Gelato print-on-demand platform redesign
Gelato is a global B2B on-demand print platform that lets businesses order branded materials printed locally at their destination, cutting delivery costs and carbon emissions.
I joined as the second product designer, working with a London-based PM and a remote development team in Ukraine.
The goal was to make the checkout fast and reliable enough that customers stopped needing workarounds, and to modernise the overall look and feel in order to make the plaftorm more marketable.
Company type
B2B SaaS, Scale-up
Time frame
6 months
Contribution
Research, Service Design, UI Design, Design System
THE PROBLEM
Gelato's checkout had grown organically and it showed. Most customers had built their own workarounds just to get through a standard order. The brief was to fix the checkout, but it quickly became clear the whole platform UI needed to move at the same time.
DISCOVERY
Before any interviews or flows, I sent a System Usability Scale survey to a sample of customers.
The score came back at 25/100. That's not just "needs improvement". It's a platform that most users find genuinely difficult to use. It also gave us an objective baseline to measure against when we shipped.
USER INTERVIEWS AND FINDINGS
Speaking with users directly through moderated user testing sessions was key to uncover their main frustrations with the app and to complement the quant data we had. I also spoke with Account managers, Customer Service and Marketing to find other common themes.
Separated flows for single and multi location orders create confusion
"It's really hard to do multi-location orders. I keep getting confused how to do it, but I figure it out eventually."
Jean Baptiste · Delachaux
The shipping and quantities are split up but should be on the same step
"It's odd that I need to select quantities in the cart. I want to see the final price at the end not before I even start the order"
Andreas · Rommelag
Missing quality-of-life features users expected as standard
"I would love to book ahead so I can do the ordering all at once without rushing at the last minute."
Keren · Kenes
PRIORITISATION
After synthesising the interviews I documented every usability issue in Airtable and scored each one across three dimensions. The severity score (criticality × impact × frequency) gave us an objective ordering. The highest-severity issues went first, regardless of which ones were easiest to design for.
DIMENSION 1
Task criticality
How important is this task to the user's core job? Scored 1 (low) to 5 (critical). Multi-location ordering scored 5, the primary use case for most customers.
DIMENSION 2
Impact on user
How much does this issue affect the user's ability to complete the task? 1 (suggestion) to 5 (full blocker). Navigation issues scored 4–5 across the board.
DIMENSION 3
Frequency
What percentage of participants encountered this issue? Issues affecting 70%+ of users were treated as must-solve for the MVP, regardless of their other scores.
DEFINITION
Desired outcomes set
before any design started
Besides the SUS baseline, I signed off the core success metrics with Product and Engineering before we started looking for a solution.
Success metrics and UX goals
Order volume increased by at least 5%
Decrease TTC by at least 67%
SUS Score above 60
Implement new features such as scheduled delivery, smart defaults for recipient addresses and delivery notes
Improve the overal look and feel of the platform by updating the UI style
COLLABORATION
A week in Tallinn with the engineering team.
The checkout complexity (multiple files, multiple addresses, variable parcel weights, international logistics) meant design couldn't happen in isolation. I flew to Tallinn with part of the London team to spend a week with the Ukraine-based developers working through solutions in person.
CONSTRAINTS IDENTIFIED
The sessions surfaced constraints that would have taken weeks to discover remotely: how the platform batched shipments, what was possible with delivery window APIs, how billing entities worked across jurisdictions.
SOLUTION ALIGNED
The result was solution alignment that let us move from wireframes to build-ready specs in weeks rather than months, with far fewer back-and-forth cycles once development started.
PROTOTYPING & TESTING
The solution was tested twice: once before development started, using a design Prototype, and again with the live product using the SUS survey as the measurement instrument.
SOLUTIONS
Delivering the new solution and styling came with frequent syncs with developers and a close collaboration with one of the front end developers who helped me find the right technical approach to the new styling implementation.
OUTCOMES
The redesign didn't just fix the checkout. It created a scalable platform foundation that made every future iteration faster to ship.
DESIGN SYSTEM FOUNDATIONS
A lightweight UI kit and built alongside the product
I built components as they were needed for the redesign, defining states, interactions, and styling patterns iteratively with the front-end developers. The result was a practical, immediately usable kit which served as a starting point for a full design system
WHAT I LEARNED
