PROCESS
My objectives were to first conduct my own qualitative research by interviewing people about their music discovery and streaming experiences. Then based on their research, I ideated different solutions which became a series of task flows. I then created a lo-fi prototype of these task flows and did a series of user testings to determine how to improve and refine my solutions. After I refined my product, I designed a visual branding identity and design system to apply to the final prototype.
RESEARCH
I focused on conducting qualitative research for the foundation of my ideation. I began by conducting seven interviews with people about their process and experience of discovering new music. In addition, I led a participatory design workshop with a group of 14 people to gather data on their desired discovery features using the mental model of Spotify.
TAKEAWAYS
100% of users said Spotify’s algorithm repeats the same music to them in its Autoplay mode.
Users said this happens either every time they use it or most times that they use it.
They are frustrated by this experience and desire more control over what is being auto-played by the algorithm.
The majority of users said they want more variety of music and an easier, faster way to discover it.
Users wanted more recommendations and search capabilities.
"Oh, I won't even hit that I “like” a song anymore. That's a guarantee I will hate that song by the end of the week because Spotify will play it so much.”
ARCHETYPE: “THE MUSIC LOVER”
I then categorized all of this data into themes and created an archetype persona to summarize who my user is. This archetype was defined as “The Music Lover” and included their Job-To-Be-Done, quotes, goals, pain points, motivating factors, inhibiting factors, possible triggers, and influencers based on the research from my interviews.
Job To Be Done: “I want to increase the quality of my daily activities with more new music.”
Quotes:
"When you find a good song, it's like finding the Holy Grail!”
"I don’t just want Folk music, I want Murder Ballads.”
"I don't want to be stuck just listening to like one thing.”
"I don't do the auto-play algorithm stuff at all on Spotify. I just get frustrated with it.”
"I don't want to hear the same thing that like has 40 million plays.”
Goals:
To have a variety of music
To find new music they like easily
To have more influence over what’s played
To have specific playlists to guide activities
To find shows
Pain Points:
Frustrated that Spotify recycles the same songs in the algorithm
It takes too much effort to find good, new music
Navigating from song to album is really confusing and takes too much time.
Genres not specific enough
Motivating Factors:
Friends
Shows
Emotional state
Activity-related
Inhibiting Factors:
Spotify’s algorithm
Spotify’s navigation
Effort and time
Possible Triggers:
Being tired of the same music
Having time to search
More specific genres
More search capabilities
Easier navigation
Recommendations
Influencers:
Friends
Spotify Radio, Discover, Similar Artists, Daily Playlists
Social Media (Instagram, Tiktok)
IDEATION
With this archetype as my guide, I then ideated and explored four different solutions to enhance music discovery: a search filter system, user-generated tags, saved search profiles, and improved recommendations. I then combined these solutions to create three task flows prioritizing the functions of search, tagging, and recommendations.
PROTOTYPE & TEST
Scenario: You’ve just downloaded Mixtape and would like to find an album different from what you’ve been listening to.
Task 1: Use the app to find an album that’s different from your previous listening.
Task 2: Once you find that album, use the app to find a way to remember it for later.
I then built lo-fi prototypes of the three task flows in Figma accompanied by three user scenarios with tasks to accompany each flow. I conducted user testing sessions with six people to gather notes about what was working and not working. I then revisited and refined my prototype to apply what I had learned from these users.
Scenario: You’ve been enjoying the album you previously found and you want to help others in the community find it.
Task: Use the app to find a way to share descriptive words about what the album sounds like and what kind of experience you have listening to it.
Scenario: You’re now desiring more specificity in what music you’d like to find.
Task 1: Use the app to find a mixtape (playlist) that meets the following categories:
Pop
Ambient
Sounds like Nirvana
Is from Australia
Has the marimba in it
Task 2: Once you’ve found a variety of mixtapes that meet these preferences, use the app to find a way to remember it for later.
To make the prototype high fidelity, I created a stylescape to define the tone, typography, colors, photography, and graphic elements. Referencing my archetype, The Music Lover, I wanted to create a sense of nostalgia for both cassette tapes and the experience of shopping at record stores, while at the same time feeling bold and modern since it offers a new form of discovering music. I used Swiss Style as inspiration to refer to cassette tapes with the use of grids, color fields, large sans serif typography, and monotone photography. I added a modern twist through the use of bright, bold colors and its unique layout design for screens.
STYLESCAPE
The logo for Mixtape is both nostalgic and modern. It uses a one line connecting the shapes to the letterforms to reference the tape that runs through a cassette tape. It consists of the same stroke width throughout so that it reads well at smaller scales as well as on top of photography.
LOGO
FINAL LOGO
SKETCHES & IDEATION
FINAL PROTOTYPE
NEXT STEPS
I would like to do another round of user testing in its high-fidelity state and refine it from there. In addition, while I was in my ideation phase, I developed another task flow that was designed for musicians uploading their album to the platform which allowed them to input music recommendations. This would allow them to choose what artists sound similar to their album, what albums inspired their album, and what other artists they recommend. This would then influence what additional listening is recommended when a user is on the artist’s album page. I would like to build out this task flow and target it toward professional musicians to test if it’s a desired feature.