Wellness products have a recognisable retention curve. Signup is strong, the first fortnight looks healthy, and then something falls off a cliff. Teams usually respond by adding notifications. That reliably makes the next cliff steeper.
Week one is borrowed motivation
In the first days the user is running on the decision they already made — to get fitter, sleep better, eat differently. The app is not motivating anyone yet; it is riding something that was there before install. Any onboarding looks good against that.
By week three the borrowed motivation is spent. Whatever the product gives back has to be worth the interaction on its own, and for a lot of wellness apps the honest answer is that it stopped giving anything back around day nine.
Streaks are a loan, not income
A streak works until it breaks, and then it works against you. The user who misses day fifteen does not restart at day one — they uninstall, because the number that represented their effort is gone and the app just told them so.
Designs that survive that moment tend to share a few properties:
- Progress that survives a bad week. Trends over windows, not unbroken chains. A rolling seven-day view forgives a missed Tuesday; a streak counter punishes it.
- Something that changes because of the data. If the plan in week three is the plan from week one, the user has learned that logging changes nothing.
- A defined smallest action. On a bad day, what is the thirty-second version? If the only path is the full flow, bad days become gaps and gaps become churn.
Wearables raise the stakes both ways
Pulling data from a watch or ring removes most of the logging burden, which helps. It also removes the user's sense of contribution, which does not. When the app fills itself in, the reason to open it has to be interpretation — telling them something they could not see themselves.
The metric worth watching
Not day-one retention, and not the streak length. Look at what fraction of users complete a meaningful action in week four having missed at least two days in week three. That number tells you whether the product recovers people, or only keeps the ones who never slipped.
