£0 to first revenue — the three build calls that mattered

James: First episode, so let's set the tone. No origin myth — just the build. Three decisions between zero and your first paying customer. Go.

Sofia: Call one: we used the most boring stack imaginable. Nothing on that list would impress an engineer. That was the point. I wanted zero surprises everywhere that wasn't our actual edge.

James: Which was?

Sofia: Call two. The matching logic — that was the whole company, so that's the one place we wrote something genuinely novel. I think of it as a novelty budget. You get to spend it once. Everything else is off-the-shelf, deliberately.

James: A lot of first-time founders spend that budget on infrastructure instead.

Sofia: That was nearly call three — and it's the mistake. We almost built for scale we didn't have. Multi-region, queues, the works. You talked me out of it. We served the first hundred users off one small box and a managed database, and it was fine for two years.

James: What did avoiding that buy you?

Sofia: About four months. Four months we spent getting to revenue instead of building for a load that didn't exist. The expensive mistakes at this stage are almost never "too simple." They're premature scale. The first invoice came from software that any senior engineer would call unremarkable — except for the one part that was the entire point.