A brand turns over £4m one year and £5.6m the next. Every channel is up. Paid social is scaling, email is compounding, the wholesale book has grown. Then the year-end accounts land and net profit is lower than it was on the smaller number.
This is the most common conversation we have with founders, and almost nobody sees it coming.
Growth hides margin compression. The three costs that eat it, paid acquisition, cost of goods and returns, all scale with volume, and every one of them sits in a different system from the revenue figure the whole team is watching.
Revenue is the least useful number you look at every day
Revenue appears in every dashboard, every board pack and every Monday update. It is also the only figure in the business that tells you nothing about whether the business is getting healthier.
A brand can grow revenue 40% by discounting harder, paying more to acquire a customer it already had, and pushing a category with a return rate it has never measured properly. All three feel like growth in the moment. Two of them destroy the year.
What actually matters is what survives to the bottom of an order. That number is assembled from four systems that do not speak to each other: Shopify holds revenue and cost of goods, Meta and Google hold the spend, the 3PL or returns portal holds the reverse logistics cost, and the accountant reconciles the lot eleven weeks after the quarter closed.
By the time the picture is complete, the decisions it should have informed were made two seasons ago.
Acquisition got more expensive and nobody rebased the target
Customer acquisition costs have risen roughly 40% between 2023 and 2026, driven by platform competition, the collapse in targeting precision after iOS privacy changes, and simple saturation in every established category.
Fashion DTC now sits at a customer acquisition cost of roughly $66 to $72. Across the sector, DTC net margins have fallen from the 8% to 15% range of the cheap-money era to somewhere between 3% and 10% today.
Here is the shape of the problem on a single order. A brand paying $70 to acquire a customer on a $95 average order value, carrying $15 of cost of goods and $12 of variable fulfilment, has around $13 left before acquisition cost. The first order is deeply negative.
That is not necessarily a broken model. It is only broken if nobody knows it, and the repeat purchase rate that is supposed to rescue it has never been measured. Most brands scaling paid social are running exactly this maths without having written it down.
Returns are a margin line, not an operations line
Returns are almost always owned by whoever runs operations, reported as a logistics metric, and reviewed quarterly if at all. In fashion they are the single largest silent drain on contribution.
Apparel return rates run between 20% and 40%, with fashion averaging around 25% overall. Footwear carries the highest subcategory rate at roughly 31%. Womenswear and fast fashion push toward the top of that range, and returns spike 30% to 50% in the weeks after the holiday season.
Compare that to physical retail, which sits around 8.7%, and the structural disadvantage of selling this category online becomes obvious.
Each returned item costs between $10 and $20 to process, before you count the margin lost on a unit that comes back unsellable at full price. Fit and sizing drive roughly 53% of apparel returns, which is the encouraging part: it is the one cause a brand can genuinely reduce through better product content, size guidance and PDP design.
The point is not that returns are too high. In fashion they will always be high. The point is that a 31% return rate on footwear is a pricing and merchandising input, and most brands file it as a warehouse statistic.
The number to actually run on: contribution margin after returns
One figure replaces the dashboard sprawl. Take the revenue on an order, then subtract cost of goods, fulfilment, payment fees, the blended cost of processing the returns that order cohort will generate, and the acquisition spend attributable to it.
What is left is contribution margin after returns, and it is the only number that answers the question a founder is actually asking: did that sale make us money.
Run it by product and the picture usually reorders the range. The best seller by units is frequently not the best seller by contribution, because it is the product being pushed hardest in prospecting and returned most often. We covered that specific trap when we looked at how to get your best seller out of your prospecting ads.
Run it by channel and the same thing happens. Wholesale looks thin on percentage margin until you price in that it carries no acquisition cost, no returns liability and no per-unit fulfilment overhead.
Why your reporting says everything is fine
Margin compression gets found late not through carelessness, but because every tool in the stack is built to flatter its own contribution.
Platform-reported return on ad spend counts orders each platform believes it influenced, so Meta and Google will jointly claim more revenue than the business actually took. Blended reporting fixes the double count but hides which channel is working.
Neither view knows what came back. Neither knows cost of goods. A campaign showing a 4x return on ad spend on a category returning at 35% is not a 4x campaign.
It is a measurement gap rather than an effort gap. The scoreboard everyone reports on has stopped matching the thing everyone is trying to grow.
Where UK, EU and GCC brands diverge
The mechanics differ enough by market that a single blended view will mislead you.
For UK and EU brands, returns are the dominant variable. Free returns are close to a consumer expectation, rates sit structurally higher, and the post-Christmas spike lands in the same quarter as the year-end stock position. Any UK brand shipping into the EU now also carries duty and handling costs per parcel that did not exist two years ago, which comes straight off contribution unless it has been rebuilt into landed price.
For GCC brands, returns run lower but the equivalent leak is delivery. Failed and repeat delivery attempts, cash on delivery handling where it is still offered, and cross-border shipping from a UK warehouse into the Gulf all carry costs that never surface in a marketing report. In-country stock changes that maths materially, which is why fulfilment speed decides GCC expansion is a margin argument as much as a customer experience one.
A brand trading in both markets needs contribution measured separately for each. Blending them produces an average that describes neither.
What to fix in the next thirty days
None of this requires new software. It requires the numbers to be in one place and looked at deliberately.
Start by getting true cost of goods into Shopify at variant level, not an estimated average across the range. Without it, every downstream calculation is a guess dressed as a figure.
Then pull your return rate by product for the last twelve months and sit it next to units sold. Expect to find two or three products carrying rates far above the range average, and check whether any of them are currently sitting in your prospecting creative.
Calculate contribution margin after returns for your top twenty products by revenue. Twenty is enough to reorder your merchandising priorities, and it can be done in a spreadsheet in an afternoon.
Finally, agree the repeat purchase rate your acquisition cost actually requires, and check whether you are hitting it. If the first order is negative by design, the second order is not a nice-to-have, it is the entire business model.
Growth that does not improve the business is just volume
Fashion brands rarely fail because they could not grow revenue. They fail because they grew it on terms that got worse as they scaled, and found out from a set of accounts rather than from a dashboard.
The brands that hold margin through a growth phase are not the ones with better ad creative. They are the ones that decided early which single number they were running on, and made every channel, product and campaign decision answer to it.
