Where does your return parcel actually go? We tracked 31 British returns (Part 6)

We bought from British shops. Our returns went to four different countries, and Britain was only one of them.

This is the last of four articles from our British study: forty nine brands, thirty one real self funded orders shipped to a German consumer address, every one of them returned through the shop's own standard customer flow. The first three covered market access, delivery promises and the return label. This one is about the routing and the money.

This is the multiplier. We bought from 31 British shops, shipped everything to a German address, sent it all back through the standard customer flow, and tracked where each parcel physically went, what the return leg cost, who paid for it, how long the money took and how much of it actually arrived. 

No customer was told any of this before buying. It is not on the product page, not in the confirmation email, not in the returns policy. And it turned out to be the single largest factor in how long anyone waited to be paid back.

As before, we name categories, never shops.

Key findings at a glance

Across 21 documented returns, customers recovered only 57% of their total spend. In the German domestic test, "over a quarter" of the money did not return. Cross border, 43% did not come back. Strip out the two anomalous cases - one refund paid on a parcel that was never sent, one on a parcel that was never delivered - and 19 clean returns yield 58% recovered and 42% lost. 

Four individual returns resulted in less than a fifth of the order value being refunded. In the worst case, the customer recovered just 10% of their spend. Three other single-order returns yielded only 11%, 19%, and 19% back to the buyer.

Return parcels from a single German address travelled to four different countries: 

the United Kingdom in 8 cases, Germany in 6, the Netherlands in 5 and Italy in 1. Destination country was documented for 20 of the 31 orders. Every one of these was a British shop, so in 12 of 20 cases the parcel never went to Britain at all - it went to whichever warehouse that brand's logistics contract points at, and the customer was told none of this before buying. This hidden routing directly affects the customer when they have to fund international shipping out of pocket or face long processing queues.

The Dutch hub was the slowest routing in the study, on both money and time. Returns routed to the Netherlands typically took 41 days to process, with customers recovering just 34% of their order value on average. By contrast, direct returns to the UK typically returned 76% of the money within 10 days.

On average, customers waited 16 days from starting a return to seeing their money, with actual wait times stretching from 4 to 62 days. By comparison, the standard in domestic German e-commerce is 1 to 2 days, and under 24 hours in the best cases.

Free returns, the German standard, did not exist here. Of 31 orders, zero offered a genuinely free return. Thirteen deducted the courier cost from the refund. Nine required the customer to arrange and fund the shipment themselves.

One shop charged 27 euro to return an order worth about 40 euro. That is 67% of the basket, paid by the customer, for the privilege of giving the goods back.

The money does not come back

Twenty one completed returns with a documented refund. Goods worth 738.00 euro went out. 422.40 euro came back. The gap is 315.60 euro, or 43% of everything spent. Remove both anomalous cases and nineteen clean consumer returns come to 680.00 euro out and 396.70 euro back: 58% recovered, 42% missing. The distribution barely shifts, which is the point. This is not one outlier dragging an average.

Some of these were multi item orders in which not every item was sent back, so this is not a pure measure of shortfall on identical goods. That cuts both ways, because partial returns do not explain the shape of the distribution.

At the bottom end, a footwear order worth 25.35 euro was refunded at 2.47 euro, which is a recovery rate of ten percent. A beauty order of 28.35 euro came back as 3.03, eleven percent. A womenswear order of 20.28 euro came back as 3.80, nineteen percent. Another womenswear order of 34.42 euro came back as 6.70, also nineteen percent. Three more sat in the mid thirties: 42.57 refunded at 14.38, 39.57 refunded at 13.41, 37.89 refunded at 13.36.

At the top end, over the same border in the same three months, a contemporary fashion brand returned 47.93 euro on a 56.79 order, which is 84%. A premium fashion house returned 64.07 on 68.39, which is 94%

Ten percent and ninety four percent. Same transaction type, same country pair, same season, same couriers.

The difference is not generosity. It is arithmetic. A cross-border return leg from Germany costs roughly the same whether the box contains a 25 euro item or a 70 euro one, and most shops deduct it at a flat rate alongside the outbound delivery fee they keep. On a 25 euro basket those two deductions are the entire refund. On a 68 euro basket they are a rounding error.

Which means the shops recovering 10% and the shops recovering 94% may be running identical policies. The customer with the small basket simply hit the floor first. Every brand in this test that landed under 20% recovery had a basket under 35 euro. Every brand above 80% had a basket over 55 euro.

Where does the difference go? Three places, and the German study named all of them.

The first is return shipping, deducted at source. Thirteen shops in this test carry the same sentence somewhere in their process: the cost of the return courier will be deducted from your refund. It is not disclosed at checkout. It surfaces at the moment the customer is already committed.

The second is the original outbound delivery fee, retained. This was the recurring pattern in Germany and it is unchanged here: refunds quietly reduced by what the customer paid to receive the parcel in the first place, explained nowhere the customer would look before shipping it back.

The third is the cross border return leg itself, priced into that deduction. A domestic German consumer return runs somewhere around 3 to 5 euro on general market rates. A cross border intra EU leg is closer to 8 to 15. From Germany back to Britain, more again. When that cost is passed to the customer through a deduction, the customer is funding the seller's warehouse geography without ever being told that is what the line item is.

Nobody disputes one euro. Somebody notices when 25 euro comes back as 2.50 and in this test that happened four times, always on the smallest baskets, because the deduction does not scale with the order.

And in neither case does the customer file a complaint. They do the arithmetic once, quietly, and then they either order again or they do not. You book the deduction against a lifetime value question you will never see the answer to.

The fix is three lines in a refund email: what came back, what did not, and why. Disclosing a deduction costs you nothing you were not already keeping - but it moves the customer from feeling cheated to feeling informed, and those two customers have very different reorder rates.

And in neither case does the customer file a complaint. They do the arithmetic once, quietly, and then they either order again or they do not. You book the deduction against a lifetime value question you will never see the answer to. The fix is three lines in a refund email: what came back, what did not, and why. Disclosing a deduction costs you nothing you were not already keeping.

Carrier, cost, and destination: What actual cross-border returns look like

Beyond routing, the customer experience was defined by who paid for the return label and which carriers were used. Here is what the return network actually looked like across the 20 orders where physical parcel destinations could be documented - including carrier flows, out-of-pocket costs, and processing delays.

Eight returns went to the United Kingdom. This is the honest routing, and it was also, surprisingly, among the fastest. Several UK bound returns refunded within 4 to 8 days, including two of the three quickest results in the entire test. The parcel crossed a customs border in both directions and still beat every continental option on the table.

Six returns stayed in Germany, using a local return address held by a British seller. The single fastest refund in the study came from here: zero days between the courier scanning the parcel and the money arriving. But this group also produced a 26 day case and a 62 day case, so a local address on its own is not the answer. What it does is raise the ceiling on how good the outcome can be.

Five returns went to the Netherlands, the continental hub model, and this was the slowest cluster in the dataset by a wide margin on both dimensions at once. Slowest to pay and thinnest when it paid.

One return went to Italy, from a premium fashion house. Refunded in 7 days at 94% of value, the best commercial outcome anywhere in the test.

Now the part that actually identifies the mechanism.

Three of those five Dutch bound brands belong to the same corporate group, share a back end and route German returns to the same Dutch facility. Their times from return initiation to money received were 42 days, 40 days and 43 days. Three brands, three separate customers, one shared operational decision, one identical outcome. Their recovery rates were 19%, 34% and 38%.

The fourth Dutch bound return came from an unrelated footwear retailer and refunded in 13 days - three times faster through the same country. Its recovery rate was 10%, the worst in the study.

So the Netherlands is not slow. A queue in the Netherlands is slow, and one group's queue is slow in a way that is stable to within three days across three separate brands and three separate customers. Meanwhile a direct route to Britain, which is physically further and crosses a real border, returned money in a median of 10 days at three quarters of value.

Whatever else is true about your returns operation, geography is not a fact you inherited. It is a decision somebody made once, almost certainly on a cost per pallet, and it is being paid for per return, every day, by a customer who was never told. And the cost is not the kilometres. It is how long the parcel waits at the other end before a human looks at it.

German shops refund in two days. Our median was sixteen.

The most quoted line from the German study was this: where the flow worked, money was back with the customer in 1 to 2 days, best case under 24 hours, and that, not the legal fourteen days, is the bar.

In our cross-border test of UK retailers, customers waited a median of 16 days to get their money back. With Shopreturns, local drop-offs and automated processing cut that turnaround down to domestic speeds.

The customer does not adjust for distance. They have been refunded in 24 hours by a domestic shop, and they are now waiting more than two weeks - and in six cases more than a month - for a British one. In their arithmetic you are not slower because you are international. You are roughly ten times slower than the market that trained them.

And the tail is worse than the median. Eight cases in this test took 32 days or more, and the slowest ran to 62 nine weeks from the customer asking for their money to the money arriving. In every one of them the delay was not carrier transit. Transit is a few days. The delay was the parcel sitting in a queue at a facility, often in a third country, waiting for somebody to look at it before anybody would authorise the money.

How the top-performing brands got cross-border returns right

One case sets the ceiling on what a return can cost the person making it.

A beauty retailer, order value roughly 40 euro. The return required the customer to pay 27 euro out of pocket for the courier - not deducted from the refund, but cash up front, before knowing whether the refund would clear at all. Our tester's note records that support offered no help of any kind and that the parcel simply had to go at her own expense. The refund, when it came, was 13 euro.

At that ratio the rational customer keeps the product. From a narrow view that looks like a saved return: no reverse logistics cost, no restocking, revenue retained. It is not a saving. It is a customer who has just learned that buying from you carries a 27 euro downside on every future order, and who will price that into the next decision they make about you.

Eight other shops also pushed the shipment onto the customer. Where the amount was recorded it ran at 6, 6, 6.95 and 8.70 euro. All disclosed, all legitimate, all in a market where the domestic alternative is zero.

Because the comparison your customer is running is not British shop against British shop. It is your checkout against a German competitor in the same category, with a free label in the box and money back inside 48 hours. That is the shop that set their expectations, and it did so long before they found you.

What the highest-performing stores did right

Three brands in this test refunded within a week at high recovery, through a flow the tester logged as simple:

• 94% of value back in 7 days
• 80% in 5 days
• 76% in 4 days

Two more refunded within a week at much lower recovery, which is its own lesson - speed and completeness are separate problems, and fixing one does not fix the other.

The three that managed both had nothing in common technologically. No shared platform, no shared carrier, no shared portal vendor. One had no returns portal at all.

They had one thing in common operationally: the parcel had a short distance to travel and somebody looked at it quickly.

A German return address, or a direct British route with a real scan to refund process behind it. The fastest case in the study, where the refund landed the same day the parcel was scanned, was a return that never left Germany. The best value case, 94% recovered in seven days, was a premium house that takes its returns seriously because its baskets are worth taking seriously.

The German study closed on exactly this, and it holds across the border. The shops that performed best were not the ones with the best returns page. They were the ones where the parcel had the shortest way to travel and someone looked at it quickly.

Check this against your own operation

Where does a German customer's return parcel physically go, and did you choose that routing or inherit it from a third party logistics contract signed for a different reason?

What is your median time from return label scan to refund issued? Not your policy. Your median, and then your ninetieth percentile.

What share of your refunds carry a deduction the customer was not clearly told about before they shipped the parcel?

If you run one continental hub for all EU returns, what does the slowest lane through it cost in days, and has anyone compared it against a direct route to the market it serves? Ours found the direct route to Britain four times faster than the hub built to avoid it.

Take your return rate, multiply it by your cross border return leg cost, and annualise it. Is that number in anybody's budget, or is it living inside a line called shipping?

The one sentence summary

In cross border British retail, customs is solved, tracking is solved and DDP is solved, while the refund arrives sixteen days after the customer asks for it, missing 43% of its value, from a warehouse the customer was never told about.

Your customers run this test on you every single day. They just do not publish the results. They churn instead.

,,Every unnecessary return mile costs margin. Redesigning your returns network brings it back.” 

Pawel Zakielarz, CEO, Shopreturns

One month of your returns on a single page: where they go, what they cost, how long they take, how much comes back.

Get in touch

About the study, methodology in one paragraph

This series is based on a live test we ran ourselves between May and August 2026. Forty nine British brands selling online direct to consumers, no marketplaces, spread across British heritage footwear and knitwear, country and outdoor and cycling, high street and mid market womenswear, premium fashion, and direct to consumer challengers in streetwear, activewear, sustainable clothing and beauty. Each was approached as an ordinary consumer with a German delivery address. Seventeen would not ship to Germany at all and one quoted delivery above the value of the product, leaving thirty one real, self funded orders with baskets of roughly 15 to 70 euro. One of those orders was fulfilled by a company we had not bought from, so thirty one orders involved thirty two businesses. Every order was then returned through that shop's standard customer return flow, with no special treatment and no contact identifying us as researchers. For each order we documented the delivery promise against actual delivery, the carrier, DDP or DDU status, cross border tracking in both directions, return label availability, the physical destination country of the return parcel, return shipping cost and who bore it, refund amount, refund timing, and every support interaction needed to close the case. Twenty six orders carried a delivery window we could measure against an actual delivery date. Twenty one produced a documented refund amount, and the recovery figures in this article are drawn from those twenty one, with the two anomalous cases identified in the text and the figures restated without them. Timing figures are drawn from the twenty orders where both a return initiation date and a refund date were recorded; one further case was excluded because its logged refund date preceded its logged return date. Return destination country was documented for twenty orders. Some were multi item orders in which not every item was returned, so those figures describe the money that came back against the money that went out, rather than a like for like shortfall on identical goods. Because one order per brand is an incident rather than a statistic, all results are reported as market and category level patterns, and no individual brand is named.