Getting it right the last time

I added a button yesterday morning. By the afternoon, I had removed it.

This is the kind of thing entrepreneurs don't like to say out loud. We track progress in one direction. We add features, add pages, add functionality. Going backwards feels like admitting you were wrong, and admitting you were wrong in front of the people who depend on you is uncomfortable.

But the button looked wrong in context. I could see it the moment I stepped back from building and looked at the thing as a user would. It was in the wrong place. It interrupted the flow. The instinct to add it was sound — the placement was not.

So I removed it.

The harder removal

There was another thing I took out yesterday — a feature that had been sitting in one of my tools for months. It worked, technically. It did exactly what the label said. The problem was that the label sent people down a path that did not actually help them. They would follow the instructions, do the thing, and the tool would respond as if everything was fine.

It was not fine. The underlying data the feature relied on was static, not live. The whole interaction was a well-intentioned lie.

I knew this for a while. I left it in because removing it felt like admitting a mistake in public. The feature existed. People had been told it existed. Pulling it back felt messier than leaving it.

That logic is backwards. Leaving a misleading thing in place to avoid the embarrassment of removing it is not prudence. It is just quieter negligence.

Andrew Hunt and David Thomas wrote: "It's not at all important to get it right the first time. It's vitally important to get it right the last time."

I do not know when the last time is. That is the difficult part. Every version feels like it might be the final one, and then it is not. You ship it, you live with it, you see it in context, and you find the thing that needs to change. This is not a sign that you built badly. It is a sign that you are paying attention.

The entrepreneurs I respect most are the ones who can remove their own work without drama. No announcement, no retrospective. Just: this was wrong, here is the corrected version.

Getting it right the last time means being willing to be wrong every time before it.