Music Discovery

ux case study


This project was undertaken during a 5-day design sprint as part of General Assembly’s User Experience Design course in London. The purpose of the project was to identify a user problem within our field of interest (I chose music discovery), and then implement user-centric design practices to deliver a solution to this problem.


Deliver a full suite of UX documentation for a digital product, from user personas and wireframes to interactive prototypes.

Research + Defining the Problem

To begin the research process a competitive analysis was undertaken to identify existing music discovery solutions on the market. Strengths, weaknesses and core features of competitor products were highlighted and mapped out to identify potential product, feature and market gaps.

One-to-one interviews with potential end users were carried out to gain insights into the current pain-points people were experiencing. The interview topics started broad and looked to uncover the ways people are already discovering music. Questions were then narrowed to focus on the most popular music discovery methods (blogs, apps, and friend recommendations) in order to uncover pain-points or unmet needs within these.

Despite most people being fairly happy with their discovery method of choice, it was interesting to hear that there was an unmet desire to try out new genres of music that was not being adequately met by existing streaming platforms (which tend to offer recommendations based on listening history), or through friends (who tend to share a similar music taste, meaning the same tracks and artists get shared between the group).

“I’m starting to get bored with the music I listen to. It makes me happy to find new tracks, and I enjoy sharing my discoveries with friends (it shows I have taste…), but I’m finding it hard to get into new types of music.”


1) People want to try out new genres, but don't always know where to start
2) People want to stop listening to the same tracks over and over again
3) People liked to send music to their immediate friend group, but don't feel comfortable sharing tracks publicly

The insights gained from user research were used to construct an end-user persona whose needs and frustrations were collated into a problem statement that would form the basis for generating design solutions:

“Chris finds it hard to branch out into new styles of music because recommendations are too genre-confined”

Insights gained were documented on post-its and categorised based on common trends


After doing a ‘crazy-eights’ ideation session (where eight potential solutions were quickly sketched out using pen and paper), the most promising (and feasible) ideas were converted into a user-flow storyboard, from which paper prototypes were made.

'Crazy Eights' pen and paper ideation

Low-fidelity prototypes used to quickly test ideas

Usability Testing

To get both conceptual and usability feedback, three rounds of testing was undertaken, with concept refinement and design iterations occurring after each. Initial 'high-level' usability testing was done using low-fidelity paper prototypes. The insights gained from paper prototyping were used to refine and adjust the medium-fidelity wireframes that were build using Sketch and InVision. From there aspects of visual design (such as typography, colour schemes and imagery), alongside further testing insights, were used to create the high-fidelity interactive prototype used for the third round of testing.

Test Tasks:

1) Create a custom playlist
2) Save a track to your library
3) Share a track with a friend


1) Input fields on home screen not clear enough. "Do I enter an artist, genre, mood…?"
2) Users want to save the songs they like to their existing music collection
3) Saved track page caused confusion as to how tracks, albums, and artists would be organised
4) Being forced to input a search phrase was causing cognitive over-load for some testers (who struggled to remember things like artist names, or genres/tracks they enjoyed). Casual listeners wanted to be able to easily generate new playlists suggestions with minimal effort.


1) Playlist phrase on the home screen re-structured, with an additional input field added
2) Mandatory Spotify sign up to allow users to save tracks to their existing music collection, and generate recommended playlists based on listening habits
3) Removal of saved track page to re-focus the app purely on music discovery
4) 'Refresh playlist' added to help users generate new playlist ideas easier

Implementing these solutions resulted in 100% of 'high-fidelity' testers completing all three test tasks with no usability issues

The Result

By putting the user-first, testing often, and following a 'less is more' approach, a simple, intuitive solution was designed.

1) After connecting with Spotify the user sets up their playlist tags using the phrase on the home screen

2) Users can 'pull-to-refresh' to generate a new suggested playlist

3) Moving the slider right increases variation within the playlist from the initial input tags

4) After each play variation increases, gradually introducing new types of music to the listener

5) Listeners can save tracks to Spotify, and send to their friends either privately or publicly