Latest / Startup Stories with Fexingo: Conversations About Founders, Funding, and Building Companies from Zero / How GitLab Built an Open Source Unicorn Without Leaving the House
Transcript
- Lucas: So GitLab went public in October 2021, valuation north of fifteen billion dollars. But here's the part that still catches people off guard — they had no headquarters. Not one office. Ever. Thirteen hundred employees in sixty-five countries, and the company had literally zero square feet of leased space. Luna: And they weren't just remote-first in the pandemic work-from-home sense. They were remote-only by design. Sid Sijbrandij, the co-founder, decided that before they even had a product. Lucas: Exactly. The origin story is pretty specific. Sijbrandij was living in the Netherlands, and his co-founder Dmitriy Zaporozhets was in Ukraine. They met through an open source project — GitLab was originally a Ruby on Rails app that Dmitriy built to help teams collaborate on code. Sid saw it, realised this could be a business, and they started working together across two time zones. Luna: So the remote structure wasn't a lifestyle choice. It was just... the only way they could work. Lucas: Right. And they leaned into it hard. Instead of treating distance as a problem to solve, Sijbrandij turned it into a competitive advantage. He wrote a public handbook — it's still online, thousands of pages — that documents every single process, every decision, every meeting note. Anyone inside the company, or outside it, can read how GitLab decides things. Luna: Is that not exhausting? Total transparency? I mean, doesn't that slow things down when every decision has to be documented? Lucas: That's the counterintuitive part. Sijbrandij's argument is that documentation is actually faster. Because once you write something down, you never have to explain it again. New hire joins? They read the handbook. Someone asks 'why do we do X?' — link to the relevant page. The company runs on what they call 'low context' communication. No meetings to re-establish shared understanding. Luna: Low context — meaning the information lives in the document, not in someone's head. Lucas: Yes. And it scales. When GitLab was ten people, you could just ask Sid a question. When it's thirteen hundred people in sixty-five countries, you cannot. So the handbook becomes the operating system. There's a page for how to expense a meal, a page for how to run a performance review, a page for how the board meetings are structured. Luna: I remember reading that they even have a page about how to have a coffee chat with a colleague. That level of detail. Lucas: It sounds absurd until you realise that in a remote company, informal hallway conversations don't happen. So you have to engineer serendipity. GitLab created something called 'coffee chat' — a bot that pairs random employees each week for a fifteen-minute video call. No agenda. Just get to know someone. Luna: That's kind of brilliant. And it's all documented. So anyone can replicate it. Lucas: That's the open source ethos applied to company culture. And it's a big reason why GitLab's S-1 filing was so unusual. When they filed to go public, they published the entire prospectus on their website with a merge request — literally a pull request where people could comment on the legal language. Luna: Wait, they invited public comments on their SEC filing? That's insane. Lucas: They did. And people actually submitted edits. Some were useful, some were silly. But it sent a signal: this is how we operate. Everything is open, everything is a draft, everything can be improved. Luna: Did any of the comments actually change the S-1? Lucas: A few. Nothing material to the financials, but some of the narrative language got tightened. The bigger point is that GitLab's market cap at IPO was roughly fifteen billion dollars, and the company had spent almost nothing on real estate. Their biggest expense was people. And because they hired globally, they could pay competitive salaries in local markets without the Silicon Valley premium. Luna: Right, so their cost structure was fundamentally different from a traditional enterprise software company. Lucas: Completely. Their operating margin at the time of the IPO was negative — they were still investing heavily in growth — but the unit economics were strong. Customer acquisition cost was relatively low because the open source community acted as a funnel. Developers try GitLab for free, love it, then convince their company to buy the enterprise version. Luna: So the product sells itself. But also, the remote structure means no one is paying for offices in San Francisco or New York. Lucas: Exactly. And that's a durable advantage. Sijbrandij once said that if GitLab had been in a traditional office, they would have needed to raise three times as much venture capital just to cover rent and furniture. Luna: I want to push back a little though. The all-remote model works for GitLab because they sell developer tools. Developers are already comfortable working async, communicating through tickets and merge requests. Would this work for a consumer company or a hardware company? Lucas: Probably not in the same way. But the principles translate. The idea that you should write things down, that you should default to transparency, that you should optimise for asynchronous communication — those are generally useful. Buffer tried a similar model. Zapier has been remote since the beginning. But GitLab is the most extreme example because they codified everything into a handbook that's literally public. Luna: And they've been public for four years now. How is the model holding up post ipo? Lucas: Mixed. The stock has been volatile, like a lot of growth tech. But the operational fundamentals are solid. Revenue has grown year-over-year, and they've maintained gross margins above eighty-five percent. The remote structure didn't crack under the pressure of being a public company. If anything, it helped them navigate the pandemic more smoothly than companies that had to scramble. Luna: It's interesting — they were already living the future while everyone else was still in offices. Lucas: Right. And now that many companies are trying to force people back to desks, GitLab is out there saying 'we never left.' They're hiring from anywhere, they're not mandating vaccines or office days, and they're still growing. It's a genuine counterpoint to the return to office narrative. Luna: But there has to be a downside. What do they miss? What's the cost? Lucas: The biggest cost is probably innovation through informal collision. In an office, you overhear a conversation, you jump in, something unexpected happens. GitLab has to manufacture those collisions through programs like the coffee chat, through team offsites — they do bring people together a few times a year — but it's not the same. Sijbrandij has acknowledged that breakthrough ideas might be harder to come by in a fully remote setting. Luna: So they trade spontaneous creativity for operational efficiency and lower costs. Lucas: That's the trade. And for a developer tools company, it's a good trade. The product itself is about making software development more efficient. So the company's culture mirrors the product's value proposition. That alignment is probably one reason it works. Luna: It's a clean story. But it also feels fragile — if the CEO leaves, does the culture survive? Lucas: That's the real test. Sijbrandij is still in charge. He's been CEO since day one. The handbook is a powerful artifact, but it's a document. It doesn't enforce itself. Culture is maintained by leadership behaviour. If a new CEO comes in who prefers meetings and memos and closed doors, that handbook becomes a historical curiosity. Luna: So GitLab's long-term bet is that the handbook itself becomes the institution, not any single person. Lucas: That's the hope. And to their credit, they've designed it to be that way. Every page has an owner, every process has a documented rationale. New people can edit the handbook. It's a living thing. But institutions are fragile. We'll see. Luna: Before we wrap, I want to come back to something you said earlier about the cost structure. You mentioned they avoid real estate costs. But there's another angle — the open source model itself. Lucas: Oh, absolutely. GitLab's core product is free and open source. The company makes money from enterprise features, support, and hosting. That's a classic open core model. But it means their R&D is partially subsidised by the community. Outside contributors submit code, fix bugs, write documentation — all for free. Luna: So the remote model and the open source model reinforce each other. Both rely on async contribution, both scale without central coordination. Lucas: Exactly. And both are culturally aligned. You can't have a closed, meeting-heavy culture and also run a thriving open source project. The values have to match. GitLab's remote culture is essentially the open source way of working applied to the whole company. Luna: It's almost tautological. They built a company that works the same way their software does. Lucas: That's the best way to put it. And it's rare. Most companies have a product culture and an internal culture that are totally separate. GitLab made them the same thing. Luna: So the question for other founders listening is: can you do that? Can you align your internal operations with your product's philosophy? Lucas: I think you can try. But it takes a level of intentionality that most founders don't have in the early days. Sijbrandij was writing the handbook before they had revenue. He was thinking about culture as a feature, not an afterthought. That's the lesson — not that every company should be remote, but that every company should ask: what does our product teach us about how we should operate? Luna: And then write it down. Publicly. Lucas: And then write it down. Publicly. That's the GitLab way. Luna: Quick honest thing — a handful of listeners chip in monthly through buy me a coffee dot com slash fexingo, and that literally funds making this many of these. Lucas: It does. No ads, no sponsors, just people who find the conversations useful. And we try to earn that support. Luna: Yeah. Anyway, back to GitLab — one last thing. What's the one number that captures their model better than any other? Lucas: I'd say it's the ratio of handbook page views to employee count. They've published that the handbook gets roughly one million page views per month. That's almost eight hundred views per employee. Every month. People are reading the manual. That tells you the culture is alive. Luna: That's a great number. Alright, I think that's enough to chew on. Lucas: Yeah. GitLab proves that a company can be built on documentation and trust. Whether that scales beyond a certain size — well, that's the experiment we're all watching.