The refactor that took a week and saved me a year

I spent a week last year doing a refactor that shipped nothing. No new features, no bug fixes, no visible change.

It was painful. Every day I felt behind. Every day I looked at it and thought this is a week I will never get back.

Then the next year, every feature took half the time it used to. The bugs got easier to find. The code got easier to hold in my head. The week paid for itself about four times over.

The trouble with this kind of work is that the payoff arrives much later than the cost, so in the moment it feels purely like loss. There is no launch, no announcement, nothing to point at.

I have to remind myself that a week of visible nothing is sometimes the most valuable week of the year.

Have you ever done a big boring refactor and been glad you did? And how do you talk yourself into the next one?