Demo marketplace that made Mirakl implementations faster and cleaner.
UX/UI Design UX Research E-commerce
My role
UX Designer / UX Researcher / UI Designer
Date
January 2021 → November 2022
Project Type
Marketplace prototype + Documentation + Workshops
Tools
Figma / Notion / Confluence / Mirakl Platform / Baymard institute
In short
Mirakl's clients kept hitting the same wall when migrating from a retail site to a marketplace: nobody had a clear picture of what changed on the front end. As the only designer in the Solution Consulting team, I built a responsive marketplace prototype plus documentation and a repeatable workshop, used internally and in front of clients to train teams and de-risk launches.
Context
Mirakl's platform turns an existing retail e-commerce site into a marketplace: it plugs into the client's CMS, adds the seller and operational layers, and lets third-party vendors sell alongside the retailer's own catalog. My team's job was to get clients from "we bought the software" to "the marketplace is live and working."
The problem
The hard part of a marketplace migration isn't the back end. It's everything the design and dev teams suddenly have to rethink on the front end: seller information, multiple offers per product, split baskets, marketplace-specific checkout and account flows. Dozens of small UI changes that ripple across desktop and mobile.
Our consultants explained these changes in meetings, but there was nothing to point to. No prototype, no documentation, no repeatable workshop. Every implementation started more or less from scratch, which meant slower rollouts, avoidable mistakes, and teams who only understood a feature once they'd already built it wrong once.
What I built
A fully responsive marketplace prototype covering the platform's main features and their underlying APIs, paired with documentation and a workshop format that walked teams through the front-end impact page by page.
Because the platform kept shipping updates, I couldn't treat this as a one-shot deliverable. I ran it as a rolling, year-long series of short sprints, each focused on whichever section was causing clients the most trouble at that moment, so my team had something usable to hand over early instead of waiting a year for a finished kit.
How I worked
I was the only designer on the project, so research, wireframes, prototypes, documentation and the workshops all ran through me. Rather than walk through a textbook process, here are the decisions that actually shaped the result.
I talked to both sides before designing anything.
Ten interviews, five from my team and five from client teams, plus shadowing live customer workshops to see what consultants had to repeat three times. That gap between what was said and what landed told me the kit had to be visual and clickable, not a slide deck.
I designed for two audiences, not one.
Affinity mapping surfaced two clearly different users: our internal consultants and the external client teams. I wrote a persona for each, which is why the deliverable works both as a live demo tool and as a leave-behind reference.
I grounded the patterns in evidence.
I benchmarked competing marketplaces and cross-referenced e-commerce UX conventions against Baymard Institute research, so the prototype reflected proven patterns instead of my personal taste.
I treated complexity as the core design problem.
Making multi-offer products and split fulfilment read as simple on screen was the recurring challenge. I built on a curated Figma e-commerce library and extended it as I went, which also kept the whole kit visually consistent.
I kept it alive.
Internal usability rounds and post-workshop client surveys fed continuous changes, and I tracked each Mirakl release so the kit never demoed a feature the software no longer had.
Outcomes
The prototype, documentation and workshop were used by both the consulting team and client teams during real implementations. Internally, consultants ran smoother rollouts because they walked into client conversations with a shared reference instead of improvising. Client teams left with a clearer grasp of the platform's front-end implications, as the post-workshop surveys consistently showed.
Takeaway
The honest challenge was doing this as a Master's student, on site only three weeks a month. My team backed me and gave me room, even though most of them had never worked with user-centered design methods. Part of the job was quietly demonstrating why the research mattered while they taught me the e-commerce and marketplace world from the inside.
It was also my first time working entirely in English, sometimes with client teams on the other side of the planet. The biggest lesson: research is the part you can't skip. In my first sprint I felt paralyzed without every technical and business requirement in hand, and learning to move forward anyway, filling the gaps as I went, turned out to be the actual skill.
