Atlassian: The Sydney Software Company That Sold Without Salespeople

Atlassian is the most commercially significant company Australia has produced in the software era, and its central innovation was a distribution model rather than a product. Selling enterprise software normally requires expensive field sales teams, long procurement cycles and executive relationships. Atlassian built products that individual teams adopted directly and let the organisation come to it. Understanding why that worked, and why it eventually stopped being sufficient, is the useful part of the story.

How did two graduates build a global software company from Sydney?

By solving their own problem and pricing it for people like them. Mike Cannon-Brookes and Scott Farquhar founded Atlassian in 2002 after university, reportedly funding the early business on credit cards, and built issue-tracking software that development teams needed and could buy directly.

The pricing was the strategy. Jira was cheap enough that a team lead could approve it on a company card without a procurement process, a business case or a sales meeting. That removed the single largest barrier to enterprise software adoption — the buying process — and let the product spread on merit rather than on relationships.

Geography reinforced it. Selling enterprise software from Sydney to customers in North America and Europe by traditional means would have required expensive offices and field teams in every market. A self-serve model made location irrelevant, which turned Australia’s distance from major markets from a handicap into something close to a non-issue.

Why did the no-sales-force model work?

Because the buyer and the user were the same person. Developers choose their own tools, evaluate them by using them, and recommend them to colleagues. A salesperson adds nothing to that process and adds substantial cost, so removing the salesperson lowered the price and improved the experience simultaneously.

The economics compound. Money that competitors spent on sales commissions and field teams, Atlassian spent on product development, which made the product better, which drove more organic adoption. Software companies typically spend 40% or more of revenue on sales and marketing; a company spending materially less on distribution can outspend everyone on engineering at the same revenue.

Land and expand did the rest. One team adopting Jira leads neighbouring teams to adopt it for compatibility, and eventually the organisation standardises on it. By the time a procurement department becomes involved, the software is already embedded and the conversation is about enterprise licensing rather than whether to buy at all.

Why did Atlassian eventually build a sales organisation?

Because the largest customers required it. Once a company has thousands of Atlassian users across multiple products, the conversation involves security review, compliance certification, data residency, integration with identity systems and a negotiated enterprise agreement. None of that happens self-serve.

The migration from server-based products to cloud forced the issue. Moving a large enterprise’s data and workflows from on-premise servers into a vendor’s cloud is a substantial project requiring planning, support and executive sponsorship, and Atlassian ended server support to complete the transition. That migration was the largest commercial risk the company has taken and it succeeded.

The numbers show the outcome. Atlassian ended Q1 FY2026 with 53,017 customers generating more than US$10,000 in cloud annualised recurring revenue, up 13% year on year, and cloud revenue reached US$998 million in that quarter. The self-serve engine still acquires customers; a sales organisation now expands them.

What is the AI story?

Rovo, Atlassian’s AI capability layered across Jira, Confluence and Jira Service Management, surpassed five million monthly active users after the company delivered its first US$1 billion cloud revenue quarter in Q2 FY2026. Monthly active users of AI capabilities had grown 50% quarter-over-quarter in the preceding quarter.

The strategic advantage is data. Atlassian holds the record of how work actually happens inside an organisation — which tickets were raised, how they were resolved, what documentation exists, who did what. An AI capability grounded in that record can answer questions no general-purpose model can, because the information exists only in the customer’s Atlassian instance.

Commercially it supports pricing. AI features are being sold as premium tiers, which raises average revenue per user without acquiring a single new customer, and Cannon-Brookes has described AI as the best thing to happen to Atlassian. For a company whose growth was moderating from 20%-plus rates, a credible upsell lever arriving at that moment is a substantial change to the trajectory.

What did Atlassian do for the Australian ecosystem?

It proved the market was reachable, and then it recycled the people and the money. Before Atlassian, the assumption in Australian technology was that a globally significant software company had to relocate to the United States. Atlassian listed on Nasdaq but kept substantial operations in Sydney and demonstrated that world-class software could be built from Australia.

The talent effect has been larger than the capital effect. A generation of Australian software engineers, product managers and operators learned how a hyper-growth software company works, and many left to found or join others. That transfer of operating knowledge is what distinguishes a functioning startup ecosystem from a collection of individual companies.

The founders have also shaped the country beyond technology. Cannon-Brookes deployed personal capital into climate and energy, most visibly through the campaign that defeated AGL’s proposed demerger and rewrote its coal closure timetable. Australian technology wealth has become a factor in Australian corporate governance, which is genuinely new.

What did the cloud migration actually involve?

Moving an entire customer base from software they ran on their own servers to software Atlassian runs on their behalf, and then ending support for the server products entirely. That is one of the riskiest transitions a software company can attempt, because every customer must be persuaded to change something that currently works.

The commercial rationale was overwhelming. Cloud subscriptions produce predictable recurring revenue, allow continuous product improvement without customers managing upgrades, and enable the data and AI capabilities that a fragmented estate of self-hosted installations cannot support. Server licences produced one-off revenue and enormous support complexity.

The risk was that large enterprises with security, compliance or data residency requirements would refuse to move and switch vendors instead. Atlassian mitigated it with enterprise cloud offerings, data residency options and migration support, and the result — cloud revenue passing US$1 billion in a quarter and growing 26% — indicates the transition worked. Companies contemplating a similar shift should note it took years and required ending the legacy product to complete.

What does the marketplace add?

A third-party ecosystem that makes the core products harder to leave. Independent developers build apps extending Jira and Confluence, customers install them to fit their own workflows, and Atlassian takes a share of the revenue while adding capability it never had to build.

The strategic value is switching cost. A team using base Jira could migrate to a competitor with effort. A team using Jira plus a dozen marketplace apps configured to its specific processes over several years is effectively locked in, because replicating that configuration elsewhere is a project nobody wants to sponsor.

It also solves a product problem elegantly. Every enterprise customer wants a feature that is essential to them and irrelevant to everyone else, and building all of them produces bloated software. A marketplace lets those requirements be met by someone else, at the customer’s expense, while the core product stays focused.

What are the risks to Atlassian now?

Competition from platform owners is the largest. Microsoft bundles collaboration, project tracking and documentation into subscriptions many customers already hold, and a bundled product that is merely adequate frequently defeats a superior standalone one on procurement economics. Atlassian’s defence is depth in software development workflows specifically.

AI cuts both ways. It supports premium pricing and deeper product value, and it also lowers the cost for a competitor to build comparable functionality. If generating a project tracking tool becomes substantially easier, the barrier that protected an established product for two decades erodes at the edges.

Growth normalisation is the third. Revenue growth has moderated from the 20%-plus rates of earlier years toward the mid-teens on some forecasts, and a company valued on growth must either sustain it or be revalued. The AI upsell is the mechanism management is relying on to keep growth ahead of that trend, and whether it does is the central question for the stock.

One organisational note worth recording. Atlassian operates a distributed work model it calls Team Anywhere, under which employees work from anywhere in a country where the company can legally employ them, with periodic in-person gatherings rather than mandated office attendance. For a company founded in Sydney and selling to the world, distributed work is not an experiment adopted after the pandemic — it is a continuation of how the business always operated across time zones. Companies debating return-to-office policies rarely consider that the answer may depend on whether the business was ever geographically concentrated in the first place.

Frequently Asked Questions