Part IThe Story
The order came in on the fax machine, overnight, in 2003. It arrived the way most things reached Sydney from America then: while the recipients were asleep. It was a purchase order from American Airlines for Jira, a bug tracker made by a company that consisted of two twenty-three-year-olds, a credit card balance, and a website. Nobody at the airline had called them, and they had not called the airline. Atlassian's own retelling, polished by two decades of repetition, preserves what the founders said to each other that morning:
"Did you talk to them?"
"No, did you?"
"Holy shit! American Airlines saw Jira on our website and bought it just like that!"
Scott Farquhar later called it the moment he knew they would make it. "That was the turning point," he told the BBC in 2019. It ended what he described as "hand-to-hand combat" for customers, the period when orders came mostly from people they already knew. By 2019 the fax hung outside his office at the company's Sydney headquarters. It is an odd thing to put on a wall. Most founders who frame a document choose a cheque, a term sheet, or the first dollar bill. Farquhar kept proof of a sale in which neither founder did anything at all.
Atlassian would grow on that idea. In the company's own words, software "should be bought, not sold." Under Farquhar it became a business with $102 million in annual revenue and no salespeople, then a Nasdaq listing worth $4.37 billion on December 10, 2015, then a company worth more than $50 billion that tracked the work behind Mars rovers, rocket launches, and Domino's deliveries. It made Farquhar and Mike Cannon-Brookes, his university friend, Australia's first technology billionaires. The New York Times called it "a very boring software company," and Farquhar seems to have taken that as a compliment.
The fax is worth starting with because Farquhar's life keeps turning on paper that arrives without being asked for, or fails to arrive when it was expected. There was an acceptance letter that came two months late. There was an email from Cannon-Brookes to an entire class, which only Farquhar answered. There was a résumé with the dates left off. There was a school essay by a twelve-year-old, which the man read aloud thirty-four years later in Canberra. There was a resignation letter that thanked a partner for being unreasonable. In 2026 there was a donation record that confused him with another man of the same name in Brisbane. He has spent most of his career in the background of his own story, as the second half of "Mike and Scott," and the documents tell it better than he has.
By the Numbers
The Farquhar Ledger
$10,000Credit card debt that financed Atlassian's founding
A$48,500The graduate bank salary the founders set out to match
$1,000Early Atlassian price point, against a $100,000 enterprise norm
$4.37BMarket capitalisation at the December 10, 2015 Nasdaq IPO
23 yearsFrom founding to his last day as co-CEO, August 31, 2024
19,000Companies signed to Pledge 1%, the movement he co-founded
US$5.22BAtlassian revenue, fiscal 2025
A Letter That Came Late
The computer was a Wang. Farquhar's father got it through work, sometime in the early 1990s, after his eldest son had spent months asking for one. The recession was hitting western Sydney hard, the family had four children, and money was short. "I remember emotionally blackmailing my parents and at one stage crying to get a computer," Farquhar said in 2014. "I feel horrible now, because my parents couldn't really afford one." Another time he said he cried himself to sleep because a friend had a computer and he did not. The machine his father brought home was so old it would not run Microsoft DOS. Farquhar spent close to a year trying to get it to play games and never succeeded.
That year is easy to skip past. A boy with a machine that would not work kept at it for twelve months, against a problem that had no solution, with no one to ask. It is an early sign of the persistence that later looked like ordinary diligence in an adult. In later interviews he did not present it as an origin story. He presented it as a failure he had stuck with.
He was born in Parramatta in December 1979, grew up around Castle Hill, shared a bedroom with his brother, and went to Castle Hill Public School. In 1991 his teacher, Mrs Crawford, asked the class to imagine the future. Thirty-four years later he could still quote from what he wrote:
"In the future adults will no longer get up at all hours of the morning to go to work. Instead, they will wake up and walk over to the home computer. The computer will be linked to a main computer at their work."
He also predicted factories run by one person and thousands of machines, which would "make computer parts, cars, televisions, even other robots!" And he predicted something he called "sinoleums," ten-storey zero-gravity entertainment halls where people could bounce around for fun. None have been built. As an adult, in what he called "a fit of optimism," he registered sinoleum.com.
He won his primary school's technology prize in Year 6. Then he went to James Ruse Agricultural High School, the selective public school in Carlingford that has been at or near the top of New South Wales' HSC rankings for a generation. He later described its computing education as terrible, "to the point of non-existent." His next real contact with computers came at university. At fourteen he wrote a résumé to get a job at Woolworths. He was in the Scouts for about a decade and repeated first-aid courses along the way, practising CPR on dummies. That detail matters later in this story.
When he finished high school in 1998, he was on course for the Australian Defence Force Academy in Canberra, which trains officers for the army, navy, and air force. The offer letter was lost in the mail. By the time it reached his parents' house, about two months late, he had accepted a place at the University of New South Wales.
— Scott Farquhar, to the BBC, 2019If the letter had turned up earlier, history may be slightly different.
Farquhar told this to the BBC as a sliding-doors anecdote. Read alongside the rest of his life, it says more. He had been heading for an institution with fixed ranks, uniforms, and a promotion ladder: a place where you are posted and assessed, and where you never sell anything because everything is assigned. He ended up building a company with no ranks to speak of, no dress code, and no sales force, because the post office lost an envelope.
The Only Reply
Mike Cannon-Brookes had a different start. He was born in New Haven, Connecticut, on November 17, 1979, about a month before Farquhar. His father, also Michael, was a British banker who came to Australia in 1984 to set up Citibank's operation and run it as managing director. The family lived in Taiwan when Mike was six months old and in Hong Kong when he was three. He was sent to boarding school in England and flew to Australia four times a year to see his family. He bought his first computer, an Amstrad PC20 that he later called "a real hunk of junk," with Qantas frequent flyer points from those trips. He was eight or nine. "Because of the flying, I had tons of Frequent Flyer points and back then points were worth something," he said. "So I got the catalogue every year and got some shit." He persuaded his parents to put the internet in the house in 1991, before the World Wide Web existed. He finished school at Cranbrook, one of Sydney's most expensive boys' schools.
So one boy got a computer from his father's employer and it never worked. The other got one out of an airline catalogue. Both ended up in 1998 in the same co-op scholarship course in business information technology at UNSW, a cohort of about forty students with industry placements built into the degree. They met on the first day and did not get along. "People from Cranbrook," Farquhar said later, "had a bit of a swagger about them before they started university."
That changed. "We came from all walks of life but we were all nerds who loved computers," said Niki Scevak, a classmate who co-founded a bookmark-management startup called The Bookmark Box with Cannon-Brookes and sold it to Blink.com in 2000. So Cannon-Brookes had already started and sold a company before he graduated. Farquhar had a part-time job and a shared house.
In 2001 Cannon-Brookes emailed his classmates to ask whether anyone wanted to start a technology company with him after graduation. Farquhar was the only one who replied. The rest of the story depends on that, and none of the coverage explains why only one person answered. The likeliest reason is that in 2001 Australia had almost no startup culture, and starting one straight out of university looked to most people like a polite way of saying you were unemployed. Everyone else in the class did the sensible thing. Farquhar was the exception.
The ambition was small. Most programming jobs in Sydney were in banks, which Farquhar later called "a bit too buttoned-down for our tastes." They had two goals. They did not want to wear suits, and they wanted to earn at least A$48,500 a year, the starting salary banks and accounting firms were offering graduates. "We figured if we could code for a living without having to wear a tie, that'd be pretty sweet," Farquhar wrote in 2017. He was living in a share house at the time. "At that stage I was living in a shared house at university and eating ramen noodles every day," he told the BBC. "We didn't have much to lose."
They financed the company with $10,000 on credit cards. Atlassian was registered as a business name in 2001. The name came from Atlas, the Titan condemned to hold up the sky, specifically the bronze statue of him at Rockefeller Center in New York. From 2011 until a rebrand in 2017 the logo was a blue X-shaped figure holding up the bottom of the sky. Both men grew up in Sydney, yet the company name came from a statue in Manhattan, and the founders were looking toward an American market from the start.
They were each other's best man. When Farquhar was asked in 2014 whether he had worried about personal liability in the early years, he said: "I think if we were bankrupt they'd have come for my Converse shoes but there wouldn't be much more than that they could've got from me."
How Old Are You?
The first version of Atlassian was a support business. Cannon-Brookes and Farquhar took customer-service calls for other companies' software, at any hour, from their bedrooms. Farquhar later called it "terrible," and Atlassian's official history is not much kinder. It was also the founders' first experience of the conventional services business that its founders would spend the rest of their careers designing out of existence.
It produced one story Farquhar still tells. One of their support clients liked their answers enough to call and ask Farquhar to come to Amsterdam and help rescue a fifty-person software project that was running late. He needed the money. "Despite not having any previous experience at this," he recalled in a 2011 graduation address at UNSW, "I said of course I could, not a problem!"
The day before he flew out, the client asked him to send a résumé.
"I said 'Sure, not a problem – I'll send it through straight away.' I remember thinking – 'I'm 22 years old, I haven't graduated from university, the closest I had been to a software project was group assignment at university, and the last time I wrote a resume was to get a job at Woolworths when I was 14.'"
He sent one with the dates left off. In the first week he was so useful that the client sat him down and asked why he was doing so much better than the consultants they had paid before. On the fourth day the boss got him alone and asked: "We noticed you didn't have any dates on your resume — how old are you?"
"At that moment," Farquhar told the graduates, "the question wasn't whether I would tell the truth or not, but instead, how much was I going to lie?"
The account that survives, a summary posted by UNSW's Centre for Innovation and Entrepreneurship on March 28, 2011, stops partway through his answer: "I stammered – 'Ah, ah, I'm —" I have not been able to find the rest. A ceremony full of new graduates presumably heard how it ended. The record we have stops with a twenty-two-year-old who had already done the job well and was about to be found out for being too young.
The support work ended because of the tools rather than the clients. The bug tracker they used was poor, and the market leader, Bugzilla, was not much better. So they wrote their own. They called it Jira, a shortened form of Gojira, the Japanese title of the 1954 Godzilla film, as a joke at the expense of Bugzilla. Jira 1.0 shipped in 2002. They quickly decided, as the company later put it, that they were "better off selling Jira than providing support services." Atlassian's later summary of the lesson was "Building your own products beats fixing other people's products." That understates how big the change was. They stopped renting out their time and started selling something that could be copied for nothing.
Confluence, a corporate wiki, followed on March 25, 2004. The release notes said the aim was a tool built "to the requirements of an enterprise knowledge management system, without losing the essential, powerful simplicity of the wiki in the process." The company's whole idea is in that sentence: enterprise requirements and simplicity in the same product, sold to large companies without their usual sales process. Its first trade-press coverage described a product that the authors of Wikis for Dummies later called "one of the most popular wikis in corporate environments."
Around then Atlassian opened a small office in New York, where most of its customers were. It did not last. The company's history dates it to 2004 and press accounts to 2005, and both say it was shut quickly. "Closing that office so soon was hard," the company wrote later, "but it was good practice in making tough calls — of which there would be many to follow." The US presence moved to San Francisco, which had more engineers to hire.
Bought, Not Sold
Plenty of people have explained Atlassian's sales model. Farquhar's own account, given to Bessemer Venture Partners in 2021, is the clearest, because it shows the model following from a few constraints the way a proof follows from its axioms:
"If we wanted to sell online in every time zone of the world, it needed to be self-service. If it's self-service, the product needs to be easy to download and evaluate and install. If it needs to be all those things, then we needed to have a very reasonable price because we didn't have a salesperson to convince you to hand over six figures."
The starting assumption was that the business would be global, because Australia was too small to support it. Global meant selling in every time zone, which meant the sale had to work while the founders slept. That required a self-service purchase. A self-service purchase required a product that customers could install themselves, and a price low enough that nobody had to negotiate it. The founders then turned that chain of constraints into a principle. "The thing that we believed that was different from our parents," Farquhar said, "was that the world was going to be flat — you'd be able to reach customers anywhere."
The phrase "a very reasonable price" will come up again in this piece. In practice it meant this, as he told UNSW graduates in 2011: "So instead of selling our software for $100,000, we sold it for $1,000. Of course, you need to sell a lot of thousand dollar software to make a real business. And if you're going to sell lots of software, you need to sell to more than just Australia. So we decided to market our products online, and buy ads on this quaint little search engine called Google."
People told them it would not work. "There were a million reasons that people told us it wouldn't work. 'People won't buy without talking to a sales person.' 'People don't want to use their credit cards online.' 'You can't sell to someone without calling them up!'" His explanation of why they kept going is very dry: "Luckily, we didn't have any money to hire expensive sales people, so we decided to keep doing what we were doing."
The other part of the thinking concerned who had to approve a purchase. Below a certain amount, an engineer could put it on a credit card without asking a manager. Atlassian priced under that line on purpose. The buyer was a developer who had already installed the free version, used it, and come to rely on it. By the time the invoice reached procurement, the product was already part of how the team worked. A newsletter writer later reported, and it fits everything else known about the period, that the founders' early survival rule was "Sell one piece of software a week, no matter what, and we'll live."
It worked better than anyone expected, including them. Atlassian got its 1,000th customer in 2005. By 2011 revenue was $102 million, a 35 percent increase, and TechCrunch's headline noted it had been done "With No Sales People." In 2014 the Wall Street Journal valued the company at $3.3 billion with the same caveat. In 2016 Atlassian still had no traditional sales force and was putting the money into research and development. The founders later called what they had done "Product Led Growth," a term invented years afterward by people who had studied them.
Farquhar's own summary to the BBC mentions a location that becomes significant later. He said they took business software sales "off the golf course" and put them directly in front of customers.
Two Chairs, One Company
By the time Fortune put them on its 40 Under 40 list in 2016, both thirty-six, the two founders had split the work in a way the magazine summed up in one sentence: "Farquhar heads engineering and research; Cannon-Brookes, product design and sales and marketing." So Cannon-Brookes was in charge of sales at a company that prided itself on having no salespeople, and Farquhar was in charge of the engineers who made that possible.
Most co-CEO arrangements fail, and fail publicly. This one lasted about twenty-two years, as long as almost any in modern business. A growth newsletter later reported that disagreements between the two at a multi-billion-dollar company were sometimes settled with rock-paper-scissors. That is entertaining, and it also says something serious: when two people with equal stakes deadlock, a random rule that both trust is better than a hierarchy that would end the partnership. In 2006 Ernst & Young named them Australian Entrepreneurs of the Year, and they were the youngest winners ever. They were twenty-six.
Kim Jackson enters the story in 2009, when she married Farquhar. Like the founders, she came to her trade through engineering. She is a systems engineer who became an investment banker and worked on infrastructure deals for hospitals, airports, and toll roads. She now runs Skip Capital, the family's investment vehicle, and heads the Skip Foundation. "My wife runs Skip Capital," Farquhar told Forbes Australia in 2024. "I help her out occasionally, but she's the one that runs all of that." In 2017 she was the buyer of record on the most expensive house the couple ever bought. Farquhar is consistent on this point. The person in the background of his account is usually him.
The honeymoon was cut short. In 2010, as Farquhar told Harry Stebbings on 20VC, a hacker got into Atlassian's systems. He flew back to Australia and ran the investigation while staff slept in the office. Atlassian was about to raise money from venture capitalists for the first time. The obvious choice was to keep quiet until the round closed. Farquhar told investors and customers before the meetings went ahead.
The round closed anyway. In July 2010 Accel Partners put $60 million into Atlassian, and the money went to buying existing shares rather than to the company. It was the founders' first outside capital, eight years after they started. A 20VC episode title later said it eventually earned Accel about $6 billion. Rich Wong, the Accel partner who joined the board, worked with Farquhar for fourteen years. "From the beginning, it was obvious that Scott was a great entrepreneur," Wong said in 2024, "but it's been an absolute pleasure to see him grow into a consummate professional and a great CEO."
In July 2012 Doug Burgum became chairman. The Register asked in a headline whether Atlassian was "heading for the exit," noting that the new directors had experience selling software companies to large buyers. It was not. In 2014 the company redomiciled its parent to the United Kingdom ahead of an IPO, and Farquhar had to defend the move in the Financial Review. In 2022 the parent moved again, to Delaware. The headquarters stayed in Sydney through both moves. The corporate address changed for investors while the company stayed where it was.
The Seed and the Oak
On a Monday night in October 2014 Farquhar gave the JJC Bradfield Lecture in Sydney, named after the civil engineer whose work, as the lecture's program put it, transformed the city in the 1920s and 1930s. Farquhar did not often give speeches. He began with electricity and the Ford plant on Piquette Avenue in 1908, which could build only eleven Model Ts in its first month. Steam-driven factories ran every machine off belts and pulleys from one engine, so machines were arranged by their power needs rather than by the order of work. Electricity let Ford give each worker a separate motor and lay the factory out as a line. Production time per car fell from twelve hours to ninety-three minutes.
"Software isn't just about software anymore," he said. "It's about everything we do." Companies now came in two kinds: "either becoming a software company, or being disrupted by one." He quoted Archimedes on the lever. Then he turned to Australia and said something that was at once a boast and a warning.
— Scott Farquhar, JJC Bradfield Lecture, October 2014We 'made it' from suburban Sydney. That's the good news. The bad news is that any other 'Mike and Scott' from any suburb in Seattle or Shanghai, Lucknow or London or Hanoi or Helsinki can make it too.
The flat world that had let two Sydney graduates sell to American Airlines would let anyone sell to anyone. Australia's distance from everything had protected local businesses and also limited them. Now the protection was gone. In the question-and-answer session he explained this with a gardening comparison he kept using for the next ten years, sometimes describing company-building itself as gardening. A small company called SafetyCulture, in Townsville, in which he had invested, was doing very well with a couple of dozen people. But when it needed a head of engineering, there would be no one in Townsville with the experience. It would have to move.
"So, I think it's easy for a seed to germinate almost anywhere; there are plenty of conditions for that to happen. There are fewer conditions for that to turn into a huge oak tree, you'd need some pretty special conditions and you do need that concentration of talent."
He was surprisingly blunt about his own industry's lobbying. Reading the submissions to a national curriculum review, he noticed that the Australian Computer Society's case for teaching computing was weaker than the arts lobby's case for making arts compulsory. "We didn't do a very good job of articulating our case," he said. And then, about government industrial policy, he said something unexpected from a businessman who had never taken a subsidy: "we should pick some winners."
He kept at it. On ABC's Lateline in September 2015 he said Vietnam started teaching computer science in Year Four and kept it compulsory through high school, and that he had just come back from there. He pressed for a technology hub at the old Australian Technology Park in Redfern. In 2017 he criticised the government's overhaul of the 457 skilled visa. In April 2019, when Parliament passed laws after the Christchurch massacre that could jail tech executives over violent content, he tweeted: "In its blind rush to legislate, the government is creating confusion and threatening jobs." The Attorney-General, Christian Porter, replied that tech executives always threatened to leave and never did: "It's like Alec Baldwin saying he's going to leave America every time a Republican is elected president." It is the only time in the record that someone in power made fun of Farquhar to his face, and the joke worked because it was accurate. Atlassian stayed.
By then the hub he had asked for was being built. In August 2018 the NSW government announced a technology precinct around Central and Redfern stations and brought Farquhar in to help shape it. Atlassian signed on as anchor tenant, and in October 2021 it got approval for a headquarters planned as the world's tallest hybrid timber tower. A man who had used an oak tree as his image of a mature company would have offices in a building made of wood.
In 2021 the Financial Review ran a headline reporting that Farquhar was now telling founders not to bootstrap. The man who had built a company on $10,000 of credit card debt, and who was turned down by banks for even small loans as a student, was advising the next generation not to do what he had done. This fits the oak-tree argument. Bootstrapping had been forced on Atlassian because the capital did not exist in Australia. He now saw his job as making sure it existed. Later he said he measured his impact by the people who left Atlassian to start companies of their own: Dovetail, Sajari, Pyn. He talked about an "Atlassian mafia" on the model of PayPal's. "I couldn't be more proud," he said.
One Percent of Nothing
Early on, Atlassian committed to giving away one percent of its equity, one percent of its profit, one percent of employee time, and one percent of its product. Farquhar later said he and Cannon-Brookes had "probably read an article" about Marc Benioff doing the same at Salesforce in 2000. "Donating one per cent of equity was easy to do back when we started," he said in another interview, "since it was one per cent of virtually nothing."
He kept making that joke because it was the point. "1% of product isn't worth anything when you don't have any product to give away," he told Forbes Australia, "and we didn't have profit for a long time." The pledge was cheapest when it mattered least, and it got more expensive at the same rate as the company's success. It worked like an option written in favour of other people. The Atlassian Foundation was set up in 2006. By the time Atlassian had about 700 employees and $150 million in revenue, the foundation had given away US$3.7 million, mostly for education, "because education has been proven to break the cycle of poverty."
In about 2014 Farquhar noticed that hardly anyone had copied it. "I looked around and it was still just Salesforce, us, and a smattering of companies that had done this." He decided the idea had a distribution problem. "Even though Salesforce had pioneered it, that was as far as they'd got. They hadn't evangelized it to anyone else." So he took the method he had used to sell Jira, a cheap and self-serve commitment that spreads by word of mouth, and applied it to philanthropy. He pulled together a few people from Atlassian, a few from Salesforce, and Seth Levine, a Colorado investor, to run a nonprofit called Pledge 1%. "Our original goal was to get 500 people signed up," Farquhar said. "We would do all the startup hustle stuff. We'd turn up to conferences, we'd pack growth-hack stuff to get people to hear about this organisation. It was on an oily rag."
Ten years later, Pledge 1% had 19,000 member companies in more than 130 countries, including more than sixty unicorns, committed to giving about US$3 billion in equity. More than 1,800 were Australian, the second-largest national group. Canva and Culture Amp were members. In 2024 Farquhar came back as chair. Asked by Bessemer what achievement mattered most to him, he named this one, above Atlassian.
The IPO came in December 2015. Atlassian had been profitable for ten straight years and reported revenue of $319.5 million for fiscal 2015. Its shares were priced at $21 and closed their first day around $28.05, up more than a third, valuing the company at about $5.85 billion. The ticker was TEAM. Twelve employees who had been with the company for more than ten years stood at the bell.
Farquhar did not know his wife was bringing their three sons from Sydney. When they appeared in New York he was, in his words, "wiping back the tears." Afterward the family went to Times Square for greasy pizza and cheap champagne. "We hadn't thought about the celebration part of it," he told the BBC. Bessemer's account adds a separate scene: the two founders at a hole-in-the-wall pizza place in New York after weeks of roadshow, eating a slice with root beer, when it occurred to them that together they were worth more than a billion dollars. The planning had covered everything except what to do once it worked.
"This is a marathon, not a sprint," he told SmartCompany that week. "We've been doing this for 13 years and there will be another 13 years ahead of us."
He stayed for nine of them.
The Fence at Point Piper
In 2017 the Fairfax family sold Elaine.
The Point Piper estate, 6,986 square metres on Sydney Harbour, had belonged to the Fairfaxes since 1891. That is longer than Australia has been a federation. The Fairfaxes owned the Sydney Morning Herald and in 1951 founded the Australian Financial Review, the paper that later reported on Atlassian's founders more closely than any other. Elaine had been empty for close to twenty years. John B. Fairfax sold it to the Farquhars for about $71 million, which reports described as including a sentimental discount, and it was a national record. The buyer of record was Kim Jackson. The next year Cannon-Brookes bought Fairwater, the estate next door that had belonged to Lady Mary Fairfax, for about $100 million, another record. The Daily Mirror reported, with some amazement, that the two co-founders now owned the two most expensive houses in Australia and lived next to each other.
A boy from the western suburbs who had begged for a computer and an English banker's son raised in Asia now owned the waterfront estates of one of Australia's oldest media families. The change in who held Australian wealth was visible in the property records of one harbourside street.
The rest of the story is less clear, and the coverage contradicts itself. In 2020 the Farquhars filed plans to knock down unsympathetic later additions to Elaine and build a $30 million contemporary house. According to the Financial Review, citing well-placed sources, those plans were derailed by "over-the-fence objections" from Elaine's only direct neighbour, Fairwater, owned by Mike and Annie Cannon-Brookes. John B. Fairfax told the paper he had known where the pushback came from. The plans were dropped. On October 4, 2024, about five weeks after Farquhar's last day as co-CEO, Elaine sold for A$130 million.
In 2024 the Sydney Morning Herald reported rumours that "Scott and Mike had finally broken up," and dismissed them: "The truth is much more boring." In August 2026 HarperCollins published Tech Bros, by the business journalists John Stensholt and Perry Williams, whose own description of the book says that "disputes over strategy, property and ambition destroyed their friendship." Neither founder cooperated with the authors, and the book says so. One reviewer found its account of the falling-out "largely third-hand and a little lifeless." In July 2025, a year before the book came out, Cannon-Brookes was in the room in Canberra to hear Farquhar speak at the National Press Club. All of these accounts are on record, and they do not agree. I cannot resolve them, and the people who could have chosen not to.
What is on record from the two men themselves is short. When Farquhar announced his resignation, Cannon-Brookes posted on X: "Mate. Thank you. It's been one hell of a ride. Couldn't be prouder." Farquhar's farewell ended with a line to his co-founder: "You really are George Bernard Shaw's Unreasonable Man, and the world and I are better as a result." Shaw's line is that the reasonable man adapts himself to the world and the unreasonable man insists the world adapt to him, so all progress depends on the unreasonable man. If Mike was the unreasonable one, Farquhar was naming his own role as well: the one who answered the email, set the reasonable price, ran the engineers, and kept things working.
The tabloids missed a smaller story. In 2022 the Farquhars bought a $10.85 million beachfront house on the Central Coast. The Financial Review's property writer, describing it, said the couple were known for "a somewhat low-key approach to their discretionary property purchases — at least among Australia's billionaire class." She added: "it's all relative."
Well, You Do It Then
In April 2022 Farquhar was in Las Vegas for Team '22, Atlassian's flagship conference, which opened on April 5 with Bob Iger among the speakers. Three nights before it started, he went to dinner with a friend who had moved from Australia to the United States, and afterward they got a table by the dance floor at Omnia, the nightclub at Caesars Palace. The club was packed and then thinned out. In the early morning Farquhar crossed the dance floor toward the bathroom and saw a man lying on his back, not moving.
To Farquhar, the man looked dead. He knelt beside him and put his cheek near the man's face to check for breath. There was none, and his chest was not moving. Farquhar started CPR, which he had practised many times over a decade in the Scouts, always on dummies. Club bouncers came over and told him to stop.
"Well, you do it then," Farquhar said.
They told him to keep going. The music stopped. People gathered to watch. Some shouted at him and some tried to help. He heard gurgling. He told CNBC later that it was overwhelming. Then, in CNBC's account, "A man in plastic gloves" arrived, and the excerpt available to me ends there. The public record shows that the man lived.
Farquhar did not publicise this. CNBC heard about it from someone with knowledge of the incident while reporting a different story, and Farquhar confirmed it only when asked. He then spoke about it in detail for one interview and did not bring it up again.
The scene fits what he had been saying for years. "The day of the lone genius is dead," he told the StartCon conference in Sydney in December 2017. In that speech he described the steelworkers in Pennsylvania studied a century earlier, whose foremen worked out the ideal weight of a shovel load and quadrupled their output. He said modern teams "absolutely should not" optimise that way, because "who here goes to work every day knowing exactly what they're going to do?" He added: "The days where the boss is the smartest person in the room are gone. My team reminds me of that all the time." Earlier that year, in an open letter to computer science graduates, he wrote: "My first professional team was a team of two. Just me and my best mate, Mike Cannon-Brookes... There's no way I could do this on my own. Not in a hundred lifetimes."
In Las Vegas he was alone with a dying stranger, and the people with official authority told him to stop. His answer amounted to: then you take over. When they would not, he kept going. His theory of teams has a limit, and it is this: nobody else would take over. Someone has to begin, and then keep going while others decide whether to help.
The same year he made the hardest product decision of his tenure. In September 2017 Atlassian launched Stride, a workplace chat product meant to compete with Slack and to replace HipChat, which it had bought five years earlier. Less than a year later, on July 26, 2018, Atlassian sold the intellectual property for both products to Slack, took a small stake in its competitor, and announced it would shut both down. Farquhar later explained the decision to Harry Stebbings:
— Scott Farquhar, on shutting down Stride, 20VCI couldn't look my staff in the eye and say the next five years of their life was going to be the biggest impact to humanity working on this product.
He judged the decision by what he could honestly say to the people building the product, not by revenue or market share. "We've launched products too late," he told the BBC the next year. "Recently we had to shut down a product that didn't make it in the market." He also told Stebbings that the nature of his mistakes changed as the company grew. A small company makes mistakes by doing the wrong thing. A large one makes them by failing to act. "My biggest mistakes," he said, "are being in the right place at the right time but not going hard enough behind the opportunities that we've had in front of us."
When the pandemic arrived, he made his biggest commitment of that era. In August 2020 Atlassian announced Team Anywhere, which let employees live and work anywhere the company had a legal entity. By April 2022 about 300 staff had moved countries. In Australia, the share of staff living outside Sydney rose from 7 percent to 22 percent. "Remote work is one of the best things that could happen to Queensland," he said in Brisbane in July 2023, where Atlassian had 170 employees and no office. He argued with Elon Musk on Twitter about Musk's "1950s" view of the office. Stewart Butterfield, Slack's CEO, said Atlassian's policy influenced his own decision to move to Aspen. Atlassian's internal data, Farquhar said, showed that deliberate in-person gatherings, which he called "intentional togetherness," raised team connection by about 30 percent for several months before it wore off. As he saw it, the fix for distance was to bring people together periodically and accept that the effect would fade.
The share price fell by half in 2022. In March 2023 the company laid off 500 people, 5 percent of its staff. Farquhar kept Team Anywhere.
Megawatts as Megabytes
In November 2023 Farquhar told the Financial Review he was as excited to run Atlassian as he had been in the early years. The headline mentioned a 100-year plan. Five and a half months later, on April 25, 2024, he posted "The journey of a lifetime" on the company blog.
"It's with a full but heavy heart that I share with you all today my decision to step down as Atlassian's co-CEO." He set out what the company had done: it had started product-led growth, it ran rockets and Mars landers and hospital procedures and pizza deliveries, it had 300,000 cloud customers. He said he wanted time with his young family, philanthropy through the Skip Foundation and Pledge 1%, investing through Skip Capital, and mentoring other tech CEOs. His last day would be August 31, 2024. He would stay on the board as a special adviser. He was forty-four. "I'm super blessed," he said later, "to be moving into the next phase of life beyond working 80 hours a week."
The Financial Review's headline on September 1 was "Farquhar's Atlassian era passes without a grand farewell." Nick Crocker of Blackbird Ventures wrote on LinkedIn: "From a kid scout to a globally admired CEO… bravo."
Several of the things he did afterward fit the earlier story almost too neatly. In June 2024, between announcing his resignation and leaving, Farquhar applied to join the Royal Sydney Golf Club. The Financial Review's gossip column covered it and noted that he had grown up sharing a bedroom and attending public schools. The man who had told the BBC he took software sales off the golf course was now applying to join one of the most exclusive golf clubs in the country. In Sydney's harbourside suburbs, as the columnist pointed out, money alone does not get you a membership.
He told Forbes Australia he was "aggressively holidaying." He signed up for an Ironman, which he called "the silly thing." He became part-owner of Factor Bikes, a British maker of high-end bicycles that supplied a $100,000 track bike to the Australian team chasing Olympic gold in Paris. Skip Capital, run by Kim Jackson, bought stakes in four airports and looked at waste, water, and data centres. He succeeded Robyn Denholm as chair of the Tech Council of Australia.
Then, on July 29, 2025, he gave the National Press Club address in Canberra. A Forbes reporter described his clothes: black t-shirt, blue jeans, an Oura ring on his right hand, and black and white Converse sneakers. Kim Jackson was there, as were his parents and Mike Cannon-Brookes. He began by saying he did not like this kind of thing: "Over my career, I've always preferred doing rather than talking." He then read from the essay he wrote for Mrs Crawford in 1991. He mentioned that he had come to Canberra as a twelve-year-old on the Year 6 trip to Questacon and Parliament House, and that his predictions had said nothing about what government would do. Now he was in the capital asking the government to act.
His proposal was that Australia, with its land, its energy, its legal system, and its position on the edge of Southeast Asia, should host the region's AI data centres, including "digital embassies" where friendly governments could keep their data on Australian soil under their own jurisdiction. "We should be exporting megawatts as megabytes," he said, "and, in return, getting paid megabucks." The pecking order of the AI era had not been settled, he said: "The scoreboard for the AI era is still blank." In 2014 he had argued it was too late to win software. He thought AI was still open.
His biggest request was about copyright. Australia lacked the fair-use and text-and-data-mining exceptions that he said the United States and Europe had, and he asked the Attorney-General to "urgently amend" the law. On ABC's 7.30 two weeks later he went further: "All AI usage of mining or searching or going across data is probably illegal under Australian law." His test for fair use was whether the result was transformative. "If someone had used my intellectual property to compete with me, then I think that is an issue," he said. "If they'd used all the intellectual property of all the software on the world to help people write software better in the future, I think that is a fair use." Guardian Australia noted that US law on the question was not settled. Capital Brief said experts disputed his reading. At the Financial Review's AI Summit in June 2026, he said that training a model in Australia under current law would require tech companies to "cut a deal with every single recording artist in the entire world." The music industry answered within days that about 80 percent of recorded music could be licensed through four deals, and the underlying publishing rights through one.
There is an irony here that his critics did not point out. Atlassian was built on removing negotiation from a purchase: a buyer finds the product, sees the price, and pays without anyone involved. Now Farquhar was arguing that AI training could not be done in Australia because it would require too many negotiations. The music industry replied that it already had something close to a standard price list. The man who had sent a $1,000 price to American Airlines was being told the licensing was already reasonable.
At the same summit he warned that the Albanese government's capital gains tax changes had doubled the effective tax rate on startups, and said he had raised it with Treasurer Jim Chalmers before the budget. In March 2026 Atlassian, now run by Cannon-Brookes alone, announced it would cut about 10 percent of its staff to concentrate on artificial intelligence and enterprise sales. The company that had famously had no salespeople was now building an enterprise sales operation, without Farquhar.
A smaller item turned up in the February 2026 political donation disclosures. Farquhar the billionaire had given $1.5 million to Climate 200. The records also seemed to show that he had given $22,250 to the Queensland Greens. That donation was from a different man, a Brisbane resident also named Scott Farquhar. At least one news site attributed it to the tech founder and then deleted the claim, and Guardian Australia admitted its own data analysis had made the same mistake. The confusion happened because the government had stopped publishing donors' addresses, so the records listed only a name. A man who had spent twenty years as the less visible half of "Mike and Scott" was now hard to tell apart from someone else with his name.
He has, predictably, become the subject of print-on-demand biographies, including Code to Capital — From Startup Underdog to Tech Billionaire and The Quiet Software Executive. They are unauthorized, and their titles describe him with confidence. He did not cooperate with them, or with Tech Bros. He has given interviews, worked his way down Bessemer's lists of lessons, and told the Amsterdam story to new graduates, and he has never written a memoir.
Back at the lectern in Canberra in July 2025, he read one more line from the essay he wrote in 1991, about buying groceries in a future he had not yet seen. "In the future there will be no supermarkets. Instead, people will be able to write out a shopping list and order over a machine. You will even be able to find out how much the shopping bill will cost before you order."
He was wearing black and white Converse.
Part IIThe Playbook
What follows is drawn from Farquhar's decisions rather than his speeches. He has given many speeches, but he described himself accurately when he said he preferred doing to talking. The principles below come from the record: the pricing, the closed offices and products, the pledge made before there was anything to pledge, and the partnership that lasted twenty-two years before a fence dispute and a book complicated it. Some of these are uncomfortable. Most run against standard advice for founders.
The principles
- 1.Let your constraints write the business model.
- 2.Price below the approval threshold.
- 3.Hire engineers where others hire closers.
- 4.Close the wrong office before it becomes a habit.
- 5.Pledge a percentage of nothing.
- 6.Tell the truth before you take the money.
- 7.Divide the crown, not the company.
- 8.Fear omission more than commission.
- 9.Kill what you can't defend to the people building it.
- 10.Audit your calendar against the four jobs.
- 11.Stop optimising for efficiency in a world you can't predict.
- 12.Build the forest, not just the tree.
Principle 1
Let your constraints write the business model
Farquhar's explanation of Atlassian's model to Bessemer is a chain of consequences. Global sales required self-service. Self-service required easy installation. Easy installation required a price nobody needed to negotiate. None of these was chosen for its own sake. Each was forced by the one before, starting from one belief, that "the world was going to be flat," and one shortage: "we didn't have any money to hire expensive sales people."
The point is not that limits produce creativity. The point is that a founder who takes a constraint seriously can follow it to a model competitors cannot copy without giving up advantages they already have. An enterprise software company with a sales force could not drop its prices to $1,000 without wrecking its own revenue. Atlassian had no such business to protect, so the limit protected it.
Tactic: Write down your two hardest constraints, follow each to its logical end, and test whether the model that results is one your best-funded competitor couldn't adopt without hurting itself.
Principle 2
Price below the approval threshold
"Instead of selling our software for $100,000, we sold it for $1,000." The cut in price was large, but the more important effect was on who made the buying decision. Below the amount an engineer could put on a credit card without asking a manager, the buyer was the person using the product, not a procurement committee. Farquhar told Bessemer he priced Atlassian so that most buyers could pay without going to their boss.
That was why the American Airlines order arrived unprompted. Nobody at the airline had to be persuaded, because the person who bought Jira was the person who needed it. Low prices also built loyalty among users who would not switch, and the upgrades came later. It was a bet on volume, and it only worked combined with Principle 1. Of course, he said, "you need to sell a lot of thousand dollar software to make a real business."
Tactic: Find the spending limit your target user can approve alone, and design your entry price and free tier to sit under it.
A
Conventional Wisdom vs. the Farquhar Model
How Atlassian inverted the enterprise software orthodoxy of the early 2000s.
| Conventional wisdom | Atlassian under Farquhar |
|---|---|
| "People won't buy without talking to a sales person." | American Airlines bought Jira off the website by fax, 2003. |
| Enterprise software costs six figures. | $1,000, below the credit-card approval line. |
| Spend heavily on sales and marketing headcount. | $102M revenue in 2011 with no salespeople; reinvest in R&D. |
| Raise venture capital early. | Eight years bootstrapped; first outside round in 2010 was secondary. |
| Philanthropy comes after success. | Pledge 1% of equity when equity was worth "virtually nothing." |
| Bring people back to the office. | Team Anywhere, 2020, plus scheduled "intentional togetherness." |
Principle 3
Hire engineers where others hire closers
Bessemer's summary of Farquhar's view is that engineering headcount is a better long-term investment than sales and marketing headcount. In practice, every dollar that a competitor spent on a salesperson, Atlassian spent on someone who made the product easier to evaluate, install, and keep using. In 2016 Fortune described the split: Farquhar ran engineering and research. As late as that year the company still had no traditional sales force.
The reasoning is that a salesperson's work stops when they stop, while a product improvement keeps producing returns. A well-designed onboarding flow makes the case to every visitor in every time zone, indefinitely. That is not an argument that sales is worthless. Atlassian is building an enterprise sales function now, at a scale where the largest customers expect one. It is an argument that sales was not needed until the product had done everything it could.
Tactic: For every proposed sales or marketing hire, ask what product improvement would remove the need for that person's job, and fund it first if you can.
Principle 4
Close the wrong office before it becomes a habit
The New York office was a reasonable decision because the customers were in America, and it was wrong almost immediately. Atlassian closed it quickly and later moved its American base to San Francisco, where the engineers were. "Closing that office so soon was hard," the company wrote, "but it was good practice in making tough calls — of which there would be many to follow."
Closing it early mattered more than closing it at all. A mistake that is a few months old has few people relying on it. One that is two years old has a lease, a local leader, a team, and a set of internal arguments for keeping it. Farquhar's willingness to reverse an obvious and well-meant decision before it took root was the same instinct, at small scale, that later shut down Stride.
Tactic: Set a review date for every new expansion at the moment you announce it, and treat reversing it as the default unless the evidence says otherwise.
Principle 5
Pledge a percentage of nothing
"Donating one per cent of equity was easy to do back when we started since it was one per cent of virtually nothing." This is the most practical point in Farquhar's philanthropy. The commitment cost nothing when it was made and grew automatically with the company's success, so it never had to be re-argued when the stakes were high. By the time the pledge was worth billions, it was already part of the company's identity and not up for debate.
He then applied his go-to-market methods to the idea itself. When he saw that almost no one had copied Salesforce, he decided the problem was distribution and set up Pledge 1% "on an oily rag" with a goal of 500 companies. It reached 19,000. He did not argue for generosity. He made it cheap to start, easy to adopt, and likely to spread.
Tactic: Write your giving commitment into your founding documents as a percentage, while the percentage is still worth almost nothing.
Principle 6
Tell the truth before you take the money
In 2010, with Atlassian about to meet venture investors for the first time, a hacker breached its systems. Farquhar cut his honeymoon short, ran the investigation from the office, and told investors and customers about it before the meetings, not after. The round closed. Accel put in $60 million, and a 20VC episode title says the investment eventually earned the firm about $6 billion.
The Amsterdam story is the other side of this. At twenty-two, asked his age, he admitted he had been working out "how much was I going to lie." The founder of 2010 had clearly settled that question. Disclosing a problem before a deal closes looks like it reduces your leverage. In fact it shows the counterparty how you will behave in every future crisis, which is the main thing an investor is trying to find out.
Tactic: If something material goes wrong before a financing or partnership closes, disclose it before the counterparty would have discovered it, and present it as evidence of how you handle trouble.
Principle 7
Divide the crown, not the company
The co-CEO arrangement lasted about twenty-two years because responsibilities were split clearly, not because the two men always agreed. Farquhar ran engineering and research. Cannon-Brookes ran product design, sales, and marketing. When they deadlocked, according to one account, they settled it with rock-paper-scissors, a random rule both accepted that kept either from overruling the other.
The arrangement also had limits, and the record shows them. It held through the IPO, the Stride shutdown, and the pandemic. Whatever went wrong between them, if anything did, seems to have happened in their private lives, over a fence at Point Piper rather than in the boardroom. A partnership can be designed to share power at work. It cannot be designed to cover everything else.
Tactic: If you share leadership, write down who decides what, agree in advance on a tiebreak mechanism neither of you controls, and keep the partnership's private affairs as clearly separated as its business ones.
Principle 8
Fear omission more than commission
Farquhar's description of how his mistakes changed is one of his most useful observations. Small companies fail by doing the wrong thing. Large companies fail by not doing the right thing. "My biggest mistakes," he told Harry Stebbings, "are being in the right place at the right time but not going hard enough behind the opportunities that we've had in front of us."
This is a problem of attention. Mistakes of commission are visible. They appear in the numbers, someone owns them, and they get a post-mortem. Mistakes of omission do not show up anywhere. Nobody writes a post-mortem on a market they never entered. A CEO at scale has to go looking for these on purpose, because the organisation's reporting will never surface them.
Tactic: Once a quarter, list the three opportunities you are best placed to pursue and are not pursuing hard enough, and require a written reason for each.
T
The Decision Ledger
Commissions, omissions, and reversals across Farquhar's tenure.
2002
Jira 1.0 ships; the founders abandon the support business for products.
2003
American Airlines buys Jira from the website, unprompted, by fax.
2004
Confluence 1.0 launches; the New York office opens and is quickly closed.
2005
ShipIt hackathons begin; Jira Service Management later emerges from one.
2010
Hack disclosed before fundraising; Accel invests $60 million in a secondary round.
2014
Pledge 1% co-founded; Bradfield Lecture on software as the new electricity.
2015
Nasdaq IPO under the ticker TEAM at a $4.37 billion market cap.
2018
Stride and HipChat sold to Slack and shut down, less than a year after Stride's launch.
2020
Team Anywhere announced; distributed work becomes company policy.
2024
Farquhar steps down as co-CEO on August 31.
2025
National Press Club address: "megawatts as megabytes."
Principle 9
Kill what you can't defend to the people building it
Atlassian shut Stride less than a year after launch. Farquhar's test was neither market share nor revenue. It was whether he could look the team in the eye and tell them the next five years of their working lives would be well spent on it. He could not, so the product went, its IP was sold to the rival it had been built to beat, and Atlassian took a small stake in that rival.
The test is useful because it is hard to fudge. Dashboards can be argued over and projections can be adjusted. It is much harder to stand in front of a team and make a promise you do not believe. The test is about the people, not the product, and talented people's time is the resource a company can least afford to waste on something that is failing.
Tactic: Before each annual planning cycle, imagine telling each product team that their next five years are best spent on what they are building; wherever you would hesitate, open a shutdown or sale discussion.
Principle 10
Audit your calendar against the four jobs
Farquhar returns every year to the same article describing the four jobs of a CEO: hire and fire the management team, set the vision, set the culture, and allocate resources. "I go back to that article every single year," he told 20VC, "to ask how much of my time is on those areas." He checks his actual diary against that list.
It is not a complicated idea, and that is why it works. A CEO's calendar fills up with whatever is urgent, and nobody else has the standing to push back. The four jobs give him a fixed standard to measure against. Anything that does not serve one of the four is probably being done by the wrong person.
C
The Four Jobs of a CEO
Farquhar's annual calendar audit, with the evidence from his tenure.
| Job | Farquhar's evidence | Type |
|---|---|---|
| Hire and fire the management team | "Enthusiastically rehire" test for every role over the next 12 months | People |
| Set the vision | "Unleash the potential of every team"; software as the new electricity | Direction |
| Set the culture | Five values, including "Open company, no bullshit" | Culture |
| Allocate resources | Engineering over sales; Stride shut down; Loom acquired for $975M | Capital |
His hiring version of the test, posted after Fortune Brainstorm Tech in 2023, is just as blunt: "If you can look at someone and know you'd enthusiastically rehire that person for the job they have to do over the next 12 months, great. If not, you need to have the tough conversation."
Tactic: Once a year, tag every recurring meeting in your calendar with one of the four jobs, and delegate or delete anything you can't tag.
Principle 11
Stop optimising for efficiency in a world you can't predict
Farquhar's story of the Pennsylvania steelworkers, whose optimised shovel loads quadrupled their output, is a deliberate warning. The method worked because "their world moved slowly." For a modern team, he argued in 2017 and again in 2018, optimising for efficiency means optimising for a world that no longer exists. "We're indoctrinated to think variability is a bad thing, and it's not — we need to learn from failure."
His alternative was to spread decision-making across the organisation and make the flow of information as fast as possible. Openness, "open company, no bullshit," was the means. He argued that a company's speed depends on how quickly information moves from one part of it to another. Team Anywhere applied the same reasoning to location. Instead of fixing where people work, it fixed when they gather, and measured the roughly 30 percent rise in connection that each gathering produced before it faded.
Tactic: Find one process your team has fully optimised and ask what would happen if the conditions it was designed for changed next quarter; build in slack and faster information flow where the answer worries you.
Principle 12
Build the forest, not just the tree
"It's easy for a seed to germinate almost anywhere," Farquhar said in 2014. "There are fewer conditions for that to turn into a huge oak tree." He has spent a decade trying to create those conditions in Australia: a curriculum that teaches computing, skilled migration, a technology precinct at Central, venture capital that can fund companies past their early rounds, and now, through the Tech Council, data centres and copyright reform. The Bradfield Lecture warned that the same distance that let "Mike and Scott" sell to the world from Sydney would let competitors anywhere do the same.
He measures his success partly by who leaves. Dovetail, Sajari, and Pyn were all started by former Atlassian employees, and Farquhar has talked openly about an "Atlassian mafia." A founder who sees departures as loss will hold onto talent. One who sees them as the ecosystem growing will train people well and send them out. Not every argument he makes on this front is sound. The music industry's rebuttal on copyright licensing was pointed. But the general method holds: a company does better if the industry around it is strong.
Tactic: Measure part of your company's impact by the founders it produces, and commit at least one senior leader's time each year to the policy, education, or capital gaps that limit your industry's growth.
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
— Scott Farquhar, UNSW graduation address, March 17, 2011So instead of selling our software for $100,000, we sold it for $1,000. Of course, you need to sell a lot of thousand dollar software to make a real business.
— Scott Farquhar, JJC Bradfield Lecture Q&A, October 2014So, I think it's easy for a seed to germinate almost anywhere; there are plenty of conditions for that to happen. There are fewer conditions for that to turn into a huge oak tree, you'd need some pretty special conditions and you do need that concentration of talent.
— Scott Farquhar, StartCon, December 2017The days where the boss is the smartest person in the room are gone. My team reminds me of that all the time.
— Scott Farquhar, to Forbes Australia, on Pledge 1%1% of product isn't worth anything when you don't have any product to give away, and we didn't have profit for a long time.
— Scott Farquhar, 20VCMy biggest mistakes are being in the right place at the right time but not going hard enough behind the opportunities that we've had in front of us.
Maxims
- Software should be bought, not sold. Design the product so the customer can make the purchase without you.
- Price under the boss. A price the user can approve alone removes the approval chain that slows every enterprise deal.
- Constraints are strategy. Follow your hardest limitation to its end and you may find a model your competitors can't copy.
- Give from zero. Pledge a percentage while it is worth nothing, so it is already decided when it is worth a great deal.
- Disclose before the deal. Telling the truth before you take the money shows investors how you will act in every later crisis.
- Close it while it's small. A mistake reversed in its first months costs a lease; one reversed after two years costs a culture.
- Watch for omission. At scale, the most expensive mistakes are the opportunities nobody acted on.
- Pass the eye-contact test. If you can't tell a team their next five years are well spent, the product should not continue.
- Efficiency suits a static world. In a changing one, value fast information flow and spread-out judgement over optimised routine.
- Grow the forest. The best measure of a company's success may be the companies its alumni go on to build.
