tool to survive making a game now built it to sell to every other team facing the same problem.
why it works
The tool had already been battle-tested under the exact pressure that matters most — a real team relying on it daily to ship a real product — so it arrived at the pivot with the features that use had forced into it (searchable history, reliable logging, cross-device access) rather than features guessed at from a blank slate. Because the same founding team and much of the underlying technology carried over, almost no time was lost re-learning a new domain, letting Slack ship within months of leftover runway instead of requiring a fresh product-market-fit search from zero. In effect, the company skipped the slow part of most startups' lives — discovering whether anyone needs the product — because they already knew people needed it: they were those people.
the payoff
Slack's private beta in August 2013 drew 8,000 signups on its first day; by February 2015 daily active users passed 500,000, a 33x increase in twelve months, and the company reached a $1 billion valuation in roughly eight months without a traditional advertising campaign — 70-80% of early adopters were switching not from a competitor but from no dedicated tool at all.
where it breaks
The pivot only works when the internal tool solves a problem common across many other organizations, not one peculiar to the failed venture's own workflow — a bespoke tool built around one company's idiosyncratic process has no broader market once that company is gone. It also requires someone to actually notice and ask the question, since most failing teams treat their internal tooling as scaffolding to discard along with the failed product rather than scrutinize it as a separate opportunity. And it depends on the team having enough people and capital left to build and launch a new product rather than simply shutting down — a venture with no runway or team left to redeploy has no way to act on the same insight even if it recognizes it.
what came after
Slack rejected roughly 8-10 acquisition offers by mid-2015, including a reported $8 billion Microsoft bid in 2016 (prompting Microsoft to build Teams as a direct competitor), before Salesforce acquired Slack for $27.7 billion in 2021 — Salesforce's largest acquisition to date — making Butterfield one of the very few founders to have built two separate billion-dollar-plus companies (Flickr, then Slack) out of the wreckage of two separate failed games.
references
- [1]The Video Game That Failed: The Origin Story of SlackStackSync, 2025stacksync.com
- [2]The death of Glitch, the birth of SlackBuilding Slack, 2024buildingslack.com