Latest / Startup Stories with Fexingo: Conversations About Founders, Funding, and Building Companies from Zero / How Atlassian Grew to 200K Customers Without a Sales Team
Transcript
- Lucas: So here's a question that sounds almost like a riddle: how do you build a multi billion dollar software company selling to massive enterprises — banks, governments, aerospace companies — without ever hiring a single person whose job is to call a prospect and close a deal? Luna: That sounds like a fantasy, honestly. Every B2B founder I talk to says sales is the hardest thing to get right. Lucas: Right. And yet Atlassian — the company behind Jira, Confluence, Trello — they did exactly that. Co-founders Mike Cannon-Brookes and Scott Farquhar started in 2002, and their founding bet was: we will never have a sales team. No cold calls, no demo pitches, no enterprise account executives. The product sells itself — or it doesn't sell. Luna: And they got away with that? Even as they went after Fortune 500 clients? Lucas: That's exactly the tension. They held that line for over a decade — through IPO in 2015, through crossing 100,000 customers. Today they have more than 200,000 customers in 190 countries. And the core team that actually sells? Still essentially zero. They have customer success people, support, solutions engineers — but nobody with a quota. Luna: So how do they get a company like, say, NASA or Tesla to sign a six-figure contract without a sales rep ever picking up the phone? Lucas: It starts with a radical product decision. Atlassian's very first product — a bug-tracking tool called Jira — was sold online, download and pay, before the term SaaS even existed. They priced it by user tier, offered a full-featured trial, and let the product do the convincing. The idea was: if the tool actually solves a painful problem, engineering teams will adopt it organically — and then they'll champion it internally to get budget approved. Luna: So it's bottom-up adoption, not top-down. The engineers become the de facto salespeople. Lucas: Exactly. And that flips the entire customer acquisition cost equation. For most enterprise software companies, customer acquisition cost is the biggest line item — sometimes 40 to 60 percent of revenue goes to sales and marketing. Atlassian's sales and marketing spend? In their early years, it was consistently below 20 percent of revenue. And their customer acquisition cost was effectively zero for the first sale — because the customer found them, downloaded the trial, and converted. Luna: But that only works if the trial converts at a high rate. What was their conversion number? Lucas: They've never been super transparent about the exact percentage, but in their 2015 S-1 filing for the IPO, they disclosed that more than half of their revenue came from existing customers upgrading or buying additional products. And the initial conversion from free trial to paid? Industry analysts estimate it was well above industry average — maybe 15 to 20 percent, compared to the typical 2 to 5 percent for SaaS. Luna: That's extraordinary. What made their trial so sticky? Lucas: Two things. First, Jira was solving a genuinely painful problem — software developers were drowning in spreadsheets and email chains to track bugs. So the value was immediate. Second, Atlassian designed the trial to be generous in time but limited in users — so as your team grows, you hit the paywall naturally. It's a classic product-led growth move, but they were doing it before anyone called it that. Luna: And they didn't have to train customers either. That's a huge hidden cost for most enterprise firms. Lucas: Right. They invested heavily in documentation, community forums, and knowledge bases — but that's a one-time content cost, not a recurring salary for a sales trainer. And because the product was designed for developers, the users were already comfortable figuring things out. By 2007, Atlassian was generating over $40 million in annual revenue with fewer than 100 employees. That's a revenue per employee figure that most SaaS companies envy even today. Luna: But there's a risk, right? If you never talk to customers, you might build the wrong thing. How did they handle product feedback without sales feedback loops? Lucas: That's a smart point. They actually used their support team as a proxy for sales feedback. Support agents were trained to listen for expansion opportunities — a customer complaining about a missing feature might actually be signaling they'd pay more for it. And they built a public feature request platform where customers could vote. It gave them a direct line to what the market wanted, without a salesperson filtering the message. Luna: So the support team became the voice of the customer — and the growth engine. Lucas: Precisely. And they were ruthless about self-service. Even for enterprise customers, the expectation was: start with a self-serve trial. If you need help, there's documentation. Only if you're a very large account — say 500-plus users — would they assign a solutions engineer. But that person's job is technical onboarding, not closing. The price is still transparent, listed right on the website. Luna: Did they ever break their no-sales rule? I seem to recall they eventually hired some enterprise relationship managers. Luna: So the culture held. That's rare — most companies that start no-sales end up building a traditional sales machine within a few years. Lucas: It helps that the founders were stubborn. Mike Cannon-Brookes once said in an interview: 'Salespeople are like nuclear weapons. If you have them, you will use them. And once you use them, you can't go back.' They saw the no-sales model as a competitive advantage — it kept their cost structure lean, kept pricing simple, and forced product quality to be the differentiator. Luna: Did the model survive the IPO? Public markets usually demand that you maximize revenue, which often means building a sales force to squeeze more out of the customer base. Lucas: It survived, but it evolved. After going public, Atlassian moved upmarket — they started selling to entire divisions of large enterprises, not just engineering teams. To do that, they introduced a 'premium' tier with dedicated support and SLAs. But the conversion path is still product-led. A team at a bank might start with a free trial, then upgrade to standard, then eventually to premium. The sales conversation happens only after the product has already won. Luna: So the product is the wedge. And that explains how they reached 200,000 customers — each one a beachhead inside a larger organization. Lucas: Exactly. And it's also why their net revenue retention is so high — over 130 percent historically. Once a team adopts Jira or Confluence, they tend to expand usage across more teams and more products. The no-sales model actually accelerates that because there's no friction: existing users can add seats or buy a new product with a credit card, no phone call required. Luna: It's almost like a consumer SaaS approach applied to enterprise. I can think of a few companies trying that now — like Canva and Figma — but Atlassian was a decade ahead. Lucas: Absolutely. And it's worth noting that not every startup can pull this off. It works best when the product has a viral quality — when one user's adoption naturally pulls in teammates. It also helps when the buyer is a technical user who prefers self-service over a sales pitch. Atlassian's product— for developers, by developers — was perfectly suited for that. Luna: What about the downside? Did they ever lose deals because they didn't have a sales team? Lucas: Probably. They've admitted they lose some large deals to competitors like Microsoft or ServiceNow that offer a more traditional enterprise sales experience. But they argue that those lost deals would have required a high-cost sales effort, and the ones they win — the ones that come through product-led adoption — are significantly more profitable. They'd rather have 200,000 happy, low-cost customers than 2,000 high-cost ones. Luna: That's a trade-off most companies wouldn't make. But it clearly worked for them — the market cap is over $60 billion as of mid 2026. Lucas: And it's a reminder that the default assumption — you need a sales team to sell enterprise software — isn't necessarily true. Atlassian proved that if you build a product that solves a real problem, price it transparently, and make it easy to try, customers will sell themselves. It's a model that's gotten more attention lately as 'product-led growth,' but they were doing it before the term existed. Luna: I think there's a lesson there for founders who are early stage and worried they can't afford a sales team. Maybe you don't need one. Lucas: If today's conversation gave you a new way to think about your own go to market strategy — whether you're building a startup or just curious about how billion-dollar businesses run — and you find value in this show staying ad-free, listener support is what makes that possible. The way to keep that going is buy me a coffee dot com slash fexingo. No pressure, just a genuine acknowledgment: if you got something useful, that's the way to keep the conversations coming. Luna: Yeah, I think that's fair. We put a lot of research into each episode, and it's nice to know it matters. Lucas: So back to Atlassian — one more thing that I find remarkable: the founders met at university in Sydney in 2001, started the company with a $10,000 credit card debt, and never took venture capital until 2010, eight years in. By then they were already profitable. The no-sales model was part of a broader philosophy of capital efficiency. Luna: That's a whole other episode — bootstrapping to profitability. But for today, the takeaway is: you can win in enterprise without a sales team, if you have a product that's genuinely indispensable. Lucas: Exactly. And that's the bar every startup should ask itself: is your product good enough that customers would miss it if it disappeared? If the answer is yes, maybe you don't need a sales team either.