Loan conversion: from 2.5% to 9.6%
conversion from application to loan
signed loans per month
was the target — we beat it
The task. 1.25 million applications a month, and only 2.5% made it to signing.
The market was crowded, and the bank had to fight for every customer. Many people downloaded Jusan specifically for a loan. Leadership set a target of 7%.
The numbers. I mapped the drop-off screen by screen, and it became clear where we were losing people.
I asked the analysts to pull data for the whole process and put the number of people who left next to each screen.
(10.7%)
(14.4%)
(<0.1%)
(7.4%)
(4.0%)
(1.8%)
(5.5%)
(<0.1%)
(51.5%)
(2.1%)
(<0.1%)
(<0.1%)
(<0.1%)
31,654 (2.5%)
How many applications were lost on each screen of the old flow. More than half were lost on the “Rejected” screen: the problem wasn’t the interface.
First, we cut the extras: we stopped asking for what we already knew. The flow got easier, but conversion barely moved — 2.4–2.5%.
The key finding. We were simply rejecting more than half of all applications.
Before the next meeting, I asked the team to find out why. There were two reasons:
- Risk rejections: people with overdue payments. The bank couldn’t approve them.
- High debt burden. By law, payments on all loans can’t exceed 50% of income, and a new loan simply didn’t fit within the limit.
The idea. If someone has too many loans, offer to combine them right in the application.
The monthly payment goes down, the limit frees up, and the person gets a new loan. The PO liked the idea, and engineering confirmed it was doable.
A clear screen. None of the 8 people understood the first version. After the changes, all 6 did.
- Split the screen in two. The first explains that refinancing is needed. The second shows the terms.
- Explained the word. Even bank employees didn’t know what “refinancing” meant. I made the word tappable, with an explanation.
- Changed the button text to “Show terms”: the person isn’t signing anything, just looking.
Result. Loan conversion grew from 2.5% to 9.6%.
- 120K signed loans per month instead of 31.7K — almost 4× more. We beat the 7% target.
- Refinancing drove almost all of the growth: 80K people accepted it.
- The bank won over some loans from other banks thanks to refinancing.
What I learned. The strongest decision here wasn’t a screen — it was the question “why are we rejecting people?”
Simplifying the screens did almost nothing — the growth came from a place design doesn’t usually look. And a single word on a button decides whether someone taps or leaves.