Rediscover your Forgotten Favorites
(Re)Wrap Your Playlists
Rediscover your Forgotten Favorites (Re)Wrap Your Playlists
Spotify (Re)Wrapped selects 10 songs from your playlists that you haven't listened to in a while to reignite your interest in them.
Executive Summary
Over 9 weeks at Stanford University, under the guidance of Frank Bentley, Director of UX Research at Google, our project focused on "Spotify (Re)Wrapped," a web app designed to help users rediscover music.
My responsibilities included conducting qualitative and quantitative user research, collaborating with engineers to develop the app, and testing it in real-world scenarios. We identified common issues such as difficulty managing large song lists, confusion with playlists, and challenges with Spotify integration.
To improve the app, we recommend clearer instructions, enhanced search features, and better integration with Spotify. This research provides insights into how people interact with music apps and suggests practical improvements for digital music services.
Insights from Literature in the Music Discovery Domain
Music discovery has progressed from basic searches by artist or genre, to using descriptive terms with visual maps, and now to advanced machine learning for personalized recommendations.
Platforms like Spotify and Deezer initially struggle to meet user expectations but improve over time, though finding new songs remains challenging.
Advanced techniques like Principal Component Analysis (PCA) and embedding methods, along with consideration of temporal listening patterns and complex user preferences, enhance music recommendation systems for more accurate and personalized suggestions.
Integrating audio analysis with social context and user-generated tags can improve music discovery, but challenges like the "cold start" problem, popularity bias, and generalizing tags need to be addressed, possibly by applying tags from well-annotated tracks to similar unannotated ones.
Users are shifting from music ownership to exploration and identification, favoring multimedia platforms like YouTube that offer a richer experience through video integration, with serendipity and visual cues like album art enhancing navigation and discovery.
Generative Research Question
How do people categorize their music on music discovery platforms ….. and why do they categorize it that way?
Methods
Conducted semi-structured interviews with 5 participants lasting up to 20 minutes, focusing on playlist creation and music discovery habits.
Participants were guided through playlist creation motivations and categorization methods.
Questions included triggers for playlist creation, decision-making in categorizing music, and platform influence on music search.
2. Led a Contextual Inquiry with Participants wherein they were asked to go through curated playlists to gain deeper insights into creation motivations and decision-making.
Real-time observation of participants creating playlists to document decision-making processes.
Some interesting User Behavior Patterns
Liked Music Library is like a Dumping Site!
Previewing music before deciding to like it..
Edra wants to remember moments with his partner, Carrie finds ideas for her art and tracks her favorite songs, and Brian saves time by using Spotify's "Like" feature instead of making playlists.
Affinity Analysis
“I wake up and I remember the tune of a song that I heard … some months prior and I'm like, I need to hear it again. Just to make sure it's as catchy as I remember. I hear it again after remembering it out of nowhere, and then it usually is as good as I remember.”
Carrie scrolled through her liked songs to “remind” her of any songs that fit the theme of her playlist
Emerging Theme: users want equal parts Nostalgia and equal parts Serendipity!
Product Idea!
What if there was a way for them to listen to a remix of songs they haven’t heard for a while and remember what they have in their “like button dumping ground” better?
Usability Study & Results
Usability Study Overview: Explored how users interact with music libraries, focusing on managing "Liked" songs.
Participants: Included Anna, Mike, and Lucy, who provided insights into usability issues and preferences related to music library management.
Key Findings: Users faced challenges with unclear instructions and navigation on the landing page. They also struggled with playlist editing and transitions between app pages and Spotify.
Detailed Findings and Recommendations: Identified the need for a confirmation button and clearer URL input labels. Recommended displaying playlist names prominently and clarifying the purpose of features like 'Sound Bite'.
Field Study
Our team conducted a field study to determine how well the WebApp integrated into users' daily lives. We provided the app to 11 participants who tested it over 10 days. During this period, we gathered both qualitative and quantitative data to gain insights into their behavior and interactions with the app.
Research Questions
How easy is it for users to rediscover old music using Spotify (Re)Wrapped?
What values did users expect to get out from using Spotify (Re)Wrapped?
Does the app help them achieve their intention?
What triggers the users to use the app? Will they continue to use the app?
What feature(s) do users find most useful?
Participants
11 participants who…
Are between 18-35 years
Have used Spotify for more than a year
Have a Spotify account linked to their email
Believe they have a large number of songs in their Spotify library
Methods
9 Days Experience Sampling (N=8)
30 minutes Post-Study Debrief ( N=5)
User Ratings Survey (N = 7)
Google Analytics (9 days)
Ease of Use
Results
71% of users found the (Re)Wrap feature easy to use. Additionally, 100% of users who visited the dashboard used the (Re)Wrap feature for their playlist or entire Spotify library
Key Value Driver
More users abandoned the Playlist (Re)Wrap compared to the Full Library (Re)Wrap. Users were more likely to (Re)Wrap their entire Spotify library.
Retention
Data privacy concerns led to higher abandonment rates. Qualitative insights revealed that some users were worried about data privacy when connecting with the platform.
🎵
Research (Re)-Wrapped
Concept Implications
People do want to rediscover old songs they’ve saved & forgotten about!
…but they don’t necessarily want to save those songs as a 10-track playlist.
“Forgotten” means different things to different users. Customization is key!
When people rediscover music, they want to listen to it right away, but save it on their own terms.
Most people’s playlists are too short or tightly curated to be worth rewrapping, but rewrapping the library is effective!
Design Implications
People do want to rediscover old songs they’ve saved & forgotten about!
…but they don’t necessarily want to save those songs as a 10-track playlist.
“Forgotten” means different things to different users. Customization is key!
When people rediscover music, they want to listen to it right away, but save it on their own terms.
Most people’s playlists are too short or tightly curated to be worth rewrapping, but rewrapping the library is effective!
References
Knees, P., Schedl, M., & Goto, M. (2020). Intelligent User Interfaces for Music Discovery. Trans. Int. Soc. Music. Inf. Retr., 3(1), 165-179.
de Assunção, W. G., & Zaina, L. A. M. (2022, October). Evaluating user experience in music discovery on deezer and spotify. In Proceedings of the 21st Brazilian Symposium on Human Factors in Computing Systems (pp. 1-11).
Xu, J., Gan, M., & Zhang, X. (2023). MMusic: a hierarchical multi-information fusion method for deep music recommendation. Journal of Intelligent Information Systems, 61(3), 795-818.
Turnbull, D. R., Barrington, L., Lanckriet, G., & Yazdani, M. (2009, July). Combining audio content and social context for semantic music discovery. In Proceedings of the 32nd international ACM SIGIR conference on Research and development in information retrieval (pp. 387-394).
Lee, J. H., Cho, H., & Kim, Y. S. (2016). Users' music information needs and behaviors: Design implications for music information retrieval systems. Journal of the association for information science and technology, 67(6), 1301-1330.