A homeware store outside Ankara ships around 900 orders a month. In January the owner rewrote the store's return shipping terms, because the rule had changed and everyone said they had to be rewritten. It took twenty minutes. Nothing else about the operation changed.
By August the carrier invoice has a shape it did not have a year earlier. There is a group of shipments on it that nobody in the company created: parcels sent from towns the store does not ship from, at rates the store does not have, with no order number attached to any of them. They are returns. Some came through the branch nearest the customer, some came collect-on-delivery from a carrier the store has never worked with, and a few arrived as ordinary parcels with a handwritten note inside.
The owner can tell you the store's return rate to one decimal place, because the e-commerce panel reports it. The owner cannot tell you what the return shipping cost per order is, which carrier the returns arrive on, or which products generate the expensive ones. Those numbers do not exist anywhere, and they did not need to exist eighteen months ago — because back then, return shipping was mostly somebody else's problem.
That is the whole change, and almost nobody has adjusted to it.
This guide is about the shipping side of returns: who pays, on whose rate, on which carrier, and how you see it. If you are looking for how to reduce the number of returns in the first place — product data, sizing, return reasons, exchanges over refunds — that is the companion piece: E-commerce Returns Management. This post picks up where its "reverse logistics costs" section stops.
What Actually Changed on 1 January 2026
An amendment to Turkey's Distance Contracts Regulation (Mesafeli Sözleşmeler Yönetmeliği) was published in the Official Gazette on 24 May 2025 and entered into force on 1 January 2026. Three things in it matter to a shipping operation.
Return shipping cost belongs to the seller, with no exception. In a return made under the consumer's right of withdrawal, the consumer cannot be made to bear the cost of sending the goods back. Not through a deduction from the refund, not through a fee, not through a policy clause they accepted at checkout.
The pre-sale information now has to name a carrier for returns. Your pre-sale information form (ön bilgilendirme formu) must tell the customer which carrier you have designated for returns. If the customer uses that carrier, they pay nothing. If you named no carrier at all, you still cannot charge them anything — you simply lose the ability to steer where the parcel goes and what it costs you. And if the carrier you named has no branch where the customer lives, you are required to collect the goods from them without charging extra for it.
Consumer electronics came back into scope. Phones, smartwatches, tablets and computers had been carved out of the withdrawal right; that carve-out was repealed. If you sell those categories, a return volume you had structurally excluded is now structurally included.
For marketplace sales there is a fourth point worth knowing even though it is not your obligation: where the intermediary platform failed to include the carrier information, or the named carrier has no local presence, the cost and the obligation sit with the platform rather than the seller.
Two clarifications that save a lot of argument later. This applies to withdrawal-right returns — the customer changing their mind within the statutory window. Defective goods were already the seller's cost and remain so. And the rule concerns who pays for the shipping; it does not oblige you to accept a return that falls outside the withdrawal right entirely.
Regulatory detail here is described as it stood at the time of writing and is not legal advice. The operational conclusions are the point; confirm the current text and your own obligations with a lawyer before you rewrite a contract page.
Free Returns Stopped Being a Differentiator
This is the part that gets missed because it is strategic rather than procedural.
Free returns used to be a choice. Stores that offered them were spending money to win a conversion advantage over stores that did not, and the whole Western returns-optimization literature is built on that trade-off: how generous should your policy be, at what point does generosity stop paying for itself, when do you introduce a fee. Some of the best-known returns platforms are still publishing that argument in 2026 — how to stop absorbing the cost of free returns, how to make the customer share it.
In Turkey that argument is closed. Every store selling to Turkish consumers offers free returns now, because every store is required to. The lever did not get more expensive; it got deleted.
Which means two things. First, none of the imported advice about return fees, restocking charges and paid-return policies applies here — worth remembering, because most of the returns content that ranks well in English assumes the opposite legal baseline. In the EU, the default is still that the consumer bears the direct cost of returning goods unless the trader agreed to cover it or failed to say so before the sale. Turkey is now stricter than that baseline, not aligned with it.
Second, and more usefully: because the whether is settled, the only competitive room left is in the how. How fast the refund lands. How little effort the return takes. And — invisibly to the customer, decisively for you — how much the return costs you compared to the store selling the same product down the road. Two stores can publish word-for-word identical return terms and pay very different amounts to honor them, because everything that decides the amount happens after the policy page.
The One Lever Left: The Carrier You Name
The regulation left sellers exactly one point of control, and it looks like a text field in a legal document. It is not. It is a procurement decision that happens to be published on your website.
When you name a carrier you hold an agreement with, a return goes back on your contracted rate, with a reference number that ties the parcel to an order, on a service you can measure. When you name nobody, returns arrive however the customer chose to send them: retail counter pricing, whichever carrier had a branch nearby, sometimes collect-on-delivery, with nothing on the parcel that connects it to anything in your system. You pay in both cases. The customer's experience is nearly identical. Your invoice is not.
Three things decide whether naming a carrier actually helps you.
Coverage is now a liability, not a convenience
Read the collection rule again: if your named carrier has no branch where the customer lives, you have to collect the goods from their address at no extra charge.
That converts a carrier's branch map into an exposure map. Every district where your named carrier is thin is a district where a return costs you a doorstep collection instead of a counter drop-off — a different service, at a different price, arranged by hand, on your initiative rather than the customer's. If you sell nationwide, this is not a rounding error. It is the single most expensive way a return can reach you, and it happens automatically wherever your coverage is weak.
The practical consequence is that the carrier you name for returns is not necessarily the carrier you use for outbound. Outbound rewards speed and price on the routes where your customers actually are. Returns reward breadth — presence in the places you would rather not have to drive to. Those are different optimization problems, and the answer is allowed to be different. Our guide to tracking carrier performance by route covers how to see the difference in your own data rather than in a coverage brochure.
A named carrier without a return product is just a name
Naming a carrier is not the same as having a return service with them.
Turkish carriers generally handle merchant returns through a dedicated product rather than a normal shipment: the seller issues a return reference, the customer takes it to a branch, and the branch uses that reference to route the parcel back to the right seller under the right agreement and bill it to the right account. Yurtiçi's Kolay İade is the clearest published example — an arrangement e-commerce sites sign specifically for returns, which then lets customers hand parcels in at delivery points as well as branches. Other carriers run equivalents under their own names. The details, including how long a return reference stays valid, vary by carrier and by agreement, so confirm yours rather than assuming the numbers you read somewhere.
The important part is what the reference does. Without one, a returned parcel is anonymous. With one, it is a shipment with an order behind it, which is the precondition for every other number in this guide.
Naming more than one
Nothing stops you naming more than one carrier for returns, and there are reasons to. Coverage gaps in one network get filled by another. Bulky items and small items rarely belong on the same return service. If you already run more than one carrier agreement on the way out, you probably have the coverage to fix the collection-liability problem on the way back.
The condition is that what you publish and what you can actually do have to match. If the pre-sale information names a carrier, the customer has to be able to use that carrier without arranging anything themselves, and your support team has to be able to issue that carrier's return reference on request. A name in a form that nobody in the company can act on is worse than no name at all: it looks like control and behaves like chaos.
The Refund Clock Starts Before the Parcel Arrives
Here is the mechanic that surprises most sellers, and it has nothing to do with cost per parcel.
Your obligation to refund runs from the moment the goods are handed to the carrier named for returns — not from the moment they land on your dock. You owe the refund within fourteen days of that hand-off, and it includes the outbound delivery charge you originally collected, not just the price of the item.
Follow the consequences.
The refund is due while the parcel is still moving. On a two-day domestic lane that is comfortable. On a slow lane, through a drop-off point that consolidates before it moves, over a long weekend, it is not — and the money leaves before the goods arrive.
You are also refunding the outbound shipping. So a single return is at minimum three separate costs before you even count what the item itself loses: the outbound leg you paid for and are now giving back to the customer, the return leg you now pay for by law, and the handling on receipt. A return-cost model that counts only the return leg is missing the outbound leg it just gave away, and that is usually a charge of the same order of magnitude.
And notice what this does to inspection. The classical returns workflow is receive, inspect, decide, refund. The clock does not wait for that sequence. In practice, for most consumer goods, the decision has to be made on the strength of what the customer told you and what your own product data says, not on what you find in the box. That is a real risk transfer, and the honest mitigations are boring ones: photograph what leaves, log the condition of what comes back, and treat a pattern of damaged returns from the same customer or the same product as a data problem to investigate rather than a refund to fight.
An Uncontrolled Return Costs You Three Times
"Uncontrolled" means a return that came back on a route you did not choose. Under the current rule you pay for it either way, so the only question is how much and whether you can see it.
The shipment you never priced
A parcel handed over at a counter by a private individual is priced at counter rates, on a carrier that may have no relationship with you, sometimes collect-on-delivery so that the charge arrives attached to the delivery rather than to an invoice line. That is the most expensive way to move a parcel in Turkey and it is the default outcome of naming nobody.
The parcel nobody can attribute
This one is quieter and worse. A return that arrives without a reference cannot be matched to an order, which means it cannot be matched to a product, a customer, a reason, or a category. Its cost lands in a total. Ask "which product line generated the most return shipping cost last quarter" and the answer is unavailable — not hard to get, genuinely unavailable, because the link was never created.
That is the difference between a cost you manage and a cost you absorb. Everything in the rest of this guide depends on the link existing.
The inspection you didn't get to do
An anonymous parcel also arrives without context. Nobody expected it, nobody knows what should be inside, and whoever opens it has to reconstruct the order from whatever is in the box. In a small operation that is an annoyance. At a few hundred returns a month it is a person's job, and it is a job created entirely by the absence of a reference number.
The Number Almost Nobody Has: Cost Per Return
Ask a store what a return costs and you usually get the return shipping rate. That is one component of four.
A defensible cost per return is:
the outbound shipping you refund + the return shipping you pay + the handling and inspection labor + the value the item loses before it is sellable again.
Every one of those four is a number your business already has or can get. None of them is on any screen next to the others, which is why the total is rarely computed.
Two refinements make it useful rather than merely interesting.
Compute it per product, not per store. A store-level average is a blend of a cheap category and an expensive one, and it hides the only actionable finding — that a small number of products generate most of the return cost. Grouping by product is what turns "our returns cost too much" into "these eleven listings cost too much."
Compare it to the item's contribution, not to its price. A return is not free if the item resells; it costs both shipping legs plus the handling plus whatever value the item shed. When that total approaches the margin on the sale, the product is not really profitable at its current return rate, and the fix is upstream — the description, the sizing information, the photographs — not in logistics at all.
You do not need a system to start. Take one month, pull the returns you can identify, compute the four components by hand for the top twenty by volume, and you will know more than the reporting in your e-commerce panel will ever tell you.
Reading Returns on the Carrier Invoice
Return shipments have a habit of being invisible on the carrier invoice in a specific way: they are billed as ordinary shipments, at the current tariff, in the same block as everything else. Nothing on the line says "this is a return."
That matters for two reasons. It makes the return line item impossible to see without a separate record on your side, and it makes a common billing error impossible to catch — a return billed twice, or billed as a fresh outbound, or billed against a failed delivery that never moved at all. If you have been through our guide to auditing the carrier invoice, you already know the general principle: you can only check an invoice against a number you wrote down before it arrived.
For returns specifically, the record you need at the moment the return reference is issued is small: order number, carrier, service, origin district, expected weight or desi band, expected cost, and the reason code. Seven fields. With them, the return block on the invoice becomes checkable and the cost-per-product analysis above becomes a query instead of a project. Without them, both are guesswork, and the invoice defends itself by being mostly correct.
What You Can Still Control
The policy lever is gone. Three operational levers are not, and they are worth more than the policy lever ever was.
Fewer returns
The largest single category of avoidable return cost in most stores is not a withdrawal return at all — it is failed deliveries that come back to you. A parcel that goes out to a bad address, a wrong phone number or an unreachable customer travels twice and sells nothing. That leg was always your bill; the difference now is that it sits alongside a regulated return cost you cannot reduce, which makes the one you can reduce worth more attention. Address and contact quality is the highest-return fix available; our guide to reducing failed deliveries covers the specific causes in order of frequency.
After that comes the genuine change-of-mind return, and that is a product-page problem: measurements, materials, photographs that show scale, and honest delivery expectations. Published figures vary widely by category and by source, but most breakdowns of return reasons put fit and expectation mismatch near the top, and both are decided before the order is placed.
Cheaper returns
Cheaper here means "on your agreement instead of at a counter." That is the carrier-designation decision from earlier, executed rather than merely written down. It also means checking whether your return rate is the same as your outbound rate — it often is not, and returns can sit on a different service class than the one you negotiated for outbound volume. When you next revisit your carrier agreement, bring the return volume as its own line rather than letting it hide inside the total.
Faster returns
Speed is the underrated one. Every day a returned item spends in transit, in a receiving pile, or waiting for a decision is a day it is not sellable — and for seasonal or fashion goods, days convert directly into markdown. The interventions are unglamorous: a return reference issued the same day it is requested, a receiving process that runs daily rather than weekly, and an inspection step short enough that it does not become a queue. None of this requires software. All of it requires somebody to own the number.
What Not to Do
Do not try to recover the cost through a restocking fee or a deduction from the refund. It does not survive contact with the rule, and it converts a cost problem into a complaint and compliance problem. The complaint side of delivery failures is expensive enough on its own.
Do not name a carrier you do not have a return agreement with. A name you cannot act on is worse than no name — the customer expects a free, easy route that your team cannot produce, and you get the cost anyway plus the argument.
Do not raise product prices reflexively to cover it. Return cost is concentrated in a minority of products. Spreading it across the catalog taxes the well-behaved items to subsidize the badly-described ones and leaves the actual problem untouched.
Do not make the customer chase the reference. Every hour between a return request and a usable return reference is an hour in which the parcel is more likely to come back some other way — at counter prices, anonymously, at your expense.
Do not restrict returns to reduce the bill. Beyond the legal exposure, it is the one move that reliably converts a shipping cost into a lost customer, and the arithmetic there has never favored the store.
Do not wait for January to look at the numbers. More on that next.
Why September and October Are the Window
Two reasons, and the first one is structural rather than seasonal.
The return terms that govern a November order are the ones you published before that order was placed. You cannot retroactively designate a carrier for a sale that has already happened. Whatever your pre-sale information says on the day a peak-season customer checks out is what you are committed to for that order's entire return window — which, for orders placed in late November, runs well into December and beyond. If your named carrier is wrong, or missing, you find out in January and you cannot fix it backwards.
And January is when the volume arrives. The peak-season shipping wave is followed by a return wave a few weeks later, and it is the one part of peak that nobody staffs for, because it lands after everyone has moved on. Our peak-season readiness guide sketches this in a single section; the point here is narrower. January 2027 will be the first full-scale peak return wave under a rule where every one of those parcels is billed to you. If you want to know what that costs before it happens, the measurement has to start on autumn volume, while there is still time for the answer to change anything.
There is also a practical deadline hiding in this. Changing the carrier named in your pre-sale information is not an afternoon's work if you do not already have the agreement — a return service is a commercial arrangement, and commercial arrangements in this industry take weeks. Starting in September means it is live for peak. Starting in November means it is live for the returns of the following year.
Where Shipink Fits
Everything above is doable without a platform, and at low volume it is genuinely fine to do it by hand. The point where it stops being fine is when returns start arriving on more than one carrier, or when someone asks a question about return cost that takes a day to answer.
What Shipink is doing in this specific problem:
- More than one carrier agreement in one place. You can run your own negotiated agreements alongside Shipink's, across 15+ carriers, which is what makes it possible to name a carrier for returns based on coverage rather than on which single integration you happen to have.
- A branded return request page that works regardless of the e-commerce infrastructure underneath, so the customer's route to a valid return reference is one link rather than a support conversation.
- Return rules — automatic approval for the cases you want approved automatically, and defined conditions for the rest, so the reference gets issued on request instead of when someone gets to the inbox.
- Return requests from the tracking page, on the same page the customer is already looking at, which is where they are when they decide.
- Return shipment tracking in the same view as the outbound, so a return in transit is visible rather than a parcel that may or may not turn up.
- Return and damage rate reporting per carrier, which is where cost per return stops being a spreadsheet exercise.
- Address correction on the way out, which is the cheapest possible reduction in the return volume you never wanted.
The honest limits: Shipink does not negotiate your return service with a carrier for you, it does not decide which carrier belongs in your pre-sale information, and it cannot tell you what a return costs unless the return went through it. The lever is still yours to pull. The platform is what makes pulling it measurable.
The Return Shipping Cost Checklist
Compliance and configuration
- Pre-sale information names a specific carrier for returns
- That carrier is one you hold an agreement with, not an aspiration
- No fee, deduction or restocking charge is applied to withdrawal returns anywhere in the flow
- Someone in the company can issue that carrier's return reference on request, today
- You know what happens when the customer's district has no branch, and who arranges the collection
Cost visibility
- Every return leaves a record at reference time: order, carrier, service, origin, expected cost, reason
- Return shipments are identifiable on the carrier invoice, not blended into outbound
- Cost per return is computed with all four components, not just the return leg
- The number is broken out by product, and the worst twenty are known by name
- Refunded outbound shipping is counted as a return cost, because it is one
Operations
- Return reference issued the same day it is requested
- Receiving runs on a daily cadence, not when the pile gets big
- Return transit time is measured, because the refund clock is running during it
- Coverage gaps in the named carrier's network are known and priced
Before peak
- Named carrier reviewed and, if it is changing, the agreement started in September
- Baseline cost per return measured on autumn volume
- Someone owns the January return wave before December ends
Start With One Month of Data
The regulation removed a decision you used to make and handed you a cost you now carry. That is not a disaster — it is the same cost every store selling into Turkey now carries, which means the competitive question is no longer who charges for returns but who runs them well.
Running them well starts with a number. Take last month's returns, find the ones you can identify, and compute what each of them actually cost across all four components. If the exercise is impossible because the parcels arrived anonymously, that is the finding, and the fix is the carrier name in your pre-sale information.
If you would rather not build the record-keeping yourself, that is what Shipink is for — branded return requests, return rules, multiple carrier agreements in one place, and per-carrier cost reporting that covers the way back as well as the way out. Get in touch and we will look at your return flow with you.
