What a streak-break cohort is for — and what it is not

Streak counters are a design choice, not a law of retention. They also create a group of people who are easy to insult in a table: those who missed a day. Many exports we see dump those people into a single “churned” bucket the morning after the miss. That bucket then gets a win-back message written as if they had left the country.

A streak-break cohort is a named group of people who completed the ritual, then skipped at least one expected day, then completed it again inside a window you choose — we often start with seven days for a daily ritual and fourteen for a weekday-only ritual. The point is to watch how they return, not to celebrate that you frightened them back with a badge at risk.

In a running-habit product, we saw that people who broke a streak on Friday and ran on Sunday had a quieter following week than people who had never been shown a streak at all. The counter was doing something. Whether it was useful was a different question, and you cannot ask it if the export never keeps the break as a state.

What a streak-break cohort is not: a moral ranking, a scare dashboard, or a reason to add a second reminder at 22:00. If the product’s own ritual includes rest days — physiotherapy, for example — a “break” that lands on a prescribed rest day is not a break. Put rest days on the canvas or you will nag people for following the clinic.

I have a mild preference against streaks in products that already carry religious or clinical weight. That is a design opinion. The cohort still needs a name even if you later remove the counter. People will miss days. You should know what happens next.

If this note names the argument in your team, send the ritual and the week people vanish.