We did not set out to build a methodology. We set out to stop arguing about when things would be ready. The rule was simple: whatever is finished on Friday goes out on Friday, and whatever is not finished waits for the next one.
The rhythm does the managing
A fixed cadence removes an entire category of conversation. Nobody negotiates a date, because the date is already known. The only remaining question is scope, which is a far more productive thing to argue about.
When the release date is fixed, work gets smaller. That is the whole trick.
What surprised us
- Estimates got better by getting shorter. Nobody can misjudge a five-day task by three weeks.
- Quality went up, not down. Small releases are easy to review and easy to roll back, so people took more risks in the right places.
- The backlog stopped being a graveyard. Anything that survived twelve weeks without shipping was quietly admitted to be something we did not actually want.
What we would change
We should have protected the last hour of Friday from the start. For the first few months the release itself ate the retrospective, and we lost the part of the loop that made the loop worth having. Ship, then talk about the shipping. In that order, every week.
