the problem
Acquisition was working. Retention was a cliff, before the wallet, 63% of first-time buyers placed a second order, but only 36% made it to a fourth. More than half of all buyers placed exactly one order and never came back.
The instinct: run a discount campaign. But a discount trains users to wait for the next one. It doesn't build a habit, it rents one. We needed something that changed why users opened the app, not just the price when they did.
The reframe: Don't reduce the cost of one order. Make not ordering feel like leaving money on the table.


the habit loop
A one-shot promotional credit would show an immediate retention boost that decays by week 6โ8. What the data actually shows is the opposite: wallet users retain slightly worse than non-wallet users in weeks 1โ2, then cross over at week 4โ6 and stay elevated through week 10.
You can't fake that crossover shape with a discount. It only happens when users are returning because the mechanic keeps pulling them, not because they received a free thing once.
product decisions
Showcase
Wallet Flow
impact
TAKEAWAYS & LEARNINGS




















