Part IThe Story
The table had held the dead before it held lunch. It was long and made of metal, and it had started life as a gurney in the morgue of the Treasure Island Naval Base, the artificial island in San Francisco Bay. In the summer of 2014 it stood in what would become the kitchen of a company that had not yet hired most of its staff. Stewart Butterfield sat at it eating potatoes smothered in chicken-fat drippings. He had been in the building less than a week. The space was a glass transept at the eastern edge of a large new building just off Third Street, with room for seventy-five people or so. On a busy day there might have been thirty.
The product those people would build was, by the admission of nearly everyone who described it that year, boring. It was a chat program. It had channels, which were rooms, and direct messages, which were conversations, and a search box that remembered everything anyone had ever typed. It did not have a sales team. It did not need one. Between February and June its daily users had gone from 10,000 to 90,000, and the company's headcount had gone from eight to twenty to thirty-four in the time it takes most startups to argue about a logo. The thing was called Slack. Its name was a backronym, Searchable Log of All Conversation and Knowledge, and the joke inside it, a "slack" product for people at work, was the kind of joke a man makes when he has stopped worrying about seeming serious.
He had stopped worrying about a lot of things. Twenty months earlier he had stood in front of his employees at a company called Tiny Speck and told them, with tears running down his face, that the video game they had spent three years building was finished. The game was called Glitch. It was strange and beautiful and nobody played it. The chat tool the team had built for itself while making the game was the only thing to survive the wreck. A decade before that he had done almost exactly the same thing: built a game nobody played, broken off one feature, and sold the feature to Yahoo as a photo-sharing site called Flickr.
So the morgue table was a joke too, though nobody seems to have said so out loud. This was a company assembled from a corpse, sitting at furniture built for corpses, run by a man who had now buried two games and was about to become a billionaire for the second burial. Six and a half years later Salesforce would pay $27.7 billion for what grew out of that room. By then the table was gone, and so, in a sense, was everything else. That's how it goes with Butterfield. The thing he set out to build is never the thing that lasts. The thing that lasts is whatever people used to talk to each other while he was building it.
By the Numbers
Two Dead Games, Two Living Companies
$22–25MYahoo's price for Ludicorp and Flickr, March 20, 2005, by Butterfield's own estimate
$17.5MVenture capital raised for Tiny Speck to build the game Glitch
8 → 34Slack employees, December 2013 to June 2014, with no sales or marketing team
$400.6MSlack revenue, fiscal year ending January 31, 2019, up from $105.2M two years earlier
$26NYSE reference price for Slack's direct listing, June 20, 2019; shares opened at $38.50
350,000IBM employees moved onto Slack in February 2020, its largest customer
$27.7BSalesforce's acquisition price, announced December 1, 2020
The Word for Child Was Noah
The highway runs out at Lund. It is a fishing village on the northern Sunshine Coast of British Columbia, the place where, as a Stanford profile put it, the Pacific Coast Highway ends. It is also where a boy named Dharma Jeremy Butterfield was born on March 21, 1973. His father had crossed into Canada from the United States to avoid being drafted for the war in Vietnam. His parents joined a commune and lived in a log cabin in the forest. For about three years the cabin had no running water and no electricity.
"My parents were definitely hippies," he told the BBC decades later, with the mild amusement of someone who has told the story often. "They wanted to live off the land, but it turns out there was a lot of work involved, so we moved back to the city."
One detail from those years outweighs the rest. The nearest other child lived a mile away, and his name was Noah. The isolation was so complete that the toddler Dharma, according to a 2015 USA Today profile, thought "Noah" was simply the word for child. One example stood for the whole category. If you wanted to design a person who would spend his adult life building places where strangers could find one another, you could hardly do better than to start him in a cabin where the entire population of other children came to one.
The family moved to Victoria when he was five. His father became a real estate developer, and the son absorbed the rhythm of that work: the long cycle of vision, then planning, then raising money, then waiting years for something to stand up out of the ground. "I started looking at the world that way," he said. His mother supplied a different lesson. When he totaled his father's car at sixteen, she told him, "Now you've learned an important safety lesson." It was not the reaction he expected, and he has repeated it ever since, perhaps because it captures the household's improbable calm about loss. A wrecked thing was a lesson, not a catastrophe.
Around 1980, when he was seven, his parents bought an Apple II, or perhaps a IIe. He is not quite sure which, and he has given both answers. He taught himself to program by typing BASIC listings out of Byte magazine, changing a couple of lines, and watching what broke. He made simple games. He also made money. At a beach near the family home he bought hot dogs at the local 7-Eleven and resold them on the sand at a markup. He was, he would later say, "among the first cohort of kids to grow up with computers," but he grew up with the lemonade-stand instinct too, and the two never fully separated.
At twelve he changed his name to Stewart.
There is no satisfying public explanation for this, and he has never seemed interested in providing one. A boy named after a Buddhist concept of cosmic order by parents who had dropped out of America decided, at the age when most children are deciding which bands to like, to call himself something you might find on a bank loan officer's nameplate. His legal name became Daniel Stewart Butterfield. Years later, a United States patent for a ranking method called "interestingness" would list its first inventor as "Butterfield; Daniel S.", one more name on the pile. The renaming fits a pattern that would become clearer with time. He keeps the material and discards the label. Tiny Spec became Tiny Speck within a week of incorporation. Game Neverending became Flickr. Glitch became Slack. Dharma became Stewart.
In February 2009, giving the opening keynote at a blogging conference in Vancouver called Northern Voice, he began with a photograph of a shack in the woods, then black-and-white pictures of his hippie parents, then a picture of himself as a boy in a trucker cap from Radio Shack Computer Camp. The phrase he kept repeating through the talk was This is who I am. By the end he had turned it around. Community, he said, changes the sentence to This is who we are.
The Big Five Philosophy Firms
He drifted away from computers in high school, at St. Michaels University School in Victoria, and toward the Grateful Dead and Phish, the long-jam bands whose fans traded tapes and drove across the continent to follow tours. Then, in 1992, as a university student in British Columbia, he found the internet, and the two interests merged.
Before the web as most people know it, there was Usenet, a sprawling set of discussion groups organized by topic. His first online community was rec.music.phish. "Using the Internet for music distribution at that point was, you find someone who will dub a cassette and mail it over in a padded envelope," he told the Globe and Mail. What astonished him was not the cassettes. It was that a kid from the edge of the map could find people who cared about exactly what he cared about, "anywhere in the world, even from such a remote place." His early handles had the clumsy specificity of the period: ui503, sbutterf, dsb26. "Ever since," he said in 2020, "the idea of using computing technology to facilitate human interaction has guided my career."
That sentence, delivered to a Stanford business school audience, makes a winding route sound straight. He did not go into computing. He went into philosophy. He earned a bachelor's degree from the University of Victoria in 1996 and a Master of Philosophy from Clare College, Cambridge, in 1998, with a thesis on nineteenth-century scientific thinkers and a focus on biology, cognitive science, and the philosophy of mind. He intended to become a professor. His own professors talked him out of it, warning that a coming glut of PhDs would flatten the academic job market. "By the time I finished my master's degree I really had no idea of what I was going to do except for be an academic," he told Bloomberg, "because, you know, the big five philosophy firms aren't always hiring."
The joke is a good one, and it hides how ordinary his reasoning was. Around 1997, he said, "people who knew how to make websites were moving to San Francisco, and I had a bunch of friends who were making twice as much, or three times as much, as what professors were making." He had taught himself HTML during university summers and made money designing websites. He did the arithmetic.
His first real job in the industry was at a Vancouver web agency whose founder had bought up hundreds of domain names, among them Dance.com and Brazil.com, betting that the explosion of the web would make them valuable. Some people call this domain squatting. It is rarely a sign of a company that's going anywhere. "It was a disaster," he said. "I quit literally two weeks before the dot-com crash. I thought I was walking away from 10 million bucks in equity." He was walking away from nothing, which in March of 2000 was the best trade available.
He then joined his friend Jason Classon at a startup called Gradfinder.com, a site for finding former classmates. They sold it, after the bubble had already burst, for what Butterfield calls "a healthy profit." Classon went to work for the acquirer. Butterfield went back to freelance design and, in his spare time, started a contest called the 5K, which challenged people to build an entire website in under five kilobytes. "It became unexpectedly huge, in every country in the world," he said. It was a game whose prize was mostly the pleasure of playing. Nobody made money from it. People found each other through it. Lots of things in his life would follow that shape.
Ludus
Caterina Fake grew up in northern New Jersey, the daughter of an American father and a Filipina mother, in a house where television was forbidden and the entertainments were poetry and classical music. She went to Choate, transferred from Smith to Vassar, and graduated in 1991 with a degree in English. Vassar had an intranet that students could reach from their dorm rooms, and she later credited it with pulling her toward the web. Before that came what she called her "post-college what-do-I-want-to-do period," which took her through jobs as a painter's assistant, an investment banker, and an employee at a dive shop. She got stuck in San Francisco while visiting her sister, taught herself to program, became art director at Salon, and in 1997 took a job managing the community forums at Netscape. Online, she had been building the same thing he had been looking for: rooms where like-minded people gathered.
They met as web designers, she in San Francisco and he in Vancouver. She moved north. They married in 2001. In the summer of 2002, just after the honeymoon, they founded a company with Jason Classon. Fake named it Ludicorp, from the Latin ludus, which means play.
They were building a game, and it was a peculiar one. Game Neverending, as they called it, had no combat and, more radically, no way to win. There was no definition of success. Players wandered around, chatted on global and location-specific channels, left notes for one another, and combined objects to make other objects whose usefulness was, in one contemporary description, "questionable." The official currency was the shekel. Sheets of colored paper, along with quires and reams, became an unofficial second currency because players decided they should be. Fake cited Neopets as an inspiration. The division of labor, as she described it to Jessica Livingston for Founders at Work: "[She] did the game design, Stewart did the interaction design and Jason did the PHP for the prototype."
The team they assembled reads now like the masthead of a later era. Cal Henderson, a British developer, ran web development; he would become Slack's chief technology officer. Eric Costello, who would also become a Slack co-founder, led client development. George Oates was the producer. Ben Cerveny was the game designer and an advisor. The advisory board included Cory Doctorow, Andrew Zolli, and Clay Shirky. These were people who thought about the internet as a social and civic space, not as a storefront. In 2002 that was exactly the wrong kind of person to be.
The earliest prototype went live in the fall of 2002 and was taken down on February 3, 2003. It had a cult following. It also had no money. "It was a really black and bleak-looking point in the history of financial markets generally," Butterfield told Reid Hoffman, "but anything as frivolous as a game is just not going to get funded." Ludicorp scraped together a Canadian government loan and got close to breaking even. Then the cash ran short enough that, in his words, "the only person on the team who got paid was the one person on the team who had kids."
If the game had a goal at all, the serious players agreed it was to build the final item, an object that was itself called a Game Neverending. The only way to finish the game was to make the game. Most people would call that a design flaw. Butterfield seems to have regarded it as a description of his life.
Puking on the Van Wyck
The founding story of Flickr is so unglamorous that it almost has to be true. Ludicorp was dying. Butterfield and Fake flew to New York for a video game conference, partly to stall, because they did not know what to tell the team. On the plane he got food poisoning.
"This sounds almost made up," he said. "But on the way, I got food poisoning on the plane, puking on the Van Wyck on the way into New York, arriving at the hotel. Just being sick all night — it's like 3:00 or 4:00 in the morning — trying to keep anything down, like ginger ale or water. And the whole idea for Flickr came to me."
The Van Wyck Expressway, which carries traffic from JFK toward Manhattan, has inspired very little in the history of American thought. On that night it produced the first big consumer success of the post-crash web. Butterfield is careful not to let the story become a myth about inspiration.
— Stewart Butterfield, Masters of ScaleIt wasn't coming from a grand vision of what photos could be. All that stuff came later. There was no insight, it was just like, "Can we not go out of business?"
The raw material was already in the building. Game Neverending had a feature that let players share photographs with each other. It worked badly, because both people had to be online at the same moment. Flickr's first version, built in eight weeks on top of the game's code, kept the real-time idea. Its centerpiece was a chat room called FlickrLive where people traded photos as they talked. Then, version by version, the team noticed that people cared less about the live room than about the shelf behind it. They wanted somewhere to put pictures, label them, and come back to them. The chat room got buried in the site map and eventually disappeared. For years afterward, URLs on Flickr ended in a file extension,
.gne, a small fossil of the game that had died to make room for the site.Flickr launched on February 10, 2004. On June 7 the company announced "Notes and Tags." Tags let anyone attach plain words to a photo, which meant the site's organization came from users rather than from administrators. That bottom-up taxonomy, later called folksonomy, became one of the defining ideas of what the industry started calling Web 2.0. Photos were public by default. Following someone did not require them to follow you back. A stranger could look at your pictures without an account. Nobody at Ludicorp could have known that these were the founding conventions of the next fifteen years of social software. They simply seemed like how a place for sharing ought to work.
Luck helped, as Butterfield is always quick to say. Google had recently bought Blogger and had no way for bloggers to host images. "When someone asked how," he recalled, "they said to use Flickr." By the end of 2004 bloggers were embedding Flickr photos across the web, and the business was profitable, which in 2004 was rare enough to be suspicious.
Yahoo acquired Ludicorp and Flickr on March 20, 2005. The price has been reported at anywhere from $22 million to $35 million. Butterfield, when asked, says "somewhere between $22 million and $25 million" and calls the higher estimates too high. It was, as a public radio profile later put it, "the first big sale of a tech company" since the bust, the deal that broke the logjam. During the week of June 26, 2005, every photo on Flickr was moved from servers in Canada to servers in the United States, and with that the data came under American federal law. The draft dodger's son had built a company in British Columbia, and its contents now crossed the border going the other direction, into the country his father had left.
In 2006 Butterfield and Fake appeared together on the cover of Newsweek and on the Time 100. They were the internet's couple. They were also, at that point, employees.
A Tinsmith in a Company That No Longer Made Tin
"It's hard to get things done in an organization of 12,000 people," Butterfield said of Yahoo, many years later, with the restraint of someone who has made peace with something. At the time it was harder.
The work was not trivial. As general manager of Flickr he watched the site become, by 2007, the nineteenth most popular website on the internet. He also watched the parent company make decisions that its users experienced as small betrayals. On January 31, 2007, Flickr told its "Old Skool" members, those who had joined before the acquisition, that they had until March 15 to link their accounts to a Yahoo identity or lose access. People complained loudly. In May, Yahoo announced it would shut down its own photo service, Yahoo Photos, and send its users to Flickr. Flickr was being absorbed and celebrated at once.
The company's corridors were full of people who would go on to run large parts of the industry. "When Flickr was acquired by Yahoo, the people that were sitting adjacent to me were Jeff Weiner, now CEO of LinkedIn, and Bradley Horowitz, who's senior VP at Google," he recalled. "It wasn't like Yahoo at that time was anything special. It's just that, if you worked at eBay, PayPal, Facebook, Apple, all your peers went on to work at or start other companies, and then other companies." One person he met there was Andrew Braccia, who would later become an investor at Accel and back Butterfield again. "What makes Stewart really unique," Braccia said, "is his desire to combine artistry and science." At Yahoo the artistry had nowhere much to go.
His private life came apart in the same period. He and Fake divorced in 2007. Their daughter was born on July 11 of that year. On June 13, 2008, he announced that he would resign effective July 12. Fake was leaving too. The press called it part of a "mass exodus" from Yahoo, and for the two of them the exodus was personal as well as corporate.
The resignation letter, addressed to Brad Garlinghouse, became the most widely circulated thing he wrote until a certain memo six years later. It was mock-formal and full of invented biography. In it he compared himself to a tinsmith in a company that no longer made tin. It ended, more or less, with a request that nobody throw him a party. The Guardian headline used the quote, "I don't need no fancy parties." Underneath the jokes, the letter said what Ludicorp's chroniclers would later describe more plainly: he felt sidelined, without much say over the thing he had made.
"It was horrible at the time," he said much later, "but I did develop character." Then, with the generosity he tends to show former employers in hindsight: "I learned a lot more than any MBA program. I was involved in things I never would have been involved with otherwise."
Flickr had a long, slow afterlife without him. In 2011 it reported more than six billion images. Facebook and Instagram took its audience. Verizon took Yahoo in 2017 and sold Flickr to a small, family-run company called SmugMug on April 20, 2018. One part of what Ludicorp built did hold up, though, and it came from the producer, not the founders. In January 2008, a few months before Butterfield left, George Oates conceived and launched "The Commons," which gave the Library of Congress and other institutions a place to share historical photographs with "no known copyright restrictions." Fourteen years later, in 2022, Oates became director of a new nonprofit, the Flickr Foundation, set up to keep Flickr's tens of billions of photographs visible for a hundred years. The game designer's producer ended up keeping the archive. The archive was the last thing anyone had meant to build.
Dr. Seuss Meets Monty Python
Then he went back and made the same game again.
That is the strangest decision in his career, and it gets the least attention. It is not quite fair to call it stubbornness. He had a theory, and the theory was reasonable. "Now there was an order of magnitude more people online," he said of 2009. In 2002 most people had no computer at home and no internet connection. The game had failed, he reasoned, because the world had not been ready for it. Now the world was ready.
The company was incorporated in Delaware on February 25, 2009, as Tiny Spec, Inc. Eight days later, on March 5, someone fixed the spelling to Tiny Speck. The company raised $1.5 million from angels in 2009, $5 million from Accel and Andreessen Horowitz in 2010, and $10.7 million more from the same firms in April 2011, for $17.5 million in all. He recruited former Ludicorp and Flickr colleagues, among them Cal Henderson, Eric Costello, and Serguei Mourachov. He also attracted unusual talent from the art-game world: Keita Takahashi, the creator of Katamari Damacy, and Robin Hunicke, a producer of Journey. He described the new game as "Dr. Seuss meets Monty Python." It was called Glitch.
Glitch was gorgeous, hand-drawn and two-dimensional, with a mythology in which the whole world existed inside the imaginations of eleven sleeping giants. Like its predecessor, it had no combat. One description of its gameplay reads like a corporate onboarding manual: "Players must learn how to find and grow resources, identify and build community and, at the higher levels of the game, proselytize to those around them." It launched on September 27, 2011. On November 30, in a move almost nobody in games had tried, Tiny Speck unlaunched it, pulling it back into beta to fix it.
What went wrong is in some ways the most honest part of the record, because Butterfield answered the question at length while the wound was fresh. "It took us until the final four months for the game to actually become really fun from a moment-to-moment gameplay point of view," he told Gamasutra in November 2012. "Sometimes we were in such a rush to complete a feature that the purpose of the feature wasn't realized. There's dozens of examples." He resisted the easy diagnosis. Players complained that the game was "just clicking," but "Civilization is just clicking, and Diablo is just clicking." Player housing never acquired a purpose. One mechanic did work. Collecting coins gave a small benefit to every other player standing nearby, and players began moving through the treetops in packs. "There was this flocking behavior," he said. "That was super fun collaborative play."
There was not enough of it. "Ultimately," he said, "if I have to identify one thing as the problem — I don't think there is just one — but if I had to choose just one, I think the game was too foreign of a concept for most people." New players needed about fifteen minutes to understand what they were looking at, and most left before then. The bucket leaked. "The people who loved the game really, really loved it," he said. "If we had figured out an easier way to get more people to that state it would have been a success."
He could have kept it alive. Glitch had revenue and devoted fans. "It would have been a fine lifestyle-business," he said, "but it was never going to become the type of business that would justify seventeen and a half million dollars of venture capital." Or, more simply: "We could have kept going, but that wouldn't have been very responsible to our investors."
The announcement came on November 14, 2012. He called an all-hands. He wept. "It was humiliating," he said. "There was a real sense I had failed all these people." To a Stanford audience years later he reached, as many people of his generation do, for Star Wars. It was, he said, "like when they blew up Alderaan." On the game's forums, one player's entire response was "Stunned. And crying." Tiny Speck spent its last weeks of running the game coaching its staff, counseling them, and introducing them to other employers. The world of Glitch closed on December 9, 2012. In January 2013 the company released most of the game's art under a Creative Commons license, so that what it had made could be used by anyone. Exactly two years after the shutdown, on December 9, 2014, a group of fans began alpha-testing a rebuilt version of the game under the name Eleven, after the sleeping giants.
The day after the shutdown announcement, a TechCrunch reporter asked Butterfield what came next. Was he planning a new photo service? He said something mysterious: "You will know it well."
— Stewart Butterfield, Lenny's Podcast, 2025The decision is about, have you exhausted the possibilities? Creating the distance so that you can make an intellectual, rational decision about it, rather than an emotional decision is essential. The reason I say you have to be coldly rational about it is because it's fucking humiliating.
Searchable Log of All Conversation and Knowledge
What survived was the tool the team had used to talk to each other while building the game. Tiny Speck had worked across Vancouver, San Francisco, and New York, and to coordinate, its engineers had built themselves an internal messaging system with channels, persistent history, and search. Nobody had considered it a product. It was plumbing. After the game was gone, the plumbing was the only thing in the building that people still used every day, and it occurred to the remaining dozen or so employees that other companies might want it too.
They spent the first half of 2013 turning it into something a stranger could understand. In August, Butterfield announced it to the world, and the headlines put the pitch more crudely than he would have. CNET: "Flickr founder plans to kill company e-mails with Slack." ReadWrite: "Die, Email, Die!" The product went out in a preview, to companies Butterfield and his colleagues more or less begged to try it, and accumulated roughly 16,000 users in its first six months without any advertising. When those early teams said something wasn't working, the company's later account of the period ran, "we fixed it — immediately." Slack opened to the public in February 2014.
Before the launch, Butterfield wrote a memo to his staff. It was titled "We Don't Sell Saddles Here." A version of it circulated publicly that February, and it became one of the most quoted product documents of its decade. The argument was that a company sells the transformation, not the equipment. Slack's customers weren't buying a chat app. They were being asked to change how their organizations worked.
"The best — maybe the only? — real, direct measure of 'innovation' is change in human behaviour," he wrote. "All products are asking things of their customers: to do things in a certain way, to think of themselves in a certain way — and usually that means changing what one does or how one does it; it often means changing how one thinks of oneself." The passage that mattered most to his engineers, though, was about tolerance:
— Stewart Butterfield, "We Don't Sell Saddles Here," 2014When you want something really bad, you will put up with a lot of flaws. But if you do not yet know you want something, your tolerance will be much lower. That's why it is especially important for us to build a beautiful, elegant and considerate piece of software. Every bit of grace, refinement, and thoughtfulness on our part will pull people along. Every petty irritation will stop them and give the impression that it is not worth it.
Read beside the Glitch post-mortem, the memo is clearly an autopsy turned into an operating rule. "Too foreign of a concept." Fifteen minutes before it made sense. A world players had not known they wanted, and so would not forgive. He had now lived through that lesson twice, and this time he wrote it down before launch rather than after the funeral.
He was also, by his own later account, ruthless about the product in a way that alarmed people who expected founders to sound proud. In 2014, the year Slack launched, MIT Technology Review asked him whether the company was working to improve it. "I feel like what we have right now is just a giant piece of shit," he said. "It's just terrible and we should be humiliated that we offer this to the public." A decade later he explained that he meant it as a standard, not as a confession: "If you can't see almost limitless opportunities to improve, then you shouldn't be designing the product."
The company's ambitions grew faster than its self-regard. When Tiny Speck first pitched the messaging product to its investors, Butterfield said, the team hoped Slack might reach $100 million a year in revenue "at our fullest potential." It passed that within a few years. On July 17, 2014, the Delaware corporation formerly called Tiny Spec, then Tiny Speck, legally became Slack Technologies, Inc. On August 28 the company's Twitter account said goodbye to the old name in a single line: "Tiny Speck is no more. We're now Slack Technologies, Inc. See @SlackHQ. Bye!"
That summer it moved into the building off Third Street, with the gurney from Treasure Island as its kitchen table.
Selling Gold
Then the money arrived, and Butterfield, who had spent 2002 unable to raise a dime for a game, treated it as an economist would.
In April 2014 Slack raised $42.75 million. In October it raised $120 million at a $1.2 billion valuation, in a round led by Kleiner Perkins and GV, the venture arm of Google's parent. In April 2015 it raised $160 million more at about $2.8 billion. The company was barely a year old as a public product. It had 750,000 daily active users and 200,000 paying ones. Most estimates called it the fastest-growing business application ever made. A New York Times columnist pointed out that Butterfield had previously said Slack had enough money, and asked why he was raising more.
"It's pretty straightforward," he answered. "I've been in this industry for 20 years. This is the best time to raise money ever. It might be the best time for any kind of business in any industry to raise money for all of history, like since the time of the ancient Egyptians." He was accepting "$160 million bucks for 5-ish percent of the company" on favorable terms, he said, and it would have been "almost imprudent" as a fiduciary to refuse. Then he offered a metaphor that explains most of his financial thinking:
"Let's say we were a gold-mining business, or maybe just a person who had a whole bunch of gold. And gold prices are going crazy. Selling some of the gold now, so you have cash in the bank in case things change in the future, is not a bad idea. It's a hedge, and a hedge that's unbelievably good for us. It's a lot of money. In the case where everything turns to [expletive], we will look pretty smart."
Two months later, talking to USA Today, he went further, with the cheerful perversity of a man who had already survived one crash by quitting two weeks before it. "We have real dollars in the bank, so in a sense the best thing that could happen to us is a giant crash," he said. "I don't wish that to happen at all, but if it did suddenly we wouldn't have to pay $75 a square foot for our office, salaries for engineers wouldn't start at $130,000 a year, ad rates would go down and there'd be less competition. That's a far better scenario for us."
It was hard to tell, in those years, whether Butterfield was a very serious man doing an impression of a casual one or the reverse. His Twitter bio contained a shruggie, ¯\_(ツ)_/¯, and the line "I'm trying my hardest!" A Globe and Mail reporter who spent an hour at lunch with him in San Francisco watched him make the shrug with his body more than once. He spoke softly and swore constantly, "not for shock value, just as a matter of habit." "I'm definitely never going to have an opportunity this big again," he said, the way someone else might mention the weather. "Why not see how far we can take it?" The only thing that matched Slack's rise for scale was the plan he described, without raising his voice: to do for communication what Windows had done for software, and become the hub through which everything else worked together.
In November 2015 WSJ Magazine named him its technology innovator. On stage, in front of Brad Pitt and Karlie Kloss, he joked that he felt as if he had won the nerd award. Robert De Niro, who followed him, did not find this charming, and ended a short rant with "I don't give a f**k who you are." Butterfield's reaction went out on Twitter: "I just literally got trash-talked by Robert DeNiro. #Lifegoals." Inc. named Slack its company of the year. Vanity Fair put him on its New Establishment list. By early 2016 Slack had 369 employees, 2.3 million daily users, 675,000 paying customers, and $64 million in annual recurring revenue. It had outgrown its office three times.
He did not, apparently, get comfortable with the money. "In truth I feel guilty spending too much money," he told the BBC in 2018, when his fortune was estimated at $650 million. "As a Canadian that world seems very strange and alien to me." He put much of his success down to luck. The cabin had no plumbing, and the man who grew up in it still talked like someone who expected to have to haul water.
The pattern under the jokes was consistent. He took the capital when it was cheap and spent his attention on the product. For seventeen consecutive weeks, he said, Slack grew five percent a week, a rate at which "it's actually really difficult to keep up," because "every new person needs to be trained and that means someone else's time." The product had become so embedded in its customers' days that their abandonment of it could serve as a political signal. Scott Walker's presidential campaign team, he noted, stopped using Slack the day before Walker dropped out of the race.
Turning Money Into Customers
Every company Butterfield built was really a group of people who enjoyed being together, and every one eventually grew too big for that to stay true. By 2017 Slack had about a thousand employees. Onstage at SaaStr that year, he described the change with a candor that most chief executives of billion-dollar companies reserve for their therapists.
"We work really hard. We're smart, and we're trying to figure out how to turn money into customers," he said. "We haven't cracked that, the scaling the business through go-to-market programs. Almost all of the growth has been people liked it. They recommended it."
The company he was describing in 2017 was moving, deliberately, into enterprise software, the world of salespeople and procurement departments and multiyear contracts. In September 2017 it raised $250 million, mostly from SoftBank's Vision Fund, roughly 45 percent of whose money had come from Saudi Arabia's Public Investment Fund. The round brought Slack's total funding to $841 million at a $5.1 billion valuation. In early 2018 Slack hired its first chief financial officer, Allen Shim. On July 26, 2018, Atlassian announced it was shutting down its competing products, HipChat and Stride, and selling their intellectual property to Slack. Atlassian also took a minority stake. One of Slack's rivals had concluded it was better off as a shareholder.
Then came the question of how to go public, and Butterfield chose not to do it the standard way. Instead of a traditional initial public offering, Slack pursued a direct listing, as Spotify had in 2018. No new shares were sold and no underwriters set a price. Existing holders could simply begin trading. The S-1, filed April 26, 2019, showed revenue of $400.6 million for the year ending January 31, 2019, up from $220.5 million the year before and $105.2 million the year before that, along with losses of $138.9 million. It also created two classes of common stock. Class B shares carried extra votes and went to insiders, which preserved founder control in a structure that was otherwise built for openness. On June 20, 2019, the New York Stock Exchange set a reference price of $26 for the ticker WORK. Shares opened at $38.50 and rose past $41 within hours. Butterfield's stake was roughly 8 percent, worth about $1.3 billion even at the low end of expectations. Cal Henderson, the Ludicorp web developer who had ported the old game code to new languages for fun, owned about 3 percent.
The listing had an unexpected legal afterlife. A shareholder named Fiyyaz Pirani sued, claiming the registration statement was misleading. Because a direct listing releases registered and unregistered shares into the market at the same moment, Slack argued he could not prove which kind he had bought, and therefore could not sue under Section 11 of the Securities Act of 1933. A federal district judge disagreed. The Ninth Circuit disagreed too, in a split decision. In June 2023 the U.S. Supreme Court unanimously sided with Slack. On remand, in 2025, the Ninth Circuit ordered the complaint dismissed with prejudice. Butterfield's preference for doing things differently had left a mark on securities law.
The best years and the most dangerous ones overlapped. On February 10, 2020, IBM said it would put all 350,000 of its employees on Slack. Weeks later, the pandemic sent the world's office workers home and made Slack's argument for it. But Slack's real opponent had arrived years earlier: Microsoft Teams, bundled at no extra charge into the Office suite that nearly every large company already paid for. In July 2020 Slack filed a complaint with the European Commission accusing Microsoft of illegally bundling Teams with Office. For a company that had grown through word of mouth, being handed to every corporate employee for free, whether they wanted it or not, was the hardest kind of competition to answer.
On November 25, 2020, the Wall Street Journal reported that Salesforce was in advanced talks to buy Slack. The deal was announced on December 1 at $27.7 billion and closed on July 21, 2021. Slack was delisted, and shareholders received Salesforce stock. In its last full fiscal year as an independent company Slack had reported revenue of $903 million and 2,545 employees.
In December 2022 Butterfield announced he was leaving. He was gone from Salesforce early in January 2023. A few weeks later the company announced it would leave its headquarters at Foundry Square, which had been designed to make workers feel as if they were among waterfalls and glaciers, and move into Salesforce Tower by the end of February. In December 2025, after Slack's chief executive left for OpenAI, its chief product officer, Rob Seaman, was named interim CEO.
When Butterfield resurfaced in public, on a podcast in November 2025, he had a theory about what happens to organizations as they grow, and it sounded like it came from long, unhappy observation. At the beginning, he said, everyone arrives each morning knowing exactly what to do and knowing it will matter. Over time "the relationship between the supply of work to do and the demand for doing work just starts to change." People keep busy anyway. He called the result "hyper-realistic work-like activity."
"People are calling meetings with their colleagues to preview the deck that they're going to show in the big meeting, to get feedback on whether they should improve some of the slides," he said. "We are sitting in a conference room, and there's something being projected up there, and we're all talking about it, and that's exactly what work is." It wasn't, he insisted, because anyone was stupid or evil. People just wanted credit for something. "Hyper-realistic worklike activity is superficially identical to work," he said. "But this is actually a fake bit of work, and it's so subtle."
The man who had built the most successful workplace software of his decade had come to suspect that much of what happened in the workplace was a game nobody could win. He would know.
— Stewart Butterfield, Stanford GSB View From The Top, 2020Making anything significant happen requires dozens of people, or hundreds, or thousands. I've been learning how to do that without creating an entirely top-down environment.
Game Neverending
There is a coda to the first game. It comes from 2008, before Tiny Speck, before Glitch, before Slack, and it is the closest anyone has come to answering the question his whole career keeps asking.
On April 1, 2008, a few weeks before Butterfield announced his departure from Yahoo, Game Neverending came back online at an address on Flickr's servers. Cal Henderson and a colleague named Myles had spent a year porting the old code to PHP. The announcement was an April Fools' parody of Jerry Yang's internal Yahoo emails. People assumed it would last a day. It lasted a little longer.
Andy Baio, who writes the blog Waxy.org, played all day and reached level seven. He built a house in a neighborhood called Fierov Heights and saved enough "making points" to attempt the final item, a Game Neverending, the object whose completion was the closest thing the game had to winning. He was about $8 million short. The recipe required, among other ingredients, Stewart, Caterina, Ben, and five other Ludicorp employees, each purchased for $1 million in the back room of a mash pub.
So the players pooled their money. At about eight that evening, someone assembled a Game Neverending out of all of them: the founders, the designer, the staff, bought and combined like paper and shekels. Then the players passed it around so everyone could hold it. When everyone had held it, they handed it back to the player who had built it, so that she could win.
The next morning the game's administrator announced the servers would shut down within the hour. "GNE is a shared temporary hallucination," one player wrote. Baio built one last copy and won the game with four minutes to spare.
Aaand, it's gone.
Part IIThe Playbook
Butterfield never followed a method, in the sense of a procedure written down beforehand and carried out. What he has is closer to a set of habits that he formed through two failures and then wrote down after the fact, sometimes in memos, sometimes in podcast monologues, often as jokes. The principles below are taken from the record in Part I. They are not universal. Several of them depend on Butterfield's particular mix of game designer, philosopher, and reluctant executive. But each has evidence behind it, and each runs against something most founders are told.
The principles
- 1.Watch what people use while you build the thing you meant to build.
- 2.Make the kill decision coldly, because it will be humiliating.
- 3.Sell the change in behavior, not the equipment.
- 4.The less people know they want it, the more perfect it must be.
- 5.Solve for comprehension, not clicks.
- 6.Institutionalize embarrassment about your own product.
- 7.Treat capital as a commodity and sell when the price is crazy.
- 8.Let affection do the selling, then admit when it stops scaling.
- 9.Match the business to the capital, or give the capital back in effort.
- 10.Supply known valuable work, or people will invent fake work.
- 11.Use candor as a brand, and make it real.
↻
The Two Pivots, Side by Side
Butterfield made the same bet twice and was saved twice by the same kind of by-product.
| Dimension | Game Neverending → Flickr | Glitch → Slack |
|---|---|---|
| Years on the game | 2002–2004 | 2009–2012 |
| Capital | Government loan; near-zero payroll | $17.5M from angels, Accel, Andreessen Horowitz |
| Why the game failed | Risk No funding after the dot-com bust | Risk "Too foreign"; 15 minutes to understand; leaky bucket |
| What was salvaged | Scan In-game photo sharing | Scan The team's internal chat tool |
| Time to new product | Try First version built in eight weeks | Try Preview within nine months of the shutdown announcement |
| Outcome | Scale Sold to Yahoo for $22–25M, 2005 | Scale Sold to Salesforce for $27.7B, 2021 |
Principle 1
Watch what people use while you build the thing you meant to build
Neither of Butterfield's great products was planned. Both were tools his teams built so they could make something else, and both turned out to be the most valuable thing in the building. Flickr came from a photo-sharing feature in Game Neverending that barely worked. Slack came from the messaging system Tiny Speck's engineers built to coordinate across three cities. In each case the warning signs had been there for a while. The pivot was not a stroke of genius. It was a matter of finally noticing.
The lesson is not "have a backup plan." It's that in any ambitious project, the team's own working habits are a running experiment, and the by-products deserve the same attention as the main product. Butterfield's teams were unusually good at building internal tools because they were game designers, people used to building small systems for play. When the main bet failed, they found a viable product in the infrastructure that had made their days bearable.
The approach has a limit. It only works if the side products are good, and they are only good if the team cares about craft even in things nobody outside will see. Tiny Speck's chat tool was not a hack. It already had persistent history and search, the features that would define Slack.
Tactic: Every quarter, list the internal tools your team relies on daily and ask, seriously, which of them a stranger would pay for.
Principle 2
Make the kill decision coldly, because it will be humiliating
Butterfield is unusual among founders in how openly he describes the shame of failure. He wept in front of his staff. He called it "humiliating." He compared it to the destruction of a planet. And his advice on pivoting is built around that shame: "The reason I say you have to be coldly rational about it is because it's fucking humiliating." The emotion is precisely why the decision needs to be made at a distance.
His test is "have you exhausted the possibilities?" and Glitch passed it honestly. The game had paying players and a passionate core. "It would have been a fine lifestyle-business." What it could not do was become the kind of business that justified its funding. He shut it down not because it was dying but because it could not reach the size its capital required. That is a harder decision than giving up on a failure, and a rarer one.
How he shut it down matters too. Tiny Speck coached and counseled employees and introduced them to other employers. It released the game's art under Creative Commons. A failure handled that way kept the trust of the people who later became Slack.
Tactic: Before deciding whether to kill a project, write down in advance the specific evidence that would prove it cannot reach the scale its funding requires, then judge against that document, not against how you feel that week.
Principle 3
Sell the change in behavior, not the equipment
"We Don't Sell Saddles Here" argued that customers do not buy a tool. They buy what the tool turns them into. Butterfield's definition is strict: "The best — maybe the only? — real, direct measure of 'innovation' is change in human behaviour." By that definition a feature that changes nothing is not an innovation, however clever it is.
This changes how a company describes itself. Slack's competitors described group chat. Slack described ending internal email, reducing the number of times you're interrupted, and reaching a searchable archive of everything your organization knew. "Be less busy," the website said, and even Butterfield admitted that the promise was ahead of the product. But the pitch pointed at the outcome people wanted, not at the mechanics of the product.
It also explains Glitch's failure. The game asked players to change how they thought of themselves, as people who play a game without combat and without winning, and it gave them no clear reason to do so. The memo is that lesson turned into a rule for the next product.
Tactic: Write one sentence describing how your customer's day will look different six months after adopting your product, and remove any marketing copy that does not support that sentence.
Principle 4
The less people know they want it, the more perfect it must be
This is the subtlest idea in the memo and the most useful. "When you want something really bad, you will put up with a lot of flaws. But if you do not yet know you want something, your tolerance will be much lower." A product that creates a new habit gets no patience from its users. Every irritation is a reason to leave, because the user never wanted it in the first place.
This reverses the usual advice to ship something rough and improve it. Rough products work when demand already exists, when people are desperate enough to tolerate a bad solution. For new categories, Butterfield argued, you need "an exceptional, near-perfect job of execution," because "every bit of grace, refinement, and thoughtfulness on our part will pull people along."
Glitch's fifteen-minute learning curve is the proof. Its devoted players proved the game could be loved, but loving it required patience that newcomers didn't have.
Tactic: Classify your product honestly as either "already wanted" or "not yet wanted," and if it's the latter, put your polish budget into the first fifteen minutes of use.
Principle 5
Solve for comprehension, not clicks
The industry's obsession with removing friction, by counting clicks and taps and trimming steps, misses the real problem, Butterfield says. "It became an assumption that you should always be trying to remove friction when the challenge is really comprehension. If your software kind of stops me and asks me to make a decision, and I don't really understand it, you make me feel stupid."
The difference matters. A three-step flow the user understands beats a one-step flow that leaves them confused. What frustrates people is not effort. It's feeling stupid. "How do I prevent people from having to think in order to use my software?" is a different question from "how do I make this faster?", and it leads to different design choices: clearer language, fewer ambiguous options, defaults that show what the product means.
The philosophy student comes through here. Butterfield thinks about software in terms of mental states: understanding, confusion, the small humiliation of not knowing what to do. The point is the user's sense of competence, not the number of steps.
Tactic: In your next usability review, stop counting clicks and mark every moment a user hesitates, then fix the hesitation before you shorten the flow.
Principle 6
Institutionalize embarrassment about your own product
In 2014, with Slack growing faster than any business software in memory, Butterfield called it "a giant piece of shit" in a magazine interview. That was not false modesty. "If you can't see almost limitless opportunities to improve," he said later, "then you shouldn't be designing the product." He called a related failure the "owner's delusion": a creator's inability to see their product the way a stranger does.
The practical value of this attitude is that it keeps the company honest about the distance between its reputation and its product. In 2015, a fast-growing startup praised by every magazine had every reason to believe its own press. Butterfield kept describing Slack the way an irritated first-time user might.
There is a balancing practice. At more than one company all-hands, he made everyone repeat a chant: "In the long run, the measure of our success will be the amount of value that we create for customers." Embarrassment about the product combined with a clear measure of value is what makes the self-criticism productive rather than paralyzing.
Tactic: Once a quarter, have each leader write the harshest honest review of the product they can, and circulate it company-wide without rebuttals.
Principle 7
Treat capital as a commodity and sell when the price is crazy
Butterfield's financial thinking is consistent and unsentimental. Equity is gold, and when gold prices go crazy, you sell some. In April 2015 he took $160 million for about 5 percent of Slack while saying he had no immediate use for it. He called it a hedge, and listed what it would buy anyway: higher stock value, credibility with enterprise customers who wanted to know Slack would survive, and leverage in recruiting.
His most counterintuitive line, "the best thing that could happen to us is a giant crash," follows from the same reasoning. A company holding cash benefits when everyone else's costs fall. Office rent, engineering salaries, and ad rates all drop, and underfunded competitors disappear. Raising money in a bubble is not a bet that the bubble will continue. It's insurance against its end.
The contrast with 2002 explains why he thinks this way. Ludicorp couldn't raise money for a game because the market had collapsed, and its payroll shrank to the one employee with children. Someone who has lived through that period does not assume capital will be available when he needs it.
$
Conventional Wisdom vs. The Butterfield Approach
Where his habits run against standard startup advice.
| Conventional wisdom | Butterfield's approach |
|---|---|
| Raise only what you need. | Raise when prices are crazy; cash is a hedge against the crash. |
| Ship fast and rough; iterate. | For unwanted-yet products, near-perfect execution is the price of entry. |
| Remove friction everywhere. | Remove confusion; some steps are fine if they're understood. |
| Project confidence; you're "crushing it." | Shrug, swear, and call your own product a piece of shit. |
| Keep a business alive while it has revenue. | Kill it if it can't justify the capital behind it. |
| Go public with a traditional IPO. | Direct listing: no new shares, no underwriter pricing. |
Tactic: When your valuation looks absurd to you, raise a modest round at it and treat the money as insurance rather than a budget.
Principle 8
Let affection do the selling, then admit when it stops scaling
Slack grew from 10,000 to 90,000 daily users in five months with no sales team and no marketing team. Flickr grew because bloggers needed somewhere to host images. Both companies were spread by people who liked the product and told colleagues, the most valuable and least controllable kind of distribution there is.
What sets Butterfield apart is that he said publicly when this stopped being enough. At SaaStr in 2017, with a thousand employees and enterprise ambitions, he said Slack was still "trying to figure out how to turn money into customers," and "we haven't cracked that." The admission came before the company hired its first CFO, before it absorbed HipChat, and before Microsoft Teams arrived in every Office license. In hindsight, word of mouth had gotten Slack into companies, but it could not, alone, defeat a competitor that came free with the software everyone already paid for.
The lesson has two parts. Affection is the best distribution a product can have, and it doesn't grow in proportion to spending. A company that grows that way has to build a sales operation while the affection is still strong.
Tactic: Measure what share of new revenue comes from referral versus paid acquisition every month, and start building the paid side well before the referral share starts to fall.
Principle 9
Match the business to the capital, or give the capital back in effort
"It would have been a fine lifestyle-business, but it was never going to become the type of business that would justify seventeen and a half million dollars of venture capital." That sentence holds a whole theory of founder responsibility. Venture money brings expectations of scale. Accepting the money means accepting the expectations, and when the business can't meet them, the honest choice is to change direction rather than keep a small business alive on a large fund's money.
Butterfield presented the Glitch shutdown as responsibility to investors: "We could have kept going, but that wouldn't have been very responsible to our investors." And because Tiny Speck still had money in the bank when it pivoted, the investors who had financed the dead game ended up financing Slack. Accel and Andreessen Horowitz, who backed Glitch, took part in Slack's later rounds. Honesty about the first bet earned him the chance at a second one.
◷
The Decision Timeline
Key turns in Butterfield's career, including the ones that looked like endings.
2002
Ludicorp founded in Vancouver, just after Butterfield and Fake's honeymoon; Game Neverending prototype goes live that fall.
2004
Flickr launches February 10, built in eight weeks from the game's code.
2005
Yahoo acquires Ludicorp and Flickr on March 20.
2008
Butterfield leaves Yahoo July 12 after a mock-formal resignation letter.
2009
Tiny Spec, Inc. incorporated February 25; renamed Tiny Speck March 5.
2011
Glitch launches September 27; "unlaunches" November 30.
2012
Shutdown announced November 14; the game world closes December 9.
2014
Slack public release in February; company renamed Slack Technologies; $1.2B valuation in October.
2019
Direct listing on the NYSE, June 20, at a $26 reference price.
2021
Salesforce closes its $27.7B acquisition July 21.
2023
Butterfield leaves Salesforce in early January.
Tactic: At every fundraise, write down the minimum size the business must reach to justify the round, and revisit it each year as a promise you are honestly either keeping or breaking.
Principle 10
Supply known valuable work, or people will invent fake work
Butterfield's late-career diagnosis of big organizations is his most original management idea. Early on, everyone has "10 things to do and every single one of them is like something I know how to do, and it's definitely going to be valuable." As the company grows, the supply of obviously valuable work runs out before the supply of people does. The gap fills with "hyper-realistic work-like activity": meetings to preview the deck for the meeting, reviews of the reviews.
The important point is that this is nobody's fault in particular. Butterfield is explicit that people doing fake work are not "stupid" or "evil." They want credit for something, and in the absence of clear priorities they excel at whatever their team already does. The fix is not to criticize individuals. It is for leaders to deliberately supply what he calls "known valuable work to do": work whose value is clear before anyone starts.
This brings the story full circle. The man who built a tool that made it much easier to look busy at work ended up arguing that looking busy is the main way large organizations fail. The irony is not lost on him. "My guess is people who use Slack are reverting to the same level of busy-ness," he said in 2015, "but they're accomplishing more."
Tactic: Each week, make sure every team lead can name at least three tasks whose value is obvious in advance, and treat a team that can't as a staffing problem, not a performance problem.
Principle 11
Use candor as a brand, and make it real
The shruggie in his Twitter bio, the profanity, the joke about the big five philosophy firms, the line about being trash-talked by Robert De Niro: Butterfield's public persona is relaxed, honest, self-deprecating, and openly unsure. Fortune observed that he "rarely falls back onto pat talking points" and "provides actual answers to questions." That came across as a strategy precisely because it didn't seem like one.
Candor helped the business in specific ways. It made his praise of the product believable, because he was also willing to call it terrible. It made his capital raises seem prudent rather than greedy, because he explained the reasoning openly. It let him say in public that Slack hadn't figured out enterprise sales without the remark becoming a crisis. And it attracted the kind of employee who wants to work for someone who tells the truth. He looked for empathy in every hire and tried to build it into the software, in his words "designing the site for politeness and courtesy."
The risk is obvious: casual candor can turn into an act. What kept Butterfield's real was that he backed the words with costly decisions, like shutting down a game with paying customers, raising money he said he didn't need, and listing shares without the usual protections.
Tactic: Once a month, say something publicly about your company's weaknesses that a competitor could quote back at you, and make sure you can show what you're doing about it.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— Stewart Butterfield, "We Don't Sell Saddles Here," 2014The best — maybe the only? — real, direct measure of "innovation" is change in human behaviour.
— Stewart Butterfield, Lenny's Podcast, 2025In the long run, the measure of our success will be the amount of value that we create for customers. Then you can put effort into demonstrating that you have created this value and stuff like that, but there's no substitute for actually having created it.
— Stewart Butterfield to Gamasutra, on the closing of Glitch, November 2012The people who loved the game really, really loved it. If we had figured out an easier way to get more people to that state it would have been a success.
— Stewart Butterfield to USA Today, June 2015We have real dollars in the bank, so in a sense the best thing that could happen to us is a giant crash.
— Stewart Butterfield to The Globe and Mail, 2014I'm definitely never going to have an opportunity this big again. Why not see how far we can take it?
Maxims
- The by-product is the product. The tool your team builds to survive the project may outlast the project.
- Decide coldly because it burns. The humiliation of a pivot is exactly why it must be decided at arm's length.
- Sell the riding, not the saddle. Customers buy who they will become, not what you made.
- Unwanted products must be perfect. People forgive flaws only in things they already crave.
- Confusion costs more than clicks. Software that makes users feel stupid loses them faster than software that makes them work.
- Stay embarrassed. If you can't see limitless room to improve, you shouldn't be the one designing it.
- Sell gold when gold is crazy. Raise in the bubble as insurance against the crash.
- Honor the money's expectations. A fine small business can still be the wrong business for venture capital.
- Feed people real work. When known valuable work runs out, fake work fills the gap, and it looks exactly like the real thing.
- Mean the shrug. Candor only works as a brand if it is backed by decisions that cost you something.
