Notebook

Web software is never done. We've built a whole religion out of it. But nobody asks whether the never-done model is good for the people building it — and the answer is often no.

The competitor ships fifty features. So the plan becomes: we need fifty-one. But the market keeps rewarding the opposite — over and over, you find companies that ship less and win anyway. The question is why.

At some point "listen to your customers" came to mean "do what your customers say." Customer obedience is order-taking. Customer obsession is caring about the problem more than the request.

When a product wins, we reverse-engineer a clean story — the roadmap choices, the focus, the execution. The luck gets edited out. But timing accounts for more than the idea, the team, and the business model combined.

"Make it more intuitive." Everyone nods. But intuitive for whom? "Intuitive" isn't something a design is — it's something that happens between a design and a particular person's history with everything else they've ever used.

I drive past two grocery stores to shop with Linda. I pay 20% more to ship with Peter. I trust a man who identifies screws by feel. Somewhere, a professor is drawing a red X through my decisions.

Judging a product designer by their visuals is like judging a chef by how well they can plate. The real work—the expensive work—is invisible.