"We're a sports brand doing about £1.5m a year. Gymwear, trainers, a couple of teamwear lines. Our own site plus TikTok Shop. When we built the plan we put returns in at 10% because that's what someone said was normal. Our actuals say roughly a quarter of orders come back, and every return seems to cost more than the refund itself. How should we model this, and where is the money really leaking?"
It is the right question, but it is not one number. A return hits contribution in five places: the margin on the unwound sale, return carriage, processing, outbound carriage and the write-down on stock that cannot be resold at full price. Sizing and fit are the number one stated reason for returns, which is the good news, because fit is fixable with information rather than discounts. Here is the direct answer, on the illustrative model below: a £1.5m sports brand running a 24% return rate loses about £11 of contribution per order placed against a plan that assumes everything sticks, and roughly £94,000 a year even after the refunds are deducted. Deducting the refunds handles the revenue side. The residue has to be modelled, priced and managed.
Every figure below is an illustrative example with synthetic planning numbers, not any client's books. Steal the method, then swap in your own data.
Here's the Short Version
- Sizing leakage is the contribution lost on the unwound sale, plus return carriage, processing, lost outbound carriage and the write-down on stock that cannot be resold at full price
- Model returns as a deduction from net revenue, at your own trailing rate, by channel, category and size. A blended average hides the one style that returns at double the rate of everything else
- UK clothing returns average 23.6% of ecommerce orders, and fit is the number one reason: 61% of consumers say wrong size or fit is why clothing or shoes went back
- Bracketing, ordering two or three sizes to try at home, is a default behaviour: 36% of consumers admit to it, and more than half of under-35s
- The costs that trail a return, carriage, processing and dead stock, come to about £15.60 per returned order in the model, on top of the margin the unwound sale never delivered
- Each point you take off the return rate is worth roughly £11,000 a year to the illustrative brand: the handling you don't spend plus the margin on the sale that now sticks
- The fixes that work are boring and cheap relative to the leak: garment-measured size charts, fit reviews with real photos, a spec check on each production run, and an exchange-first returns flow
What Sizing Leakage Actually Means
Founders treat a return as one event: money back to the customer, stock back to the warehouse, life continues. In the accounts it has several moving parts, and only the first is visible on a settlement statement.
One: the sale unwinds. The revenue comes off, the stock goes back on the balance sheet, and the contribution you had banked disappears. If your plan assumed every order sticks, this is where the first gap opens.
Two: the handling. Somebody pays to move the parcel back, somebody inspects, re-bags and re-shelves it, and the outbound delivery you paid on the original order is gone for nothing. On TikTok Shop, change of mind carriage is covered by TikTok or by you depending on who arranged the return, and a product issue lands on you. On your own site, who pays postage depends on what you told the customer before they bought.
Three: the dead stock. A returned item is not automatically sellable: worn, washed, missing tags or out of season, a slice of everything that comes back gets written down or liquidated, and that slice is where the compounding losses live.
Four: the planning and cash timing. This is not another cost to add to the sale unwinding. It is where the same loss appears in your marketing maths, stock plan and cash forecast. The acquisition cost and outbound delivery are already spent, but the expected contribution did not arrive. Cash leaves when you refund, and stock only becomes available again after it is received, checked and reprocessed.
The Numbers Behind The Leak
Retail Economics and ZigZag's UK returns benchmark forecasts £25.1bn of non-food online purchases sent back in 2025, down from £26.7bn in 2024, and ZigZag's report puts clothing returns at an average of 23.6% of ecommerce orders. Clothing and footwear lead every market studied: in Rithum's 2025 Global Returns and Profit Impact survey, 68% of consumers had sent items from that category back in the past year.
In the same survey, 61% of consumers said wrong size or fit was the top reason they returned clothing or shoes, and a third said the item did not match the description or photos. Better information is the fix they are asking for: 39% say improved size and fit recommendations would significantly reduce their returns, and 31% say real life customer photos would make them less likely to send something back. Customers are telling you what would stop the leak.
Sizing is also a behaviour, not just a mistake. Bracketing, buying two or three sizes and returning the extras, is admitted by 36% of consumers globally and more than half of under-35s. A style can convert well and still return badly, so judge it on kept sales and contribution after returns. In sports kit the problem concentrates: compression fits read as small, footwear half sizes and width are guesswork. The table below shows what a blended rate hides.
| Line (illustrative) | Return rate | What it is telling you |
|---|---|---|
| Gym tees and shorts | 16% | Sizes are behaving, pictures and descriptions are doing their job |
| Compression tops | 34% | Fit anxiety: customers order two sizes and keep one |
| Trainers | 29% | Half sizes and width need more detail before the click |
| Blended | 24% | The number nobody can act on |
How To Model It In Your Numbers
Five rules do most of the work. None of them need new software.
Start from net revenue, not gross. Sales returns are a deduction from the sale they came from, not a cost line invented later. Use delivery dates, not order dates. The statutory cancellation window runs for 14 days after delivery. For eligible products, TikTok Shop lets customers start a change of mind return within 30 days of delivery.
Track three rates, not one. Your return rate by value, your cost per return, and your recovery rate on returned stock. A blended return rate with no cost per return tells you something came back, never what it cost.
Split it three ways. By channel, because a TikTok Shop return and a Shopify return behave differently, and the fee mechanics differ, as we broke down in the cost stack of a £1m TikTok Shop sports brand. By category, because trainers and tees are not comparable. And by size, because that is where the actionable detail lives. One size of one style returning at twice the blended rate is a fixable information problem hiding inside every average you report.
Model exchanges separately. An exchange keeps the sale alive: you burn the return carriage, the processing and a second outbound label, and you keep the margin. A refund loses the margin, the handling and often the customer's momentum. Track the split, or you cannot tell whether your returns process is protecting revenue or just processing losses.
Review monthly, at each management accounts date. Set the refund estimate from your trailing rate at each reporting date, the way we walked through for a supplement brand in returns and expiry accruals. The mechanics are the same; in sports, more of the exposure is sizing driven.
The Illustrative Example: A £1.5m Sports Brand
Here is the model for the brand in the opening scene. Prices are till prices with VAT stripped out the way they land in your accounts: a £60 order shows as £50 of net revenue. Landed cost is £20 a unit, so a kept order contributes £30 before return costs. A 24% return rate by value means 24 orders in every 100 come back, treated as an average order returning in full for simplicity. Synthetic planning numbers, not a client's books.
| Line, per 100 orders placed | Amount | What it is |
|---|---|---|
| Net sales, accounts basis | £3,800 | 76 kept orders at £50 |
| Cost of goods on kept orders | £1,520 | 76 units at £20 landed |
| Contribution before return costs | £2,280 | The real starting point of your plan |
| Return carriage | £84 | 24 returns at £3.50 a parcel, blended across channels |
| Processing and repacking | £144 | Inspection, steaming, re-bagging, put-away at £6 a return |
| Outbound carriage lost | £84 | You paid to send 24 orders out for nothing |
| Dead stock write-down | £62 | One in five returned units cannot be resold at full price, recovered at 35p in the pound |
| Return-adjusted contribution | £1,906 | What the hundred orders actually delivered |
Two plans meet these numbers. Plan A assumed every order sticks: it expected £3,000 of contribution and got £1,906, a gap of £1,094, or about £11 an order. Plan B deducted the refunds from revenue but nothing else: it expected £2,280 and got £1,906, a gap of £374, or £3.74 of hard cost for every order placed. Most brands model Plan B and believe returns are handled: it nets the revenue, but it does not touch the cost of the return.
Now scale it. If £1.5m is annual checkout value before returns and the average order is £60 including VAT, that is 25,000 orders a year, so multiply the per-100 model by 250. Plan A misses by about £274,000 of contribution. Plan B still misses by about £94,000 a year. Those costs are usually spread across fulfilment, warehouse and stock write-down lines rather than shown as one returns line. Take a point off the return rate and you avoid 250 returns: £15.60 of carriage, processing and dead stock saved on each, plus the margin on sales that now stick. About £11,000 a year per point. That is the number that should decide your fit budget.
What It Does To Your Pricing And Your Marketing
Returns quietly rewrite two numbers you rely on. The first is your cost per order. If 24 orders in 100 come back, the spend that produced 100 orders produced 76 keepers, so your acquisition cost per keeper is 31.6% higher than the blended figure. At £18 a blended order, that is £23.70 a keeper. Judge campaigns on kept orders, or budget the difference somewhere honest.
The second is your price. With a 24% return rate and £15.60 of carriage, processing and dead stock cost per return, every order carries £3.74 of expected return cost before margin. That cost either sits inside your price or eats your margin, which is why the brands with the best return rates can afford the sharper price. On VAT: refunds come off your output VAT in the period you pay them, not when the order was placed.
The Fixes That Take Points Off The Return Rate
Start with the data you already own: pull 90 days of returns by style and size and rank the worst offenders. The sizes that run hot, the styles whose return reason is consistently fit, the gap between "fits small" in the reviews and a size chart that says nothing: the fix list falls out of it.
Fix the information first, because it is nearly free. Size charts with garment measurements in centimetres, not just body measurements, so a customer can measure a top they own and match it. Model details with the size worn and the model's height. Real customer photos carrying a real fit signal. Rithum's report credits Zalando with cutting size-related returns by 10% after rolling out more precise sizing tools, and 39% of shoppers are asking for exactly this.
Then fix the product. Approve a measurement spec per garment and check a sample from every production run against it. Sizing drift between runs, one factory cutting a shade tight, is a silent generator of return spikes nobody connects back to production. For footwear, publish width and half size guidance and stop treating "true to size" as an assumption. And for anything sold on fit psychology, such as compression or shaping, write the fit note that tells the customer how it will feel, not just what size to buy.
Then fix the process. Make the exchange the easiest path in the returns flow. An exchange keeps the revenue, but it still costs return carriage, processing and a second outbound shipment. And a style with a 34% return rate needs a price change or a fit fix before it gets more ad spend: scaling a leaking style scales the leak.
The Legal And Platform Floors
Your policy is your exposure, so know the floor before you raise it. Online, the customer can cancel for any reason within 14 days of the goods arriving, with another 14 days to send them back, and you refund within 14 days of getting the goods back or seeing evidence they were sent, standard delivery included. You cannot charge a fee for making the refund. You can reduce the refund only for loss in value caused by handling beyond what a shop would reasonably allow, up to the contract price. If you never told the customer about their cancellation right, the cancellation period can be extended by up to 12 months and you cannot make that deduction.
Fail to make return postage terms clear before the sale and the direct cost of the return is legally yours. TikTok Shop goes further than the statute for eligible products. Customers can start a change of mind return within 30 days of delivery. TikTok covers change of mind return shipping when it arranges the return, and you cover it when you arrange the return. You pay return shipping for a product issue. A refund can be held until you receive the goods or the customer provides proof of return, so do not assume every refund leaves exactly 14 days after the request. On Amazon UK, refunding a customer returns most of your referral fee, minus a refund administration fee of the lesser of £5 or 20% of that fee. Do not promise more than you have modelled: a generous promise has a price.
FAQ
What return rate should I model for a sports brand?
Start from your own trailing six months, by channel and category, before borrowing anyone else's benchmark. With no history, use the UK clothing average of 23.6% as a sense-check, not as your single planning assumption. Run a range and set higher cases for fit-critical lines such as compression kit and footwear.
How do I work out what a single return costs me?
Add four things: the return carriage (whoever ends up paying it), the handling time to inspect and re-sell the item, the outbound delivery you never get back, and the write-down on stock that cannot be resold at full price. In the illustrative model that comes to £15.60 per returned order, before the margin you lose on the sale itself. Replace the planning figures with your own courier quotes and warehouse timings.
Do I have to accept every return?
For online sales, the 14 day change of mind right comes off only in specific cases, like personalised items or sealed goods that cannot be returned for hygiene reasons once opened. Faulty goods are different: customers have strong statutory rights there and you cannot contract out of them. On TikTok Shop, eligible items can be returned within 30 days of delivery. Read the exceptions properly before building a policy on them, and never advertise a no returns line: restricting or misleading customers about their rights is against the law.
The Bottom Line
Model returns like this: net revenue from your own trailing rate, a cost per return covering carriage, processing and dead stock, an annual view that turns the rate into contribution, and a size level report that finds the leak. On the illustrative numbers, a 24% return rate costs about £94,000 a year beyond a plan that deducts refunds but ignores the hard costs of returns, and every point off it is worth around £11,000. The refund was never the problem. The problem is the costs that travel with it, and the pricing that never accounted for them. Get the size data right and the rest follows, because fit is the rare cost line you can fix with better information rather than more discounting. Turnover is vanity, profit is sanity, and cash is reality.
If you want your return rate translated into contribution, size by size and channel by channel, that is the work we do. We are specialist social commerce accountants for UK ecommerce brands, and we covered the wider picture in how a 7-figure sports brand should structure their P&L. When you are ready, Book a call and bring your returns report and your last few months of channel statements.