Making Music Social
3 min read
3 min read
Thanksgiving Break, 2017. Rather than spending quality time with our parent like good children who adhere their cultural values of filial piety, my best friend and I decided to spend some time tackling a questions thats been on both of our minds recently as audiojunkies –
How can we make music more seamless in our lives?
I was involved in all aspects of the design sprint from beginning to end over the 5 days.
Spotify Party is a way for users to quickly create an algorithmically generated playlist made up of the favorite music of everyone in the party. The party leader simply selects a style of music and adds members to their party. Spotify then uses the listening data of everyone in the party to curate a group playlist. While all members may upvote/downvote songs, only the party leader can play songs and reorder the playlist.
As a college student living in fairly large suite, my room is often the hangout place. After a Friday night of going out, everyone would gather in the living room and to unwind and share stories about the night. Like most people, we like to have some light music in the background when we hangout.
However, people within our friend group have fairly diverse music tastes and everyone wants to have a say in the songs that play. With multiple opinions and only one DJ, this becomes a difficult situation to manage.
Enter: Spotify Party. With this feature, everyone will feel like they have voice in the music that's being played. Since the playlist is automatically generated from the music tastes of people around the room according to the selected mood, all the DJ needs to do now is press "Play".
To help us structure our time over the 5 days, we decided to follow an modified version of the GV design sprint. As such, we spent Day 1 learning more about our question’s environment.
After researching and sharing our knowledge for a few hours, we realized that our question was far too broad and we needed to narrow it down if we wanted to continue with the process.
As such, we focused instead on detailing the different type of music listeners and the means with which they listened to their music.
We quickly decided to interview 5 users about their experience with music instead of interviewing stakeholders and experts within Spotify itself.
After these interviews, we did a round of affinity-mapping with our notes and discovered some valuable insights that would heavily influence the direction of the rest of the sprint.
The back-to-back interviews were exhausting, but the information gleamed was extraordinary valuable.
Day 2 was about diving deeper into our product idea and figuring out how it could potentially exist and how we could feasibly create something testable to validate our idea.
As such, we spent the day bouncing ideas off of each other from public geo-fenced hubs for creating music for any physical space to wearable apps that allow you to see the mood of your close friends based off of their listening history.
Needless to say, my brain was fried. 🍳
After a short lunch and a much needed break to replenish lost gray matter, we got back to work and started critiquing our ideas and designs to identify the ones with the msot potential.
Inspired by the recent release of Spotify.me which showcases the amount of data Spotify has on its users, we ended up moving forward with the concept that seemed the most promising:
A way for friends to source music from those around them seamlessly through their Spotify data.
Deciding to build off of Spotify's ecosystem was a tough decision but we decided to optimize for a solution, rather than a viable independent business idea.
On the third day, God said
On the third day, we went back to brainstorming, but this time for how we could potentially test this idea with a concept that could validate/invalidate its core principles.
We ended up going with a mobile concept that would exist inside of the app itself. Friends would be able to add other friends to a party room where they can simply create the best playlist that suits everyone’s music tastes based on their Spotify data.
Day 4 was all about building that prototype and getting the testing environment set up. Principle's simple UX was real blessing here.
On day 5, D-Day, we gathered a group of five students to test the feature using the Principle prototype.
From these five tests, we learned that while people loved the idea of an algorithmically generated playlist to match the music tastes of the room, we needed to re-evaluate the way in which users are given control over this playlist. The current method of upvoting/downvoting adds a trivial benefit in exchnage for a nontrivial amount of UX complexity. We agreed that for any potential iteration 2, that feature would need to be axed.
© Albert Dong, 2019. All Rights Reserved.