Track My Thesis: From Music Nerd to Product Owner

Research & Discovery

To kick off the project, I conducted both primary and secondary research to understand:

• How users currently discover new music
• What they like or dislike about Spotify’s existing recommendation experience
• Their familiarity with audio features like danceability, energy, or valence

Key Insights

• Most users feel that Spotify's autoplay feature plays songs tend to be repetitive.
• Lack of customization features in Spotify song search system.
• The Spotify discovery feature has a fairly broad scope of classification.

User Persona

Meet Kanina Araya — a 22-year-old music enthusiast from Yogyakarta who listens to music up to six hours a day. As a college student and regular concertgoer, she’s always on the hunt for new indie tracks with specific audio vibes. Her biggest struggle? Spotify’s recommendations often repeat familiar songs and can’t filter music by the unique characteristics she craves. This led us to design Brokoli around her habits, goals, and frustrations — which you can see reflected in her user journey below.

User Journey

To understand how users interact with Brokoli, we mapped out a typical journey. It starts with boredom — Kanina gets tired of Spotify’s autoplay repeating familiar songs. She then opens Brokoli, chooses a discovery mode, and receives personalized recommendations, which sparks excitement.

As she tweaks filters, the results feel more relevant, and she saves the playlist to her Spotify. The next day, she returns to explore new playlists with adjusted filters. This journey reflects an emotional shift from feeling stuck to satisfied, highlighting Brokoli’s role in refreshing her music experience.

Design Solutions

The product was designed to let users:

1. Sign in via Spotify
2. Choose a reference point (artist or playlist)
3. Manually set audio feature ranges
4. Receive a new, custom playlist based on that input
5. Save the playlist or refine the filter further

This flow puts users in control of the discovery process, enabling experimentation while maintaining Spotify integration for ease of use.

Wireframing

Low-fidelity wireframes were created to sketch out each screen and iterate quickly on layout and logic:

• Homepage with simple CTA
• Reference selection screen
• Audio filter slider UI
• Playlist preview and save states

The interface was kept intentionally minimal to allow the core concept — tuning your own recommendations — to shine.

High Fidelity UI

Once the flow and structure were validated, I translated them into high-fidelity designs using Figma.

I applied a music-tech visual language inspired by Spotify’s brand but distinct enough to stand alone.

Key design choices:

• Soft green palette for accessibility and harmony (and to get that Spotify familiarity)
• Modular cards for song previews
• Sliders for each audio feature (e.g., energy, valence, acousticness)

Development

Results & Learning

• Successfully deployed a working MVP that generates playable Spotify playlists
• Learned to balance user customization with good UX (not overwhelming users with too many sliders)
• Validated that productizing audio features can be a compelling hook for music lovers

This project was a pivotal learning experience where I owned the entire product lifecycle — from identifying a user need, translating it into product strategy, leading design, and shipping a usable product.

Through it, I proved that I can think like a product owner, design like a user advocate, and ship like a builder.

Next Steps?

If developed further, the product could benefit from:

• AI-assisted filters (suggesting optimal values for mood-based playlists)
• A community feature to share and remix filters
• Improved performance and responsiveness for mobile users