Back to all
At a glance
- 33% of shoppers have returned something because it did not match the description or photos. The listing caused that return, not the courier.
- Listing accuracy is the one large returns driver you control. Bracketing, review sentiment and carrier performance are not.
- In the UK and Europe you answer for accuracy and safety whichever party is seller of record. Handing a brand the transaction does not hand over the liability.
Every returns dashboard has a category that should embarrass the business more than it does: item not as described. The customer did not change their mind and the courier did not damage the box. You told them the product was one thing, they paid for it, and something else arrived.
You control that number, which is what separates it from everything else on the dashboard. You cannot change how many shoppers order three sizes meaning to keep one, what reviewers write, or how a carrier performs on a Tuesday. You decide what your listings say. We surveyed more than 6,000 shoppers across the US, Canada, UK, France and Germany for our 2025 Global Returns & Profit Impact Report: 33% had returned an item that missed its description or photos, and 37% want more detail than your listings give them today.
Where the wrong data comes from
If you sell third-party brands, look at onboarding, or product induction. Product information reaches your site through a chain of manual translation: the brand’s spreadsheet, an enrichment team’s interpretation of it, a copy-paste into the PIM. Every hop lets a dimension drift, a colour name mistranslate, a specification land on the wrong SKU.
One product ends up wearing another product’s attributes. The listing becomes the defect, and every unit you sell against it is a return already in transit.
The refund is the cheapest part of what follows. Reverse logistics, inspection, repackaging and markdown eat margin on every unit that comes back. Marketplaces also read elevated “not as described” rates as a quality signal, which cuts your visibility and then your sales.
Why the cost never reaches the board
Most retailers code returns by what happened to the unit. It comes back, gets inspected, gets graded, gets resold or written off. Almost nothing records that a wrong attribute started it, so nobody owns the cost and no line item names it. Merchandising books it as a supply chain problem and supply chain books it as a merchandising one. Our 2026 Commerce Readiness Index surveyed 200 retail and brand executives across the US and UK, and nearly three in four admit they make decisions on outdated or incomplete data.
In the UK and Europe you are liable either way
European retailers increasingly describe third-party trade by who is merchant and seller of record rather than by dropship or marketplace. Most do not want the brand as seller of record, for two reasons.
Revenue comes first. Seller of record decides who books the top line.
The law comes second. Sell through your own channel and you answer for the accuracy and safety of what you sell, even when the customer’s card statement names the brand. A regulator will not accept a redirection to the supplier. That responsibility stays with you when the transaction does not.
Giving up revenue therefore buys you nothing. You carry the accuracy risk whichever commercial model you pick, whether traditional wholesale, cost price or commission, so keep the top line and control the data. This is why European induction templates ask for so much: teams capture every attribute in writing rather than answer later for a wrong one.
Why European returns cost more than American ones
Return behaviour runs hotter in Europe, against thinner margins. Our analysis of millions of EMEA marketplace orders puts refund rates in Austria and Poland above 60%, driven by bracketing. German shoppers bracket more than any other market we surveyed at 43%, with the UK at 38% against a 36% global average. Rates swing by up to 20 percentage points between countries.
The same wrong attribute costs you more here than it costs a US retailer. European commerce teams ask for a technical fix at the point where their US counterparts still fund another ten people.
Guardrails beat cleanup
Retailers with the lowest description-driven return rates do not run bigger QA teams. They validate at the point of entry. Brand data has to clear their taxonomy and standards before a listing can exist, so a product cannot go live wearing another product’s dimensions. Uniform, complete data is also why the best product experiences in European retail feel consistent no matter which of a thousand brands supplied the item.
Cleanup after go-live never catches up. Your catalogue grows faster than your audit team, and every week a wrong listing stays live it manufactures more returns. Zalando, a Rithum customer, cut size-related returns by 10% after putting more precise sizing on its product pages. Correcting a field costs pennies. Processing the returns that field generates costs orders of magnitude more.
Three numbers to pull this week
- Your “not as described” return rate by category, benchmarked against your overall rate.
- Your fully loaded cost per return, including reverse logistics, inspection and markdown, multiplied out. This is the size of the prize.
- The share of those returns tracing to third-party brand data rather than your own content team. It is almost always the former.
Rithum works with some of the world’s largest retailers on getting products from third-party brands and suppliers live faster, whether the commercial terms behind the assortment are traditional wholesale, cost price or commission. Fewer wrong attributes at induction means fewer boxes coming back. Talk to our team.
This is part two of our product induction series. Part one covered the headcount behind manual onboarding. Part three covers why time-to-online is the growth metric nobody owns.
