At a glance
- App category
- Journaling
- Industry
- WellTech
- Services
- Growth engineering Conversion rate optimization
- Channels
- iOS and Android apps, backend
- Period
- 2024 -
- Markets
- Worldwide, United States primarily
- Key results
- +131% LTV. Crash-free users on Android from 82.5% to 98.5% within the first 90 days. The app rewritten from UIKit to SwiftUI, with new activation and monetisation flows adding roughly $174,000 in revenue
The challenge
Five Minute Journal had over 800,000 downloads and a growth problem that was not a growth problem. The development approach could not support the pace the product needed, and customer support was not keeping up. Both show up in the same place: in revenue, with no obvious cause attached.
An app this size does not stall because of one bad decision. It stalls because crashes, slow releases and unanswered tickets each take a little, continuously, and none of them look like the reason.
What we did
1. Stood up three development tracks at once. iOS, Android and backend, staffed quickly, because a product waiting for a team is losing the same revenue every week it waits.
2. Fixed the bugs that were costing money first, then everything else. The crash and defect work was ordered by revenue impact rather than by severity labels.
Reasoning and learning: in a subscription app a crash is not a quality problem, it is a churn event. The fix list should be sorted by what it is costing, not by how bad it looks.
3. Rebuilt the roadmap around a growth target. The updated roadmap was written to support at least 60% year-on-year revenue growth, so engineering work was selected against a number rather than against a wish list.
4. Connected product, marketing and development. The three had been working in sequence. Putting them on the same cadence is what made release speed a growth lever rather than a constraint.
5. Integrated a design system. It saves build time on every screen after it, and it is what makes A/B testing cheap enough to do often.
Reasoning and learning: a design system is a testing tool. The cost of a variant is what decides how many variants a team runs.
6. Rewrote the app from UIKit to SwiftUI. Done alongside the growth work, not instead of it.
7. Built new activation and monetisation flows. The onboarding and paywall structure were rebuilt once the platform could carry changes quickly.
Results
LTV rose 131%. Crash-free users on Android went from 82.5% to 98.5% within the first 90 days, with similar movement on iOS. The design system shortened development and testing cycles, and the rebuilt activation and monetisation flows added roughly $174,000 in revenue.
131%
LTV increase
Key learnings
Stability is a revenue metric. Sixteen points of crash-free rate on Android is not an engineering statistic, it is users who stopped losing their entries, and it sits underneath the 131% LTV gain rather than beside it.
Sort the backlog by what it costs, not by how it is labelled. The expensive bugs and the severe bugs are rarely the same list.
A design system pays for itself in tests, not in screens. It lowers the price of a variant, and the price of a variant decides how much a team learns.
Conclusion
Five Minute Journal was not short of users. It was short of a platform that could ship. Three development tracks, a revenue-ordered fix list, a design system and a move to SwiftUI turned the app into something that could be improved continuously, and the rebuilt onboarding and paywall flows turned that capability into revenue.
This is a separate project from our onboarding redesign for the same app.
More on our conversion rate optimization and health and fitness app marketing.





