My Process

Discover

How do users typically share and discover music? Those were the two primary questions I had going into the research phase. I started by looking at competitors in the US music streaming business and created some provisional personas based off of past experience and blog posts I have read.

Competitor Analysis

Click to Enlarge

Survey & Interviews

I conducted 4 in-person or phone interviews and issued a survey that received 17 responses from individuals in my social network. Some responses from the survey and the interview are below.

Click to Enlarge

Interview Excerpts

Define

As a result of the interviews and survey, I discovered that there are generally two camps of users: those that are casual listeners (i.e. they like to turn on playlists with any set of songs or listen to songs suggested by their friends), and those that are music aficionados (i.e. they love to listen to remixes and underground artists, and are keen on sharing specific songs with individuals).

Personas

I used these findings to create two personas: Jennica the casual listener, and Antonio the aficionado. From real feedback and user details I crafted the personas’ pain points, goals, and user flows. Users like Antonio would like to share music with their friends nearby more easily, even if they do not have Spotify. Users like Jennica are more concerned about finding music was recommended to them more easily.

Click to Enlarge

User Flows

How Might We Statements

With these established, I brainstormed using how might we (HMW) statements to find solutions that would connect these two disparate types of users.

Click to Enlarge

Design

Considering the tasks that both Jennica and Antonio needed to accomplish, I decided to design three different features: commenting, news feed, and a broadcasting feature. 

The news feed and past shared feeds would allow Jennica to find songs that she heard in the past more easily.

The broadcasting feature would help Antonio share music with friends nearby, even if they do not have Spotify. Sharing to multiple people nearby would automatically save that song in the Shared history feed.

The commenting feature was a solution for both users to be able to share their opinions on music on the platform directly, and was a desired feature from the user feedback.

Below are some of my original mock ups and then the final prototype product built in InVision.

Wireframes

Deliver

I used remote testing on Maze.design to gather feedback. I chose remote testing because I wanted to get a larger volume of testers to try the prototype in a short amount of time. The test consisted of four missions, or scenarios, that explored how users would:

1. Share a Song to Friends nearby (Broadcast feature)

2.  Comment on a Song

3. See what Songs their friends liked

4. Find songs recently shared to them

I received 14 responses in 3 days. The main results of the tests show that users were confused about the Broadcast feature and found the newsfeed mission confusing.

I observed the heatmaps, misclick rate, bounce rate, and user comments on the prototype to improve on my mock up screens. Overall, the changes I made were on font size, logo placement, and wording (for both the missions and the social feed and broadcast feature).

Test Results

Iterations and Adjustments

Lessons Learned & Next Steps

I conducted a second set of tests and gathered feedback from 13 more users online. After modifying the UI and the wording of the tests, the success rates of the missions increased overall.

I conducted a second set of tests and gathered feedback from users online. From using primarily remote testing method, I learned:

Wording and names have a huge impact on the design

Users were introduced to new features in an already familiar app, making it challenging to guide them into the right direction. I hypothesize that some of the users deviated from the path because of the way the scenarios were worded.

All user pain points cannot be solved with one feature

The original task was to create one new social feature, but after uncovering and developing my personas, it was difficult to fit all problems into one box. At that time I had to think of features that would become compatible with each other.

In the future, I look forward to seeing if this feature will be considered in Spotify’s web community for enhancements.