Had a really great conversation yesterday about how to think about retention problems in consumer apps. My experience is primarily with consumer, mobile-first, media & content centric products (music, reading, social). This post is a way to capture some learnings - many of these can be tied back to step function increases .

Cultural Learnings

  • Retention isn’t about any one thing, it’s about the compounding effects of many small things. To make this possible, any team has to be committed to the exercise of continual experimentation and learning. At Wattpad, our data team pioneered the experiment council - a weekly review of the experiment pipeline, which is still in place today and enforced this discipline. This is a cultural exercise, not a technical one.
  • Retention is a growth exercise, which is new, so talk to the best in the game. There are no books on this (yet), but there are very active communities that share best practices. Reforge is a great example (I’ve done an early Reforge course and still refer back to it often).

Product Specific Learnings

  • You need to understand your existing retention patterns and base rates. At Wattpad, we knew we were a multi-day-a-week product, and for every session people spent about 30 minutes. This helped us baseline our experiments and be specific with our goals.
  • The world is busy. You need to understand the “a-ha” moment of your product, and deliver to people as quickly and as often as possible. This is the moment your product delivers exactly what your users are looking for (and in the best cases, didn’t know they were looking for).
  • If your product enables any kind of user interaction - comments, liking etc - keep that interaction as “close” to content as possible. The closer it is, the more likely users will engage. In media products, this manifests as in-story comments (Wattpad), in-track comments (Soundcloud). This also happens to create a unique experience that can’t be replicated by simply moving the content from one place to another.
  • Discover is nonlinear and needs to be optimized across channels. People find your product in many ways, don’t assume content discover is going to happen from your “discover” screens. It rarely does.
  • Creating rabbit holes of discovery. Youtube’s biggest source of discovery is the “related videos” next to each video. They have spent years optimizing that list to encourage you to continue watching. Today those “related videos” aren’t just related to the video you’re watching, they are personal recommendations, editorials, ads, etc. A personal note - This method works but as an industry we have to go beyond capturing attention just for the sake of doing it.
  • Don’t waste the moments in which users are most likely to be happiest - e.g. right after they’ve watched a video, or given a positive rating to some content. That is a potential opportunity to encourage another action (potentially a conversion to being a paid user), or even a simple positive app review.

These are only some learnings and I’ll update this as I think of more.