We raised twice and stayed at nineteen people. People assume this was restraint. It was closer to the opposite: we wanted to move faster, and headcount was the thing slowing us down.

The constraints we chose

Three rules held the whole thing together. No team owns fewer than two surfaces. No project runs longer than six weeks without shipping something a customer can touch. And anyone can say no to a meeting, including the founders.

None of these are clever. What made them work was that we enforced them when it was expensive to do so, not only when it was convenient.

What it cost

We turned down two enterprise deals in 2025 because the integration work would have consumed a quarter. Those were real revenue decisions and they were genuinely painful to make.

What it bought

Decisions that took other companies a week took us an afternoon, because the people who understood the problem were also the people who could ship the fix. That compounding is hard to see on a spreadsheet and impossible to miss from the inside.