Atlassian: how two Sydney graduates built a global software empire without a sales force

Somewhere around 2015, I started noticing Atlassian's name on the internal tooling of almost every tech company I was reporting on — not because anyone had sold it to them, but because developers had simply started using it. That detail stuck with me. In an industry where enterprise software is typically shifted by armies of account managers with generous expense accounts, here was an Australian company growing at scale with almost none of that. It was worth understanding properly.

The short version is this: Mike Cannon-Brookes and Scott Farquhar met at the University of New South Wales in the early 2000s, decided they'd rather build something than take graduate jobs, and borrowed roughly $10,000 on a credit card to start Atlassian. Two decades later the company was valued at well over $50 billion on the Nasdaq. But the short version doesn't explain the how, and the how is the interesting part.

The credit card origin — and what it actually meant

The $10,000 credit card story gets told a lot, sometimes to the point of becoming mythology. What it actually signals is more practical than romantic. Cannon-Brookes and Farquhar were not venture-backed from the start. They were not building for an exit on someone else's timeline. They needed the product to generate revenue, quickly, or they'd go under. That constraint shaped everything that followed.

Their first product was Jira, a bug and issue tracker aimed at software development teams. It was not a glamorous product category. But it was a genuinely useful one, and they priced it in a way that put it within reach of small teams — not just enterprise procurement budgets. The pricing model, combined with a product that developers actually wanted to use, meant that Jira spread through organisations from the bottom up. A team of five would adopt it; six months later the whole company was on it. No sales call required.

Building without a sales force

Atlassian's decision to grow without a traditional enterprise sales operation is the thing that most distinguishes its early history from comparable software companies. For a long time, industry observers treated it as a quirk, or a limitation that would eventually have to be fixed. It was neither.

The logic was straightforward, even if the execution was not. Sales teams are expensive. A direct sales motion requires hiring, training, managing, and paying commission to people who then convince customers to buy something they may or may not need. Atlassian's alternative was to make products that developers discovered themselves, adopted, and then advocated for internally. The company invested in the product and in self-serve purchasing infrastructure instead. Customers could download, trial, and buy without ever speaking to a human being at Atlassian.

This is sometimes described as a product-led growth model, a term that has become fashionable enough to be almost meaningless. What it meant in practice was that Atlassian's cost of acquiring a customer was structurally lower than most of its competitors. That difference showed up in the margins, which in turn funded more product development. It was a flywheel, and once it was turning, it was hard to stop.

Confluence, the team wiki product, followed Jira and was similarly distributed. Jira Service Management came later. Each new product could be sold into the existing customer base without a sales organisation, because those customers were already inside the Atlassian ecosystem. The company was also early to build a partner and reseller channel, which provided market coverage without the fixed cost of a direct sales headcount. Honestly, I think the no-sales-force story sometimes overshadows how much Atlassian leaned on that channel to reach enterprise accounts — it was never quite as pure as the narrative suggests.

Sydney headquarters, global ambitions

The company was headquartered in Sydney for most of its first decade, operating out of offices that over time occupied space in the inner city. It expanded internationally early — a San Francisco office opened well before the Nasdaq listing — but the founders stayed Australian and the company retained a distinctly Australian identity in its culture and communications. Cannon-Brookes in particular became one of the more outspoken technology figures in the country, eventually becoming as well known for his views on energy policy as for his software company.

That global expansion was driven partly by necessity. The Australian domestic market for enterprise software is simply not large enough to support a company at Atlassian's ambitions. To build a genuinely global software business, you needed American customers, European customers, and eventually customers across Asia. The product-led distribution model helped here too: software that spread by developer word-of-mouth crossed borders without needing a local sales office in every market first.

By the time of its Nasdaq listing in December 2015, Atlassian had tens of thousands of customers across well over a hundred countries. The IPO priced shares at $21, and the company was valued at around $5.8 billion at listing — a substantial figure for an Australian technology company at the time. It was the largest US tech IPO of 2015 by some measures, and it was conducted without the company having raised traditional venture capital in the way most comparable businesses had. The founders retained significant ownership as a result.

The Nasdaq listing and what came after

Listing on the Nasdaq rather than the ASX was a deliberate choice, and one that drew some commentary at the time. The logic was access to a deeper pool of technology-focused institutional investors and better comparables for valuation purposes. Australian institutional investors have less familiarity with pure software multiples; American tech investors price these businesses differently. Whether that was the right call for Australia's capital markets ecosystem is a separate argument — one that has been running ever since and shows no sign of resolving — but for Atlassian's own trajectory it clearly worked.

Post-listing, Atlassian continued to grow both organically and through acquisition. The purchase of Trello in early 2017 for a reported $425 million was one of the more significant moves, adding a lighter-weight project management tool that reached a different segment of the market. The company also acquired a number of smaller businesses in collaboration software, IT service management, and communications.

The shift to cloud was the defining strategic move of the late 2010s and early 2020s. Atlassian had historically sold a significant proportion of its products as on-premises, self-hosted software — customers ran the servers themselves. The migration of that customer base to cloud-hosted subscriptions was a substantial undertaking, and one that involved some friction with long-standing customers who preferred the control and cost certainty of the old model. The company eventually announced it would end sales of server products, forcing customers to either move to the cloud offering or to Data Center, the self-managed alternative for larger enterprises. That transition was bumpy in places, but the recurring revenue model that came out the other side was financially transformative.

Scale and structure today

Atlassian today employs tens of thousands of people globally and generates annual revenue well north of $4 billion, according to its published financial results. The product portfolio spans issue tracking, project management, team wikis, IT service management, and communications tools. It competes across those categories with Microsoft, ServiceNow, Monday.com, and others — a list that would have seemed implausible when two UNSW graduates were charging their startup to a credit card.

The company also made a widely-reported structural decision to go fully distributed during and after the COVID-19 period, allowing employees to work from anywhere. That move was backed by genuine conviction from the founders, not just pandemic necessity, and it reshaped Atlassian's real estate footprint and hiring strategy. It also became a data point in the broader argument about remote work productivity, one that Cannon-Brookes and Farquhar were publicly willing to defend.

The Nasdaq listing gave the market a live valuation on the company's fortunes. Through the technology bull run of 2020 and 2021, Atlassian's market capitalisation reached extraordinary figures; through the subsequent correction of 2022 it fell sharply, as did most high-multiple software businesses. That volatility is the cost of public markets, and it doesn't change the underlying story of what the company built. A lot of software companies with venture-funded sales armies and decade-long head starts never got anywhere near the scale Atlassian reached.

What the model proves — and doesn't

The temptation with Atlassian is to conclude that any sufficiently good software product can be distributed this way. I'd push back on that. The no-sales-force model works when the buyer and the user are the same person, or close to it — when the developer or team lead who wants the product can also acquire it. In enterprise software categories where purchasing decisions sit with a CIO or a procurement committee who will never use the product themselves, this model is much harder to execute. Atlassian chose its initial product categories well, perhaps deliberately, perhaps partly by luck.

What it does prove is that Australian companies can build globally competitive software businesses without being American, without relocating to Silicon Valley, and without following the standard venture-backed playbook. That's worth stating plainly, because for a long time it wasn't obvious it was possible. Atlassian and, more recently, Canva have changed that picture substantially — the latter pulling off something comparable from Perth, of all places.

For the broader story of how Australian companies build durable competitive positions internationally, it's worth reading alongside the Outliers & Global Empires archive. The patterns that show up there — patient capital, distribution advantage, genuine product differentiation — all appear in Atlassian's history too, even if the specific form they take is different from, say, Cochlear's dominance in hearing implants.

The $10,000 credit card is a good story. The more interesting story is the twenty years of decisions that followed it.

— Priya Naidu, Outliers & Global Empires