Improving financial app retention through better transaction entry
PO, PM, Tech Lead, Developers, QA
Context
Kimi is a financial app for microbusinesses and solo entrepreneurs, helping them track income, expenses, and transfers to monitor their financial health.
Problem
When I joined the project, retention was low. I led usability reviews and user interviews, uncovering that the operation entry flow was confusing:
- the outdated Short vs. Long mode puzzled users,
- transfers worked in only one mode,
- the dated, inconsistent interface led to hesitation, errors, and drop-offs.
Challenges
There was a some version of design system, legacy code, and a few corner-cases like recurring payments floating in scope.
Here’s what made this tricky:
- I needed to simplify without oversimplifying — users needed speed and flexibility.
- I had to make Short vs. Long modes immediately understandable.
- The visual structure lacked clarity, everything was inconsistent.
Results
We couldn’t fix retention overnight, but we could remove daily friction by making transaction entry fast, clear, and trustable. So we rebuilt the flow into two distinct modes, cleaned up the layout, standardized interactions, and instrumented key events to track real-world use — all while untangling product and design debt alongside the dev team.
We saw early signs of improvement even before full rollout. In test sessions users understood when to use Short vs. Long and told us it just “felt faster and easier.” Also, we set up event-based tracking for key actions to measure task time and drop-offs after release.
Impact & Learnings
The biggest takeaway for me was that small, focused interaction fixes around core daily actions deliver disproportionate business value. It wasn’t onboarding or shiny reports that moved retention — it was making the app effortless for the one thing people needed every day.
And it reminded me that product risk hides in the unglamorous basics; if logging transactions doesn’t work, nothing else matters. Would do it the same way again. Probably faster now.
