COD Return Rates: What They Cost and How to Cut Them
A returned parcel costs you both shipping legs and earns nothing. This is how to work out what your return rate is really costing per order, where returns concentrate, and the five changes that move the number.
Tassyir Team
8 min read
Key takeaways
A return costs you the outbound fee, the return fee and a week of stock — with no revenue against any of it. Budget it per order, not per month.
Returns concentrate. A handful of wilayas and one or two products usually account for most of the damage, so the fix is targeted rather than general.
The most common causes are addressing failures and price surprise at the door — both preventable at confirmation, not at delivery.
Product pages that oversell drive returns that no operational discipline can fix. If one product returns at twice your average, the photos are the problem.
Track return rate by wilaya and by product weekly. Monthly reconciliation finds the problem after you have already shipped a month of it.
In prepaid e‑commerce a return is annoying. In cash on delivery it is the whole game. A refused parcel earns nothing and costs you the trip out, the trip back, and a unit of stock that is unavailable to sell for a week or more.
Most stores know their return rate as a single monthly percentage. That number is almost useless on its own — it hides where the returns actually come from, and by the time you see it you have already shipped a month of them.
What a return really costs
Take a typical order: 3,500 DZD product, 400 DZD delivery fee, 1,200 DZD gross margin. Delivered, it earns 1,200. Returned, it costs you the 400 outbound and roughly another 400 on the return leg — so the delivered orders have to carry that 800 before anything is left.
At scale that arithmetic is unforgiving:
Return rate | Cost of returns per 100 orders | Effective margin per delivered order |
|---|---|---|
5% | 4,000 DZD | 1,158 DZD |
10% | 8,000 DZD | 1,111 DZD |
20% | 16,000 DZD | 1,000 DZD |
30% | 24,000 DZD | 857 DZD |
Those numbers assume the return leg is charged at the same rate as delivery, which is the common case. Going from a 10% to a 30% return rate hands back roughly a fifth of your margin — and it does so silently, because nothing in your order list looks different.
Returns concentrate — find where
A store‑wide return rate is an average of very different populations. Split it three ways before you change anything:
By wilaya. Carrier coverage is not uniform. Two or three regions frequently account for a disproportionate share of failures, and the fix there is pricing or carrier choice, not confirmation.
By product. If one product returns at twice your store average, look at its page rather than your operations. Something in the photos or the description is promising more than the parcel delivers.
By traffic source. Broad interest‑targeted campaigns produce weaker intent than retargeting or search. If one campaign’s orders return at 30% while the rest sit at 8%, that campaign is not profitable at any ROAS.
Once you can see those three cuts weekly, most of the work becomes obvious. Without them you are left adjusting things store‑wide and hoping.
Five changes that move the number
1. Say the total out loud at confirmation
Price surprise at the door is one of the largest single causes of refusal, and one of the easiest to remove. State the product price, the delivery fee and the total as one sentence on every confirmation, and make the customer say yes to that number specifically.
2. Re‑read the address, not just the wilaya
Commune, street and a landmark. A courier who cannot find the address marks it undelivered exactly as if the customer had refused, and it counts against you identically.
3. Give a delivery window
People answer the courier’s call when they are expecting a parcel. A customer who ordered eight days ago and has heard nothing since treats the call as spam.
4. Price high‑return regions differently
If a wilaya returns at 25% against a store average of 9%, its delivery fee should reflect that. Raising the fee there both covers the expected loss and filters out the weakest orders, which is exactly the effect you want.
5. Fix the product page before blaming the courier
Sizing charts, real photographs rather than supplier renders, honest descriptions of material and finish. A product whose page sets accurate expectations gets accepted at the door. One that oversells is generating refusals before the parcel is even packed.
What to do with the parcels that come back
Returned stock has a habit of quietly disappearing from inventory. Give it a defined path: inspect on arrival, restock what is sellable, write off what is not, and reconcile the count against what the courier says it sent back.
The reconciliation matters more than it sounds. Discrepancies between parcels you dispatched, parcels the courier delivered and parcels returned to you are where money leaks in COD — and they are invisible unless somebody checks every remittance cycle.
Watch it weekly, not monthly
Return rate is a lagging indicator with a long tail: a parcel dispatched today may not be marked returned for two weeks. That delay is exactly why monthly review is too slow — by the time the number moves, you have shipped another few hundred parcels on the same settings.
Weekly, by wilaya and by product, is the cadence that lets you catch a bad region or a bad product while it is still a small number.
Frequently asked questions
What is a normal return rate for cash on delivery?
Under 10% is healthy, 10–20% is workable if your margins are wide, and above 20% most COD businesses are losing money on growth without realising it. Rates vary by category — apparel and anything size‑dependent runs higher than electronics accessories.
How much does a COD return actually cost?
Roughly twice the delivery fee, plus the working capital tied up while the parcel travels out and back, plus the opportunity cost of that unit being unavailable to sell for a week or more. On a 400 DZD delivery fee that is 800 DZD of direct cost against an order that earned nothing.
Why do customers refuse COD parcels at the door?
In order of frequency: they were not expecting the delivery fee, nobody is home or reachable, the address was wrong, they changed their mind in the days since ordering, or the product does not look like the photos. Only the last one is a product problem — the rest are process problems.
Can you charge customers for failed deliveries?
In practice, no. There is no payment instrument to charge against, which is the defining constraint of COD. The only real defences are preventing the failure and choosing not to ship orders that look likely to fail.
Should you stop shipping to high‑return regions?
Rarely stop outright — usually price differently. Raise the delivery fee to cover the expected return cost in those wilayas, or require a stronger confirmation before dispatching there. Blanket restriction throws away profitable orders alongside the bad ones.
Tassyir Team
E‑commerce operations, Tassyir
We build the platform Algerian merchants use to run cash-on-delivery stores — orders, stock, couriers and finances in one place. Everything here comes out of the operational data and merchant conversations behind that product.
