Pratyusha’s approach
Lets define user engagement for spotify. Is it the total time spent per user on app per day? no. of songs listened per user per day? or anything else?
Assuming its total time spent per user on app per day. i.e (total time spent by all active users) / ((total active users)* (no. of days taken into consideration))
To increase this metric, we need to increase daily sum of session times per user. lets evaluate which of the potential features increase this metric the most
-
Daily mixes 2.0 with new algorithms - if we are improving the model behind Daily mixes that allows users to find more interesting songs -> higher satisfaction with relevant songs = higher continuity in listening on spotify = higher return rate and higher session times . This seems to be a strong feature for the goal
-
Podcast Transcription & Search - helps podcast listeners read the podcast when words are unclear and find relevant content. We should check what % of daily active users are podcast listeners and estimate what % of them could benefit from transcriptions on podcasts. This estimate can be done by the number of times a user rewinded 10 seconds back in podcast episodes. i believe the % of daily active podcast listeners would be a much smaller fraction of total users as compared to total DAU. above this, the % of users needing help with transcription and leaving the app due to lack of it would be much lesser. so I would deprioritise this feature as compare the #1
-
Artist LiveStream integration - allows artists to do publicity for their original songs. the target users would be the artist's existing fans & followers. this would help gain traction for specific songs- which can increase total user time for the artist's fans but not necessarily for all users.
Given the above reasoning, I would prioritise #1 strongly. If we need an order. 1 >3 >2