We built the API we wished existed
when we were the retailer.
PerfectFit is a B2B size recommendation API that helps fashion retailers cut size-related returns. No proprietary sizing standards required — it works with your existing size guides.
The problem is bigger than it looks
The EU fashion industry ships around 5 billion parcels per year. Estimates put return rates at 25–40% for online orders — and size fit accounts for a third to a half of those returns. That's not just a logistics cost: it's a carbon footprint, a customer experience failure, and a profitability gap that compounds every quarter.
Existing solutions are either built for giants (multi-month integrations, custom SDKs, enterprise contracts) or inaccurate (generic "pick the nearest size" logic). Neither works for a mid-size retailer who wants results in weeks, not quarters.
PerfectFit bridges that gap: accuracy at the level of custom solutions, simplicity at the level of a JavaScript snippet.
Team
Pharm.D. + EM Lyon MBA. 8 years in pharmaceutical brand management (specialty care, product launches, market access). Transitioned to fashion tech after identifying the return-rate problem on a consulting project for a European fashion brand.
Building PerfectFit as a solo founder — fast feedback loop, no committees, direct line to every pilot partner.
We are not hiring yet — but if you're an ML engineer or fashion industry expert interested in joining when we close our first customers, reach out.
Timeline
While working on return data for a fashion brand, Arnaud saw how large a share of returns came down to the wrong size — and that existing tools were either too complex or too generic.
Built a category-aware size engine working across brands, and validated it internally on a Celio-style catalog.
useperfectfit.com goes live. B2B API + embeddable widget + tenant dashboard. First pilot conversations with fashion retailers in France.
Looking for 3–5 fashion retailers willing to run a 3-month pilot on one product category. Free during pilot. We share return rate data, you provide feedback.
Values
A wrong recommendation costs more than a slow one. Every algorithm decision is made with false-positive rate as the north-star metric.
We explain every recommendation with a confidence score and contributing factors. No black box — retailers deserve to understand the logic.
Our widget deploys in a single script tag. Our API needs one POST call. Complexity belongs in our code, not in yours.
We integrate directly with your data team, align on KPIs, and share the upside of reduced returns. We win when you win.
Ready to cut your return rate?
We offer a free 3-month pilot for qualified retailers. No contract, no credit card — just a shared commitment to measuring the impact.