Scaling the Right Sales: Growth That Doesn't Come Back

A sale that comes back three days later, half-worn and repackaged, was never really a sale. It was a loan with a carbon cost. And for the woman at the other end of it, it wasn't a good experience either. She spent money she wanted to spend well, waited for a parcel she hoped would work, and ended up back at the post office with a bag she has to queue to send away.

That moment, repeated across millions of orders, is the quiet problem underneath online fashion. The industry talks about growth in gross terms, but a growing share of that growth is provisional. It has to be scaled back before it counts. The more useful question, for the shopper and the brand alike, is how to grow the sales that stay.

The sale that stays starts with her

The women we work with don't want to return things. Nobody enjoys it. When a dress goes back, it isn't a verdict on her taste. It's usually a sign that she was asked to guess, and the guess didn't hold up.

Online clothes shopping has a return problem the industry has never really solved. Research collected in our white paper on fashion returns puts the average online apparel return rate somewhere between 20 and 40 percent, and higher than 50 percent in some segments. Compare that with physical retail, where returns sit closer to 8 to 10 percent. Roughly four times as much clothing goes back when it's bought online, and in the worst-affected European markets nearly half of every order comes home only to leave again.

The woman living inside those numbers isn't careless. She's navigating a system that gives her almost nothing to work with. A size label that means one thing at one brand and something else at the next. A photo of a model whose body she can only loosely relate to her own. A returns policy generous enough to make over-ordering feel rational. Of course she buys three sizes. The system taught her to.

Behaviour, not just fit

Here's the part most sizing tools miss, and it's the heart of the white paper's argument. Returns are not simply a fit problem. They are a behaviour problem wearing a fit problem's clothes.

The widely-quoted figure is that around 70 percent of returns are due to size and fit. Our white paper deliberately questions that number. It's recycled across industry reports without much independent validation, and it almost certainly flatters the truth. When a woman selects "wrong size" from a dropdown, she's picking the least awkward option. What often happened is that she ordered several sizes on purpose, always intending to keep one and send the rest back. Coresight found that 69 percent of Gen Z shoppers over-order with exactly that expectation. More than 90 percent of European consumers say free returns matter to them. The behaviour is baked in, and no amount of measurement alone will unpick it.

This matters for any brand trying to grow, because it reframes the goal. You cannot reduce returns by getting the size recommendation right and stopping there. You reduce them by helping a woman buy with intent in the first place: choosing garments that suit her shape, fit her measurements, and fill a real gap in her wardrobe rather than a passing impulse. Confidence at the point of decision is what turns a tentative order into a kept one.

We've seen what that does to the numbers. Across a community of around a thousand women, working on style, wardrobe and buying habits alongside fit, we observed return rates fall to about 5 percent over one to two years. One garment in twenty, against an industry norm of roughly one in three. That gap isn't a sizing algorithm. It's what happens when a purchase is considered, validated and genuinely wanted before the card is charged.

Beyond the try-on

This is why we think fashion AI has to grow up, and grow past the virtual try-on.

The try-on had a good run. Upload a selfie, watch a dress appear on your body, feel for a second like the fitting room came home with you. It's a lovely trick and it sells the idea of certainty. But a render only answers one question, and not the important one. It shows a woman what a garment might look like. It says nothing about whether it will fit, how the fabric will sit, or whether the cut actually suits the way her body is built. A flattering image on a screen and a garment that works in real life are not the same thing, and the gap between them is exactly where returns live.

What a woman needs is closer to a proper fitting than a photo filter. That means three things working together, not one in isolation.

It means visuals she can trust, grounded in her real proportions rather than a smoothed-over outline. It means accurate size guidance built from a genuine three-dimensional understanding of her body, talking to the real fabric, ease and cut of the specific garment in front of her, so the recommendation reflects this dress, in this material, on her. And it means personalised style advice: help understanding which shapes, lengths and lines suit her, so the choice is right before fit is ever in question.

A body scan makes all three possible. Two minutes on her phone, filmed once, and the measurements belong to her afterwards. The video is deleted immediately. What stays is a profile that every future purchase can draw on, getting a little smarter about her each time. That's the shift we're building toward: from a novelty that decorates the shopping experience to a full virtual fitting that actually informs it.

Why the right sales compound

For the woman, the payoff is simple and human. Fewer disappointing parcels. Fewer trips to return something. A wardrobe that slowly fills with clothes she reaches for instead of clothes she meant to send back. Shopping that asks her to guess a little less each time.

For a brand, the same shift shows up on the balance sheet and in the relationship. A sale that stays is worth more than two that don't, once you count the reverse logistics, the repackaging, the items that never make it back onto a shelf, and the emissions attached to all of it. Returns generate more than 25 billion kilograms of waste globally every year, and return transport alone accounts for over 10 million tonnes of CO2 in Europe. Every avoided return is money and carbon that never had to be spent.

But the deeper reward is trust. A woman who opens three parcels in a row and finds that each one fits stops hedging. She stops ordering two sizes as insurance. She buys with intent, keeps what she buys, and comes back, because the brand has quietly proven it understands her. That is what scaling the right sales actually looks like. Not more orders and more returns in lockstep, but a slower, sturdier kind of growth built on purchases that were right the first time.

The brands that win the next decade of online fashion won't be the ones that sell the most. They'll be the ones that sell the best: the ones that helped a woman find the thing that suited her, fit her, and stayed in her wardrobe. Growth like that doesn't need to be scaled back. It just adds up.