Kumo checkout
A coffee subscription lost a third of its buyers at the address step. Three fields went, one page went, and the order button stopped moving.
Result
+18%
completed orders in eight weeks
- Client
- Kumo, a coffee subscription
- Role
- Lead designer
- Year
- 2025
- Tools
- Figma, Maze, A/B tests
The problem
Kumo sells coffee by subscription. A third of the people who chose a coffee never finished the order, and most of them left on the same page: the address form, asked for before the price was clear.
What I did
I watched forty recordings of real checkouts and tested three versions with twelve customers a week. Company name, the second address line and confirm email went; the address moved after the price; the order button stopped jumping when the postcode was looked up.
What changed
Completed orders rose by 18 percent in eight weeks, and support emails about wrong addresses fell by half. The new form became the pattern for every form in the shop.