FocusA self-directed case study in progressive personalization, grounded in App Store review mining and real usage friction.

FitOn offers free, celebrity-led fitness videos and builds a "personalized plan" from a handful of upfront preferences, but gives users almost no way to adjust that plan once it exists. As a regular user, and a busy mom with limited workout windows, I used this as a self-directed case study in progressive personalization.
Once a plan is generated, there's little users can do to refine it further. Favoriting a workout drops it into a separate section that's disconnected from the plan itself, and as new workouts are uploaded, that favorites list becomes harder to navigate. Users can't delete workouts they don't want or reorder the ones they do.
I combined personal usage friction with a pass through App Store reviews, which surfaced the same three requests repeatedly:



Deeper plan customization, unsolved by either app
4.9★ · 138K ratings
4.8★ · 419K ratings
Even the larger competitor hadn't solved deeper plan customization, leaving real room for FitOn to differentiate.
Wireframes:





High-fidelity final designs:





I scoped delivery in four phases by build cost and risk, rather than proposing everything at once:
FitOn already uses this taxonomy elsewhere in the app, so it ships first.
Test against the existing plan-creation flow, 2–4 weeks to measure engagement and return visits.
Quick to test once footage exists, and reusable across the rest of the app.
FitOn didn't yet track the signals a recommendation engine would need, so it waits.