Pandora invented the personalized radio station. It pioneered the Music Genome Project and introduced a generation to the idea that an app could learn what you want to hear. But as competitors built their interfaces around the music, Pandora kept building around the app. This redesign asked one question: what would it look like if the music came first?

Challenge

Users weren't losing trust in the algorithm. They were losing patience with the interface built around it. Core interactions like thumbs up and down, Pandora's most direct feedback mechanism, became inconsistent and hard to hit across device sizes. The information architecture had accumulated years of feature additions without consolidation, leaving the sidebar fragmented and visually noisy. And the now playing state, the primary reason a user opens a music app, was visually subordinate to everything around it.

Solution

Navigation was consolidated to four items reflecting actual user behavior patterns, not feature inventory. Feedback controls were repositioned as the dominant interaction layer in the now playing view, with no competing visual weight. The card and grid system was replaced with full-bleed editorial photography, giving the listening experience visual hierarchy that matches its functional priority. Mood-based discovery made the recommendation logic legible to the user, shifting the experience from opaque personalization to transparent curation.