Looking into why, we found that people didn't fully get our offering, didn't understand our license model, and struggled to find the right music to hook them. One problem was the first view users landed on, the discover view. Most people never scrolled past the first screen. The content wasn't relevant enough, it felt static, and the covers were so big that seeing anything took a lot of scrolling.
I led the design and the implementation. We started wide: research and workshops to understand what first-time users needed, and work with tech in parallel on what was possible with the content and the system we had. Implementation was mostly about finding a technical solution that held. There were a lot of constraints, so much of the design work went into making things feel more dynamic than they really were. We tested with users and iterated, and a second round of workshops brought the strongest parts together into one direction.
A three-step setup flow got to know first-time users early, enough to surface relevant suggestions from the start. We also added clearer guidance on what was needed before publishing a track, more hand-holding than before to clear up the license confusion.
The discover view itself was redesigned around carousels, a familiar pattern that made our catalogue easier to scan and understand. Content was grouped as a mix of curated, new, and inspirational, so the breadth of what we offered came through clearly. On top of that, we created a new playlist format to bring in more personalised content.
I conceptualised and drove a new playlist format built around community-based picks and tracks matched to similar content, updated weekly. The goal was to make discovery feel more personal and less algorithmic, something people could trust and return to.
Success rate improved: the time it took to find a track and save or download it got noticeably shorter. More people also completed the steps needed to publish a track, which is where people used to get stuck. Our store rating went up in the same period too, though that's harder to tie directly to this work. The setup flow was later rolled out on web as well.
Understanding the technical limits early opened up creativity rather than shutting it down. Once we knew what the system could actually do, we could put the effort into getting the most out of it instead of designing for something we were never going to get. Working close to the engineers was what made that possible.
