Scott Farquhar on Building Atlassian, CEO Decisions and Distributed Work

Atlassian co-founder and co-CEO Scott Farquhar tells Harry Stebbings that he and Mike Cannon-Brookes started the company after university because neither wanted a conventional job. They tried several software ideas before building Jira. He says the company had grown to roughly 250,000 customers and 11,000 staff by the 2023 interview.

Farquhar believes people need to enjoy their work, but says competence can also deepen that enjoyment. He loved creating software, although he would not necessarily have chosen bug tracking as a lifelong mission. The broader challenge of building a company and helping teams work better sustained his interest.

Atlassian started after the dot-com crash, when venture capital was hard to raise. Farquhar says the company bootstrapped for many years before taking venture investment. He believes financing choices depend on competition and market conditions: a small niche can still support bootstrapping, while a heavily funded market may be different.

He challenges the idea that startups are glamorous, pointing to survivorship bias and the frequency of failure. Early on, he recalls, telling people he worked on a startup was often interpreted as saying he was unemployed.

Farquhar recounts a 2010 security incident that interrupted his honeymoon in Botswana. A message from Cannon-Brookes reached him through a long chain of contacts, and he spent a night unsure what had happened. On calling back, he learned that an attacker had accessed one of the company’s systems.

The team had worked around the clock to investigate. Farquhar returned to Australia so he and Cannon-Brookes could take turns managing the response. They informed customers and prospective investors before the planned fundraising meetings, because they considered transparency important.

He recalls the business then operating at roughly a $60 million annual revenue run rate and an estimated 20–30% profit margin; he presents these as recollections, not audited figures. Atlassian invited several investors it already knew and requested sealed bids for its first venture financing.

Farquhar says the process was designed to elicit each investor’s best offer rather than negotiate through successive indicative prices. The founders had spent time with the prospective partners beforehand and were comfortable working with any of them. Accel made the winning offer and invested $60 million, according to his account.

He says Atlassian also offered Benchmark an opportunity to participate if it matched Accel’s price, but it did not do so. His retrospective estimates of what investors might have earned are hypothetical and depend on the actual investment and distribution history. He says a similar bidding process was later used before the IPO.

As a small company, Farquhar says, mistakes often involved doing the wrong thing; at greater scale, his biggest regrets are opportunities not pursued aggressively enough. He cites Bitbucket and the company’s workplace messaging products, arguing that Atlassian had promising products but did not invest enough behind them.

Investment has to come from somewhere, so funding a new idea means reducing resources elsewhere. Atlassian’s Point A program incubates products such as Atlas and Compass. Farquhar says the aim is to support promising opportunities enough that the company does not later regret leaving them underdeveloped.

He returns to four CEO responsibilities: hire the management team, set the vision, establish culture and allocate resources. He periodically reviews his calendar against those duties. He says resource allocation need not be a daily exercise; Atlassian moved from an annual to a quarterly cadence.

Farquhar identifies systems thinking as a strength: understanding how changing one part of a business affects the whole. He connects it to curiosity across fields and to understanding customers, people and markets. His weakness, he says, is prioritization, because so many topics can capture his attention.

On AI, he compares the moment with earlier platform shifts such as personal computers, the internet, mobile and cloud. He predicts that AI will change the competitive landscape, while acknowledging that the eventual leaders are uncertain. As Atlassian’s co-CEO, he has a commercial interest in its AI strategy.

He expects simple text summarization to become common and sees greater potential in combining data and workflows. Examples include finding an answer across support material or linking an error to the code and people involved. He says the right interfaces and actual implementation were still unclear.

Farquhar says hundreds of Atlassian engineers were working on AI at the time. Customer data belongs to customers, he says, so the company’s task is to help them use it rather than build a large model across companies’ data. He favors learning through experiments, including discovering what does not work.

Small companies can move quickly, but Farquhar says a larger company has advantages too: existing customers and their long histories can help it test ideas with them. This is his assessment of Atlassian’s position, not evidence that size always produces better AI products.

Atlassian chose early in the pandemic not to require employees to return to an office, allowing it to hire in locations without offices. Farquhar says the policy improved access to talent and retention, and cites 92% of employees saying it helped them do their best work. These are company-reported results.

He says remote work still requires intentional in-person gatherings. According to Atlassian’s internal surveys, team and company connectedness rose roughly 30% after gatherings and returned toward baseline over four to five months. He treats these as company findings and says the team is still testing which activities build the strongest bonds.

The remote policy became a major commitment because many employees no longer lived near offices; reversing it could damage trust. Another consequential choice was shutting down Stride, Atlassian’s workplace messaging competitor, after years of investment and a rewrite.

Farquhar identifies three lessons from the messaging business. Its paid entry price was low, but Slack’s free tier made it easier to try; he wishes Atlassian had changed its pricing. Doubling a small team did not double productivity because coordination costs grew, and he believes the opportunity needed a larger investment.

He says the product was growing several times year over year, but in hindsight the market opportunity demanded even faster growth. These are his retrospective assessments rather than a controlled comparison. Although employees had invested years of effort, he concluded they could have greater impact on other products.

Farquhar monitors competitors but says winning often comes from listening to customers and developing a different approach rather than copying rivals. He cites Jira Product Discovery as an example of entering an existing market with a distinct product.

Stebbings raises criticism that Jira can be complex for new users while simpler competitors appeal to small teams. Farquhar acknowledges complexity as a trade-off of a product designed to scale. He says some complexity is inherent and some comes from how customers configure it; Atlassian continues to simplify the product.

He argues that Jira’s ability to integrate with code, build and security tools is valuable as teams grow, while acknowledging that a product built for five people may be simpler. This is the company leader’s account of its competitive strengths.

Asked about AI and per-seat pricing, he agrees AI could put pressure on the model. If a support task can be solved without a human agent, the number of paid seats may no longer reflect the value delivered. He expects pricing to evolve but says no one yet knows the best replacement.

The effect may depend on price and whether AI augments or replaces a particular role, he says. A company might need fewer seats but pay more for added value, while other use cases could require a different model. These are scenarios, not predictions of specific pricing or job outcomes.

On money, Farquhar describes a changing trade-off between time and expense. As a student he had more time than money; later he had to learn when paying for convenience was worthwhile. He values buying back time to spend with his children.

He says raising children amid wealth while teaching humility and the value of money is difficult, and he does not claim to have solved it. Even when parents try to shield children, other people may discuss the family’s wealth. He expects it will take years to know how well he has done.

Farquhar credits his marriage to admiration for his wife Kim, shared experiences and learning to apologize quickly when wrong. He presents this as their experience and mentions research discussed in a book, without establishing a universal formula for relationships.

He describes an early Atlassian pricing experiment: for five days the company sold a five-user version for $5, donating the proceeds. He says it raised just over $100,000 in small transactions, revealing demand and beginning a relationship with the literacy nonprofit Room to Read.

That philanthropic experience helped inspire Pledge 1%, which encourages companies to commit portions of equity, product, profit or employee time to nonprofit causes. Farquhar cites about 18,000 companies having pledged at the time; this is his reported figure.

He says he has become more willing to accept that another person may be right and to disagree and commit, rather than insist on proving his own view. He also says markets had been more resilient than he expected. These comments reflect his assessment at the time of the interview.

Daily exercise and spending time in bed for eight hours help him remain calm, he says. Asked about fatherhood, he wants his children to make good choices and still want to spend time with him when they are adults.

Farquhar acknowledges a trade-off between work, health and family: the same 168 hours must accommodate all of them. He believes devoting all his time to either work or parenting might make him better at that one role, and tries to balance the responsibilities he has chosen.

He says one of the kindest things people can do is believe in someone’s potential. He credits Cannon-Brookes and an early chairperson with doing that for him. He describes his co-founder relationship as intense and complementary: Cannon-Brookes generates many ideas, while he is more methodical.

Looking ten years ahead, Farquhar hopes Atlassian can reduce administrative work and help teams collaborate more effectively, through AI, training and other tools. This is his vision for the company, not a demonstrated outcome. The conversation closes with both thanking each other.