How I unified the way people pay on Payoneer, on web and native mobile.
The new flow on desktop and phone, side by sidepayoneer-hero.png
At a glance
Role
Senior Product Designer
Timeline
2023 to 2026 (with one long pause)
Platforms
Web, iOS and Android
Status
Ships August 2026
2023Mapping, research, wireframes
PausedEngineering moved to a company migration (I worked on Withdraw to Bank meanwhile)
2025Back, with a design system and Maze tests
Sep 2025Handed off web (full walkthrough, open edge cases listed), moved to mobile
May 2026Every mobile flow done
Aug 2026Ships
01The problem
Two products. They never talked.
The 2 old flows, side by side
Wall of fields
Names cut off
Long, technical names
problem-old-flows.png
One paid into a Payoneer balance (instant, if you had their email). The other paid into a bank account (days to verify, then days to send).
People would type in someone's bank details, find out they had a Payoneer account too, and start over. And every new feature got built twice, once per product.
02Research
Flow mapping: internal first, external second
Wise flow mapresearch-wise.png
Revolut flow mapresearch-revolut.png
PayPal flow mapresearch-paypal.png
Mercury flow mapresearch-mercury.png
Airwallex flow mapresearch-airwallex.png
I mapped our 3 existing flows first. Then the same flows at Wise, Revolut, PayPal, Mercury and Airwallex. Every payment came down to 4 questions:
01Who are you paying?
02Where's the money coming from?
03What are the details?
04How do you confirm it?
03Finding the shape
Three structures, a mashed-up winner
V1 wireframes, 5 stepsshape-v1.png
V15-step flow, scheduling lived inside the amount step.
V2 wireframes, 5 stepsshape-v2.png
V25-step flow, scheduling got its own step.
V3 wireframes, 4 stepsshape-v3.png
V34-step flow, recipient and transfer type together, behind a mega dropdown.
No clear winner, but we made a combo. The big plan (3 products and batch payments, all in one place) didn't fly. Research kept withdrawals separate, batch got saved for later, and single payments became this project.
04The real scope
It looked like two flows. It was closer to twenty.
Recipient
Saved
New
Not on Payoneer yet
Currency
Same
An exchange
Timing
Now
Scheduled
Recurring
Paid from
Balance
Card
Bank
Company budget
Paid into
One method
Several
Roles
One person
A contributor plus an approver
I mapped every combination, for first-time and returning users, plus errors and payment tracking.
05Round two
Back in 2025. New design system, new rules.
Contact picker, beforepicker-before.png
Before
Contact picker, afterpicker-after.png
After
Solid components, but anything new needed the design system team's OK. Testing a prototype in Maze showed people weren't sure which contact to pick, or where the money would land. I made changes, but we never got to test them again.
06Mobile
Same flow. Smaller screen.
Old way to select, 1select-old-1.png
Old
Old way to select, 2select-old-2.png
Old
Old way to select, 3select-old-3.png
Old
The new way to selectselect-new.png
New
The flow in phone framesmobile-flow.png
In September 2025 I handed the web flow to another designer and joined the mobile team, right as this flow was heading there.
The app's design system was broken. No autolayout, a dark mode that didn't match light mode, 3 ways to select the same thing. A UX consultant agreed we should start fresh, management said yes, so I rebuilt it. This flow was one of the first things built on it.
07Where it stands
Two products. Two mental models. Now one.
The finished flow, end to endfinished-flow.png
Ships August 2026. Metrics stay internal, so no numbers here.