Tally · Fintech · 2025
Rebuilding a banking app around one question
Tally’s 900,000 members kept asking support the same thing: how much can I actually spend? I led the redesign of the home screen and spending insights so the answer takes three seconds, not three screens.
- Client
- Tally
- Role
- Lead product designer
- Timeline
- 7 months
- Platform
- iOS and Android
- Team
- 1 PM, 6 engineers, 1 researcher
- User research
- Information architecture
- Interaction design
- UI design
- Usability testing
Results
- weekly active members
- +31%
- balance-related support tickets
- −44%
- App Store rating, up from 3.9
- 4.8
The problem
Tally’s home screen tried to show everything at once: the account balance, pending card payments, three savings pots and a feed of offers. Members opened the app four times a day on average and still didn’t trust what they saw. The balance included money already promised to rent and bills, so people overspent, got declined at the till and called support.
One in four support conversations was a version of the same question: how much can I actually spend?
Research
I ran 18 interviews and a two-week diary study with members living close to their limits, then sat with the support team to tag 600 recent tickets. A pattern came through quickly. People didn’t think in balances. They thought in what was left until payday.
“I don’t want to know what’s in my account. I want to know what’s mine.”
We turned that into a test we could repeat with every concept: how many seconds does it take to answer “can I afford this?” The old home screen averaged 11.
Design
The new home screen leads with one number, Safe to spend: the balance minus upcoming bills and savings goals until the next payday. A single sentence underneath explains the math, and the old balance stays one tap away, because trust also means letting people check.
- Upcoming bills sit in a timeline, each one tappable to pause, move or split.
- Spending insights moved out of their own tab and into the moments that need them: a nudge when a category runs hot, a short recap on Sunday evenings.
- Every state was designed, including the hard ones: payday moving, a bill bigger than the balance, and an account with no income detected yet.
I tested four concepts with 40 members in unmoderated sessions, then refined the winner with engineering, who built a forecasting model for recurring payments alongside the interface.
Outcome
We released to 10% of members first. The three-second test dropped from 11 seconds to 2.8, and the full launch followed six weeks later. Balance-related support tickets fell by 44%, weekly active members grew by 31%, and the App Store rating climbed from 3.9 to 4.8.
“Iris didn’t just redesign our home screen, she changed how the whole team talks about money. Every roadmap debate now starts with her three-second test.”