Selling apparel on Ozon: returns, sizing and the traps that eat the margin

Apparel is the category sellers ask for most and the one experienced sellers warn against hardest. In an interview with Korneliy Pugachov, founder of the Noviy Market platform, fashion is called the biggest category and the loudest request from sellers; the apparel sellers at the same table answer that it is the hardest and most expensive niche on a marketplace. These are sellers on their own experience, not Ozon documentation; each number belongs to when it was said.

Why the buyout rate is lower in apparel

The mechanism is specific to the category. As one Ozon training course puts it, there are categories where a 100% buyout rate is unreachable, and clothing is first among them: it fits some and not others, and somebody orders five T-shirts to take one. The Uzbek/Uzum podcast explains it through habit — the marketplace replaced the fitting room. At a bazaar, trying something on and saying no thanks is awkward; online it costs nothing. An order in apparel is often a fitting, not a sale.

In a podcast recorded in January 2026, Anton Zheleznyakov of the Pro Ozon Telegram channel reports the shift from his own analytics: entering bags roughly two years earlier, the Ozon buyout rate was 80-85%; at the time of the conversation, 50-60%. He blames payment on collection and loose return rules, saying used goods can come back without seller approval — while noting he never tested this himself.

The last row is not an estimate but the platform's gate: in Ozon's webinar, promo eligibility rests on complaints under 10%, rating above 3, and returns plus uncollected orders under 10%, of which at least two must hold. A high return rate cuts off traffic, not only margin.

Sizing, size charts and the card

Two sources contradict each other here and both deserve writing down. Ozon's webinar is unambiguous: put your size grid in the photographs too, because the grid uploaded through card settings has to be clicked and not everybody clicks, whereas everybody sees the photos. The seller course flags the opposite-looking mistake — putting the size grid on the second photo of a dress, which it calls thoroughly odd, because the second slide exists to hold attention. The two reconcile: in the photo set, just not in frame two.

Shevchenko's teardown of a real card shows it in practice. Sizes appear as the abbreviations OG, OT, OB with bare numbers beside them — ninety of what? A second grid sits at the bottom, where nobody scrolls. The same course notes that showing available sizes on the main photo carries weight in apparel.

The persuasive image does not come from the seller. Buyers are used to seller shoots on models where every size sits perfectly, so in clothing they drop straight to the reviews to see how the garment falls on real people. The webinar also advises merging colour and size variants into one card, which brings an apparel-specific oddity: asked why sizes of one item get different promo discounts, Ozon answers that this happens when prices differed per size during the median window, or variants drifted into different navigation categories. Discounts always come from the 30-day median.

How returns eat the margin

A return is not a cancelled sale but a second logistics invoice: Zheleznyakov and the course's unit-economics sheet agree that reverse logistics costs the same as the outbound leg. The structure he describes for Ozon in January 2026: a base fee by volume in litres, a surcharge as a percentage of price when average delivery time is poor, plus last mile at roughly 5.5%. On a small item of his own, logistics runs about 100 roubles at 30-33 hours and approaches 180 at 50-60 hours once the back-and-forth is counted; at 80% buyout it totals 10-15% of price. He quotes from memory; tariffs move every couple of weeks.

The multiplier is the buyout rate, typed into the unit economics by hand — the course stresses it comes from no tariff table. In Shevchenko's case the seller has no answer when asked for hers.

The costliest version shows up in advertising. Shevchenko explains why pay-per-order campaigns trap apparel sellers: you pay 20-21% of the price per order, but that is only an order; at a 33% buyout rate you pay three times for one sale, so on a 1,000-rouble item 600 goes to ads. Zheleznyakov runs the same tool only at launch, four or five days.

Pricing follows. Shevchenko's rule: against a cost price of 800 roubles, sell at three times that — expensive logistics, low buyout. Zheleznyakov is harsher: apparel works only if you buy at around 500 roubles and mark up seven or eight times, or drive external traffic from bloggers, TikTok and VK, which raises the buyout rate. Competitive niches carry 10-12% margin, he says; in clothing even that is gone. He puts fashion commission at 13% rising to 43-46%, but figures from 8% to 55% circulate across these videos — check Ozon's tariff page instead.

The second quiet margin-killer is automatic promotions. Shevchenko's seller set a minimum price of 1,600 roubles and found the item selling at 834 — against 800 cost, a loss. Opting out lasts only a month; skip the repeat and the item slides back in. Discount arithmetic misleads too: an apparel producer modelled a 10% cut (a Wildberries club discount) lifting revenue 20%, and margin fell from 28% to 24% — on 100,000 roubles, profit rose by 728.

Season: where the cash freezes

Both of Zheleznyakov's million-rouble burials are seasonal. In the first, a wrong batch and negative reviews ruin the card's conversion and the stock then cannot be sold, because nobody wants a big black bag in summer; roughly a million froze and he recovered 700-800 thousand. In the second, 2,500 bags sat in fulfilment for six months and cleared at a 700,000-rouble loss.

Stock freezes in transit too: delivery to Novosibirsk takes 18-20 days, so shipping against a one-month turnover can mean summer is over on arrival. Geo-promos, which let sellers price a distant region out of reach, have been removed, so holding a good average delivery time is harder. One exception helps: the delivery-time surcharge does not apply for a new item's first 28 days — a real window for a seasonal launch.

Ozon's high-season webinar, for the 2023 season, adds the calendar: goods arriving 15 September to 14 October earn litre bonuses, later arrivals do not. Localise supply — 72% of 11.11 sales in 2022 were regional.

Finally, depth versus breadth. Zheleznyakov's closing advice is direct: do not enter niches with many sizes and many colours, because very large amounts of money freeze there. Shevchenko says the same through his case — two colours on a 100-piece run was a mistake; black sells best in basic apparel in Russia.

Mistakes the sellers report themselves

• Mismatching budget and category: entering apparel with 200,000 roubles, when apparel needs volume.
• Not knowing your own numbers: buyout rate, add-to-cart and order conversion.
• Leaving the card unpromoted. A new card's first two or three weeks are an investment; a high ad-spend ratio there is normal.
• Choosing a niche by turnover: bags show some 1.2 billion roubles of monthly order turnover, but hold 40,000 sellers.

Short summary

• An order in apparel is often a fitting; the buyout rate is structurally low and is the main multiplier, entered by hand.
• Reported rates are dated: 30-50% for fashion (Zheleznyakov, January 2026), 50% at best and 33% typical (Shevchenko, 2025).
• Past 10% returns plus uncollected orders, Ozon promos get hard to enter.
• Reverse logistics costs the same as the outbound leg and grows with delivery time.
• At a low buyout rate, pay-per-order advertising charges several times for one sale.
• Size chart in the photos, not in frame two; buyers' own photos persuade more than seller shoots.
• Seasons freeze cash, transit included; a new item's first 28 days carry no delivery-time surcharge.
• Do not run many colours and sizes on a small batch.

These traps become report lines once you track buyout rate, size mix and return share per card — which is what the Ozon panel in Commbot watches.