.png)
Back to the list
Field notes from two days in London with people who own the sizing problem.
Somewhere between our third conversation about PLM integrations, it became clear that this year's Fashion Tech Show Europe had a theme nobody had officially announced.
Everyone was talking about fit.
Not in the vague, aspirational way the industry talks about sustainability. In the specific, operational, "this is costing us real money" way. Pattern makers, digital transformation leads, and product development teams carrying cross-brand responsibility for sizing standards. People who've been in this industry for twenty years, sitting in a room together, asking the same question: why is this still so hard?
It's a fair question. When you break down a product's cost structure - sampling iterations, size availability gaps, inventory trapped in the return loop, overstock from inaccurate size curves, conversions lost to size insecurity, discounting driven by skewed ratios - sizing and fit-related costs account for roughly 34% of a typical product's price.
Hidden across multiple teams and budget lines, rarely owned by anyone in full, it's the kind of cost that never shows up as a single line on a P&L, which is exactly why it persists. For a brand earning $1 billion in annual revenue, that's a $100 million margin opportunity sitting untouched.
A year ago, every session at every fashion tech event was about AI. What it could do, what it might do, the demos, the possibilities, and the potential.
This year felt different. According to McKinsey's State of Fashion report, more than 35% of executives are deploying AI in specific functions - customer service, product discovery, copywriting, and image creation. The question has shifted from whether to how. Specifically: how do you embed intelligent systems into workflows that were built long before any of this existed?
One panelist put it plainly: "Last year everyone ran to AI, now they really need it, and the question is how to put it into the digital workflow."
Underneath that question was another one, asked quietly in nearly every conversation: what needs to be true before any of this actually works? The answer, almost every time, was the same.
The data has to be right first.
.png)
This is obvious when you say it out loud. The gap between knowing it and having solved it is where most brands currently live.
The problem isn't a shortage of data. It's that the data exists in silos that don't talk to each other. Returns signals sit in one system, product measurements in another, and customer body measurement data somewhere else entirely. Nobody has a complete picture, so decisions - on size curves, on grading, on inventory splits - still get made on instinct more often than anyone wants to admit.
We heard this reflected in almost every conversation on the show floor:
The same challenge is arriving via regulation. The EU's Ecodesign for Sustainable Products Regulation - the legal foundation for Digital Product Passports - entered into force in late 2024, with textile-specific requirements expected to roll out through 2026 and 2027. Brands are being asked to structure their product data in ways they never had to before. For many, it's forcing an overdue conversation about data infrastructure. Fit data, structured properly, belongs at the center of it.
This is the reframe that seemed to land most in the room.
Returns are the most visible symptom of a sizing problem, and the fashion industry does produce an estimated 92 million tonnes of textile waste each year, much of it linked to overproduction driven by inaccurate size curves and poor demand signals. However, returns are just the surface cost.
The less visible ones are where the real margin lives: missed sales because the right size wasn't available, capital tied up in returned inventory that can't be restocked at full price, customers who abandon their basket because they're not confident enough to commit, and discounting triggered by getting the size ratios wrong at the buy stage. These costs don't show up in one line on a P&L; they're distributed across teams, buried in different budget lines, and rarely connected back to a root cause.
When you add them up, the picture changes. What looks like a logistics problem or a product issue is actually a data problem, and data problems have data solutions.
Something else was consistent across two days of conversation. The organizations making the most progress weren't the ones with the most sophisticated technology, they were the ones who had decided to start.
"Brands need to see what they need to evolve and they have to start and try out solutions to then have workflows and developments to improve." Simple, but it cuts through the analysis paralysis that still holds a lot of companies back.
The brands who stood out had already done the first hard thing: they'd admitted that gut feel wasn't enough, and they'd started building the foundation to do better.
The Fashion Tech Show Europe 2026 confirmed something important: the industry has moved past the debate about whether fit intelligence matters. The people doing the actual work - the pattern makers, the garment technologists, the digital leads - already know it does.
The conversation now is about infrastructure, about building a clean data foundation that connects what's already there, and making decisions that can be shared, questioned, and improved across the whole organization rather than made by one person in one room.
That's the hard work, and it's the interesting work.
If you're in the middle of it and want to compare notes, we'd love to talk.
SAIZ hosted a Think Tank at Fashion Tech Show 2026 exploring the decision infrastructure behind autonomous fashion intelligence - Sizing & Fit edition.
