Goal
Design a mobile bank that people can actually use without thinking about it. The brief asked for something modern and calm, close to Revolut in spirit: only what you need, nothing extra.
Problem
The interface tried to do everything on one screen. Every function sat at the same level, and the simple things were buried under menus. A transfer is what people do most, but it took too many steps to get there.
1
People leave after something breaks
Nobody switches banks because another app looks nicer. They switch after something goes wrong. One person moved banks after his accounts got frozen the moment his mobile internet dropped, with no warning and no way in.
2
Some users don't want to figure anything out
Not everyone wants to learn a banking app. In the interviews people said “I don't get into it” and “I don't really understand this stuff”. For them the app has to work on the first try, or they're gone.
3
Transfers are the whole point
Every person named transfers as the thing they use the app for most. And slow transfers were the thing they complained about most. If transfers are heavy, the app is heavy.
4
Design and comfort
People notice design and comfort, and it's a real reason to stay. But it never comes first. If the basics don't work, no amount of polish saves it.
Research
Benchmarking
I went through direct competitors screen by screen: T-Bank, Alfa-Bank, Raiffeisen, and Revolut as the reference from the brief. I looked at sign-up, the main screen, transfers, and how each app groups its functions.
- Sign-up is long everywhere, four to six steps before you reach the app
- Most apps put every function on one screen with no grouping
- Revolut stays minimal, but some screens feel unfinished
- Older banking apps have the features but look and move like 2015
Interviews
I ran 3 in-depth interviews with people who use a mobile bank every day. Different ages, different countries, different comfort with tech. All of them had changed banks in the last few years.
What i understand
- Send money fast, without extra steps
- See where the money goes
- Get help without waiting or explaining a lot
Personas
I pulled it into 3 personas. Two come straight from the interviews. The third is built on the brief's core audience, the people the app is really for.
You can view the entire research
Process
1
Brief and spec
Read the brief and wrote the tech spec, the full list of screens and flows
2
Analysis
Went through competitors screen by screen and mapped every function the app would need
3
Structure
Built the flows and navigation before the screens, starting from the split between doing something and looking something up
4
Screens
Designed the app on top of that structure, balance first, actions second, the rest one tap away
5
Design system
variables, tokens, components, and all states, so every screen stays consistent
6
Testing
Ran A/B and 5-second tests, then reworked the screens that lost
Hypotheses
[1] Send to several people at once
Hypothesis
If you can pick several contacts and send in one action, then splitting a bill stops being a chore.
Solution
Choose several contacts, split the amount or set it per person, confirm once
Impact
Cuts out the transfer people repeat most, and makes splitting costs a normal part of the app.
[2] Quick questions in support
Hypothesis
If the support chat opens with the most common questions ready to tap, then most people get an answer without writing anything.
Solution
The chat starts with preset questions that cover what people ask most, each one leading straight to an answer.
Impact
Faster help for the user, less load on real support for the questions that repeat.
[3] Cashback as a shared pot
Hypothesis
If several people put their cashback into one shared goal, then it turns from small change into something worth watching.
Solution
A shared goal,a trip, a device, paying off a loan early, that everyone's cashback feeds into, with progress everyone can see.
Impact
Gives cashback a point, and a reason to come back and check it.
Wanted to build
[1] Anti-subscriptions
The bank finds subscriptions you barely use and offers to cancel them in one tap.
[2] Future balance
Shows what's left in 7, 14, and 30 days, counting regular payments, subscriptions, and usual spending.
Implementation
1
Registration
- Kept the flow to the steps people already know: phone number, sms code, a four-digit passcode
- Added Face ID from the second launch, so the user gets in with one look instead of typing a code every time
- Drew the error states into the flow, wrong code and code not received, so sign-up never turns into a dead end
2
Main screen
- Put a map of nearby ATMs right on the main screen, so when the user needs cash they don't have to search for it
- Added a spending block with recent transactions, so the last movements are visible without opening history
- Put the balance and the card at the top, the first thing people open the app to check
3
Account screen
- Split the screen into two tabs, Operations and Information, because the A/B test showed one flat screen with every function was too much at once
- Kept the actions in Operations and everything you look up in Information, so doing something and checking something are two different modes
- Drew the error states into the flow, wrong code and code not received, so sign-up never turns into a dead end
4
Transfers
- Added every way people actually send money: by phone number, card, QR, or between their own accounts
- Kept the flow short, pick a contact, type the amount, confirm, so the most common action takes the fewest taps
- Put contacts and templates on the transfer screen, so a repeated transfer doesn't start from scratch
A/B testing
I tested screens against each other and ran 5-second tests to see if people find the right button fast.
Test 1 — Account screen
A: everything on one screen — balance, cards, send, deposit, receive, activity, cashback, details, settings, statement.
B: the same functions split into two tabs — Operations and Information.
Test 2 — Balance on the main screen
A: total account balance only.
B: total balance plus the balance on each card.
Result
Echo went from a brief to a full product: research, structure, flows, screens, and a design system. The main calls were checked against real people, not taste.
Two screens changed because of testing. The account screen dropped half its noise once the functions moved into tabs. The main screen got per-card balance, because that's what people looked for first.
Learnings
Testing changed the design more than any opinion did. The account screen I liked at first was the one that lost, and the data made, not me