Pou Is Here
final_ff.png

Find Friends (Old)

 
 

Find Friends

 

Smule.  01/2016

In January of 2018 only 1 in 5 new users on Sing! follow. The Find Friends (FF) experience was outdated and half baked. The feature was buried in the Feed, and the experience itself often recommended users who were irrelevant or difficult for users to evaluate.

Our hypothesis: Improving usability of the Find Friends feature, and quality of accounts/recommendations surfaced, will increase follows, which in turn will help engagement, social activity, and ultimately retention.

 

Team

  • Design Lead and Strategist: Pou Dimitrijevich (me)

  • Product Manager: Winnie Lee

  • Visual Designer for Android for Phase I: Oscar Corral

  • UX, IxD and Visual Design for Android Phase II: Li-Tien Ou

  • UX, IxD and Visual Design for iOS Phase I and II: Pou Dimitrijevich (me)

 

Game Plan

PHASE I - Improve UX/UI and discoverability.

  • Make Feed more contextual. If a user is not following anyone, the Feed should be empty with only a CTA to ‘Find Friends’. Currently only 17% of D1 users ever click-through to Find Friends.

  • Improve quality of recommendations.

    • Prioritize real-life friends and social users.

    • Only show users with a profile photo.

    • Filter out new and dormant accounts.

    • Filter out spammers.

  • Improve Find Friends cell.

    • Include contextual components such as: reason for connection and profile blurb. (i.e. Also likes <genre>, Friends on Facebook, You both sang <song title>, Popular in your region, Hot Influencer!, Follows you, Power user, In your contacts (phone/email)

  • Surface ‘Friend Recommendation’ in areas outside the Feed.

    • Initially: Feed and Songbook

    • Future possibility: Explore and Profile.

PHASE II - Connect users with their real-life friends earlier.

  • Surface FB friends during on boarding if the account was created using FB auth. (Estimated 30% of Sing users register with Facebook)

PHASE III - Expand Find Friends to phone and email contacts.

  • 79% of Sing iOS users and 30% for Sing Android users register with their email or phone number.

 

Risk and Challenges

PHASE I

  • Content shown on the feed is 90 days and younger, so if you follow a user that has content older than 90 days, then there may also be nothing to show.

  • If user is following only 1 person, follower may have no content. We can counter this by creating a ‘Recommended for you’ module that displays w/in the Feed.

PHASE II

  • Additional steps to on boarding may hurts throughput.

PHASE III

  • GDPR and permission settings.

 
 
ff_userflow.png
 
 

Existing UI / UX

ANDROID

iOS

 

Wireframes

ANDROID

 

Research / Learnings

RESEARCH

I did a competitive analysis of apps that put a lot of emphasis on community; i.e. Instagram, Twitter and Musical.ly. What I learned was this: there are two things users care about the most when deciding who to follow; the first is content, the second context.

SPLIT TEST

To mitigate the main risk of Phase 1–show an empty Feed when a user has 0 follows–we ran a 50/50 rollout via our split test platform. The test ran for 2 days on Android and 1 day on iOS before it hit significance. The results were positive. We saw a lift to % of D1 users who follow, D1 dap/dau, and D2 Retention in the test group.

 

Final Design

iOS

 

Retrospective

KEY METRICS

PHASE I

  • 9.4% increase of new users that follow on Sing Android.

  • 5.5% increase on DAP/DAU on Sing Android.

PHASE II

  • 24% increase of new users that follow on Sing Android.

  • 238% increase on Follows/Follower on Sing Android–reflect improvement in accounts surfaced.

  • 14% increase on D1 DAP/DAU on Sing Android–reflect improvement in social engagement.

  • Almost 1 in 3 new users follow.

PHASE III

  • Coming soon...

 

IMPROVEMENTS

  • Increase visibility to Find Friends.

  • Real-time notifications.

  • Recommend common connections.

  • Negative signals to improve recommendations.