Latest / The CTO Podcast with Fexingo: Technical Leadership, Architecture, and Engineering Org / How Basecamp Runs Engineering With Two Designers
Transcript
- Lucas: So Basecamp has roughly forty people in total, and the engineering org is maybe fourteen or fifteen of those. And they ship software that millions of people rely on every day — project management, team chat, email — with just two designers. Luna: Two? That's a ten-to-one engineering to designer ratio. Most companies I've talked to are closer to five-to-one, sometimes three-to-one. Lucas: Right. And they don't have a dedicated QA team. There's no formal product management layer in the traditional sense. The founders — David Heinemeier Hansson and Jason Fried — they're still writing code and making product decisions directly. Luna: So how does that not turn into chaos? I mean, David is the creator of Ruby on Rails, so he's got serious technical depth, but still — two designers for everything? Lucas: The key is that Basecamp enforces extreme focus. They don't try to do everything at once. Each six-week cycle, they pick a handful of features — usually three to five — and assign one small team per feature. That team owns it end to end: design, implementation, testing, shipping. Luna: One team per feature. So the designer is embedded in that team for the cycle. Lucas: Exactly. And because the feature scope is small — they call it 'shaping' before the cycle — the designer can stay heads-down on that one thing. No context-switching across five different projects. The engineer and the designer are sitting together, talking through trade-offs in real time. Luna: That feels like it only works if you're disciplined about saying no to scope creep. Lucas: They are ruthless about it. If a feature doesn't fit in six weeks, they cut scope — they don't extend the deadline. They'd rather ship a smaller version that works perfectly than a bloated version that's buggy. David has a line he uses: 'the best way to get something done is to stop working on it.' Luna: That's almost Zen for software. But what about the async communication piece? Basecamp is famously remote-first, even before the pandemic. Lucas: Their entire communication model is built around their own product — Basecamp. No Slack, no real-time chat for urgent stuff. They use a concept called 'the cool-down period' — after they send a message, they wait a few hours before expecting a reply. The assumption is that everyone is doing deep work, not monitoring a chat window. Luna: So no standup meetings either? I've heard they don't do daily standups. Lucas: Correct. Instead, each person writes a short update — what they worked on, what's next, any blockers — in a shared document at the end of their day. Everyone reads it at their own pace the next morning. It's asynchronous standup. They claim it saves about thirty minutes per person per week. Luna: And that adds up. For a team of fifteen, that's over a hundred hours a year not spent in meetings. Lucas: And it's not just time — it's cognitive load. You don't have to context-switch into a meeting at nine a.m. You process the updates when you're ready. Now, the obvious question is whether this works at scale. Basecamp has been around since 1999 and has deliberately stayed small. They've never raised venture capital, so they're not under pressure to grow headcount. Luna: That's the key distinction. Most CTOs listening are probably at companies that are growing fast — they need to hire twenty engineers this year, not stay flat. Can you take Basecamp's model and apply it to a hundred-person engineering org? Lucas: I think you can take the principles, not the exact ratios. The idea of one team per feature, with a dedicated designer, and a six-week cycle — that scales if you have multiple teams each running their own six-week cycle. The async communication norms definitely scale — in fact, Slack becomes noise at scale, so ditching real-time chat helps more as you grow. Luna: But the no qa thing? That seems risky when you have a hundred engineers shipping code. Lucas: Basecamp's argument is that QA is everyone's job. The engineer who wrote the code is responsible for testing it. The designer reviews the implementation against the mockups. And they have automated tests — not just unit tests, but integration tests that run against the full stack. David has said that if you need a separate QA team, it means your code isn't testable enough. Luna: That's a strong opinion. And one that not every CTO would agree with. But it does force you to write better code up front. Lucas: Exactly. It creates a culture of ownership. The person who writes the code is also the person who has to fix it if it breaks in production. There's no safety net. That aligns incentives in a powerful way. Luna: Speaking of aligning incentives — if these conversations are useful for what you're building or running, a couple of dollars a month is genuinely what keeps these going. Buy me a coffee dot com slash fexingo, if you've gotten something out of them. Lucas: Yeah, it's a small thing that makes a real difference for us. Keeps the podcast ad-free and lets us dig into details like this. Luna: Alright. Back to Basecamp. So they ship without QA, they have two designers, they use async standups. What's the one thing that a CTO at a larger company could implement starting Monday? Lucas: The cool-down period. Just try it for one week: after you send a message, don't expect a reply for four hours. Let your team know that it's okay to ignore notifications for a block of deep work. The productivity gain from that single change is massive. Luna: And what about the fear that something urgent will get missed? Lucas: Basecamp's answer is that almost nothing is truly urgent. If a server is on fire, you call someone — you don't send a chat message. For everything else, it can wait a few hours. And when you do have an emergency, you have an explicit escalation path. But those should be rare. Luna: So the lesson is: design your communication norms to protect deep work, not to maximize responsiveness. Lucas: Exactly. And Basecamp proves it's possible with a small team. The challenge for a CTO is to bring that same discipline to a larger org — to resist the gravitational pull toward real-time everything. Luna: I wonder if we'll see more companies adopt this as remote work becomes more established. The cost of synchronous communication is higher when everyone is in different time zones. Lucas: That's actually the trend. Many of the newer remote-first companies are borrowing from Basecamp's playbook. GitLab, for example, has a similar async-first culture, though they have a formal QA team. But the underlying philosophy — write things down, assume good intent, work at your own pace — that's spreading. Luna: And it starts at the top. If the CTO sends messages at 10 p.m. and expects replies, the culture will follow. Lucas: Right. David and Jason model the behavior they want. They're known for taking weekends off, posting about their hobbies, and not apologizing for it. That sends a signal that work-life balance is not just tolerated — it's expected. Luna: Maybe the biggest takeaway for any engineering leader is that constraints — small teams, fixed cycles, limited designers — can actually produce better outcomes than unlimited resources. Lucas: I think that's the counterintuitive insight. More people, more meetings, more tools — that often slows you down. Basecamp shows that less can be more, if you have the discipline to enforce it. Luna: Alright, that's a good note to end on. Next time we'll look at how a company that grew from ten to a hundred engineers kept its culture intact — or didn't.