Pou Is Here
FF_full phone@2x.png

Find Friends

 

STRATEGY • UX • IxD • VISUAL DESIGN


Find Friends

The Find Friends feature was designed to surface interesting accounts for new users to follow. However, in Q1 of 2018 only 1 in 5 users ‘followed’ on the first day. We know that when new users follow, they are:

  • More engaged

  • Retain at a higher rate

  • More socially active

This low engagement of the Find Friends feature was a serious cause for concern.

 
 

TEAM

  • Product Design Pou Dimitrijevich

  • Product Manager Winnie Lee

 

INITIAL DESIGN

 
 
 

RESEARCH | COMPETITIVE ANALYSIS

 

We looked at other music and social network apps for inspiration and learnings.

f_logo_RGB-Hex-Blue_512.png
Twitter_Logo_Blue.jpg
Spotify_Icon_RGB_Green.png
download.jpg
 
 
 

RESEARCH | QUANTITATIVE ANALYSIS

Then we looked at the data and formed a hypothesis based on our learnings.

  • 17% of D1 users visited the page = Low visibility of the feature

  • 11% of users follow anyone from the page = No clear value proposition, or incentive 

  • Recommendations are irrelevant, and difficult to evaluate

 

 

OUR HYPOTHESIS

By improving discoverability, quality of accounts, UX/UI, and the recommendation logic of the Find Friends feature, we can drive more users to follow.

 

DISCOVERABILITY

Part I

Provide an incentive to follow by deprecating the Feed if a user is not following anyone; replace it with a CTA to ‘Find Friends’.

Android_FF_2.png

Part II

Increase visibility by surfacing ‘Find Friends’ recommendations outside just the Feed.

Android_FF_3.png
 
  • QUALITY

    • Remove ‘New on the scene’

      New users most likely don’t have content, and wouldn’t be engaging for other new users to interact with.

    • Remove inactive and spam accounts

      If a user hasn’t uploaded an profile photo, and has been on the app more than a week, data show they are inactive/spam accounts

    • Get specific

      ‘Singer you may like’ are users who share a common musical interest with you. However the copy is vague, and doesn’t encourage user to follow.  

  • RECOMMENDATION LOGIC

    • Improve back-end logic

      Recommend users interest: by region, social activity, and quality of accounts. 

    • Improve strings

      Clearly communicate recommendations

      • You both sang <song title>

      • Also likes <musical interest>

      • Popular in <region>

      • Hot Influencer!

      • Power user

      • Follows you

      • Friends on Facebook

      • Followed by <username>

  • PERSONALIZATION

    • Allow users to speak for themselves by including their profile blurb. 

Current Find Friends.png
 

USERFLOW

 
ff_userflow.png
 
 

FINAL DESIGN

3.0 iOS_feed_recommended.png
1.0 iOS_FF_recommened.png
2.0 iOS_feed_recommended.png
1.0 iOS_FF_recommened.png
 

METRICS

 

Increase in new users that follow

  • 9.4% increase of new users that follow

  • Almost 1 in 3 new users follow.

 

Increased engagement on the first day and overall

  • 5.5% increase on DAP/DAU.

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

 

Increase in quality of accounts followed

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

 

IMPROVEMENTS

 
  • Increase visibility to Find Friends.

  • Real-time notifications.

  • Negative signals to improve recommendations.