Homer is an early learning app for children ages 2–6, with a library of 1,000+ lessons across reading, math, and social-emotional development. At the time of this project, the app had strong acquisition numbers but was losing ground on retention — kids were opening the app, but not coming back.
I was the sole product designer on this initiative, working directly with the VP of Product, two engineers, and the brand team. I owned the full design lifecycle: research, wireframes, prototypes, usability testing, design system, and final UI across iOS, Android, and web.
Hero overview showing the full product context.
I ran 14 moderated usability sessions with children ages 2–6 and their parents across in-home, daycare, and remote settings. The goal wasn't to validate — it was to understand what was actually happening before we touched anything.
Research artifacts from 14 usability sessions.
The content library had over 1,000 lessons. Kids couldn't find anything. We rebuilt content discovery from scratch — moving from a flat grid to a curated, character-led home screen.
Design explorations for objective 01.
Every navigation element relied on text labels that 60% of our users couldn't read. We rebuilt the nav system around iconography, character cues, and audio feedback.
Design explorations for objective 02.
The app felt static. We introduced motion, character reactions, and celebration moments — designing for joy, not addiction.
Design explorations for objective 03.
The final design centered on three core changes: a curated home screen driven by Homer's character cast, a fully icon-based navigation system that worked without text, and a motion design language that made the app feel alive.
Final shipped product screens.
We prototyped a recommendation engine, a learning path flow, and a parent dashboard in the first two weeks. All three got cut after testing. Killing them early saved roughly 3 sprints of engineering time.
I built a component library of 40+ components across iOS, Android, and web — the first time Homer had a unified system across platforms. The system cut per-feature design and build time by roughly 30%.
Design system components and user flows.
Final shipped product.
I'd push for a longer research phase with more children under 3. I'd also involve engineering earlier in the design system work — I built it largely in isolation and some component specs needed significant revision before they could ship.
The decision to cut the recommendation engine. Every stakeholder wanted it. The data didn't support it. Holding that line — with research to back it up — was the right call and probably saved the project from scope creep.