Six months of use earns a product

FamilyLock isn't a product. It's where I find out which of my ideas survive contact with an ordinary Tuesday.

Everything in there got built on the day the friction became unignorable. Voice to tasks, because I kept losing the ten-second thoughts I had while holding a drill. Personal Kanban, because the paper list kept lying to me. Spaced repetition, because I read things and then don't have them. FieldBase, because tracking jobs across sites in my head quietly stopped working.

The rule I use

If I'm still using it six months later, and not out of loyalty to the version of me who built it, it has earned the right to be rebuilt as a product. If I'm not, it stays in the lab and I've learned something a lot cheaper than a launch.

Building for a user I understand completely means the feedback loop is same day.

The personal Kanban is the clearest case. It started with swimlanes, WIP limits, effort points and three views. It lasted about eleven days. What survived is three columns and a rule that anything untouched for two weeks goes back to the top of the list to be defended or dropped.

What I got wrong

For a long stretch I treated I built it as if it meant it works. Those are two different claims and only one of them is checkable.

A tool I abandoned in week three isn't a failed product. It's a product I didn't have to launch to find out about.