App Masters · Dec 21Do This If Your App Retention Sucks
Lifted day 7 retention from under 10% to over 40% in a couple of months through many small experiments guided by a public manifesto. Messaging like "read later, not never, set your reading goals" got a much higher response than a mission pitch about being happier and smarter, and a freshness test only showed results at around 600 users.
Tips
- Run A/A tests first to size how long and big experiments must be▶ 4:08Listory used A/A tests to plan experiment length and traffic before running their retention experiments.
- Plan for roughly 600 users per experiment before reading results▶ 5:24In a content freshness experiment (stories trending in newsletters last week vs last month), results only appeared at around 600 users.
- Test messaging high in the funnel, such as store listings, to learn faster▶ 9:28
- Lead with concrete, actionable value instead of your mission statement▶ 10:42Utilitarian messaging about the real 'read later' beat aspirational mission messaging for Listory.
- Align acquisition messaging with the feature you are testing▶ 13:33
- Write a manifesto of what you believe and do not believe▶ 6:44Listory credits the manifesto with helping the team prioritize features and drive actions.
- Publish your manifesto publicly to build trust and commit the team▶ 8:01
- Expect retention gains from many small experiments, not one big win▶ 8:30Retention rose from under 10% to over 40% over a couple of months of continuous experimentation.
- Find the funnel bottleneck and re-check messaging after each change▶ 14:10
- Limit content to a finite, curated set so people consume more▶ 1:41
- Let data decide, even on a highly opinionated team▶ 3:09Experimentation was the main strategy behind the retention lift.
- Aim to lift early retention from under 10% to over 40%▶ 2:18Retention went from less than 10% to more than 40% after the experimentation program.
- Check that you measure second day retention in the right place▶ 14:43
