To reduce your cash-on-delivery return rate, stop looking for one big fix: it does not exist. What works in our stores is improving every lever we control by about 5%, from the product and the price to the ads, the courier, the warehouse and the order checks. Stacked together, those small gains bring the return rate down by 5 to 10 points per market.
This guide is for COD merchants who already have orders coming in and watch too many parcels come back. By the end, you will have the full list of levers we work on, what we do on each one, and a simple way to measure progress market by market.
Why should the return rate be the first thing you optimise?
Because it kills your margin. When a COD order comes in, you have already paid for it twice: the ad spend on Meta that brought the customer, and the shipping to send the parcel. A sale only exists once the customer pays at the door. Before that, nothing has happened.
We ran the numbers on the same store, with the same ad spend and the same shipping costs, once at a 30% return rate and once at 25%. The margin goes from 340 to 475. That is about 40% more profit for 5 fewer points of returns. Your exact numbers depend on your average order value and your costs, but the direction never changes.
The mistake we see most: merchants celebrate a big day of orders and never check the return rate. An order is not revenue; a delivered and paid order is.
For a sense of where you start, our data at COD Confirm shows COD refusal rates from 11% to 33% depending on the country.
Which return-rate factors can you control?
We split every factor into two families before we touch anything:
| Family | What it means | What you can do |
|---|---|---|
| Controllable | You can remove the cause | Fix it: product, price, funnel, comments, courier, warehouse, order checks |
| Structural | It will happen anyway | Reduce the impact: customer not home, no exact cash, a courier who does not really try |
We also separate front-end factors (product, price, ads, funnel) from back-end factors (courier, warehouse, checks, messages). Put your energy on what you control, and accept early that a 0% return rate does not exist.
The way we see it, your return rate is not a fact of life. It is twelve decisions, and you control most of them.
The checklist: every lever, what we do, what we see
| Lever | What we do | Impact we see |
|---|---|---|
| Product choice | Sell products with a strong pain point, a real problem the customer wants solved fast | The products we see scale always have a low return rate |
| Gift products | Favour products people buy for someone else (their pet, their kids, a relative) | Tend to come back less often |
| Impulse products | Avoid very compulsive purchases | The customer decides at the door they don't need it; returns go up a lot |
| Price | Test a lower price when the marketing is not strong enough | In our A/B tests, selling at 25 instead of 30 gave a clearly lower return rate |
| Ad funnel | Send traffic from the ad to an advertorial (listicle, blog-style page) before the product page | Lower return rate in our split tests |
| Ad comments | Delete negative comments on every ad, in every market; hide keywords on the Facebook page | Buyers waiting for their parcel don't get talked out of it |
| Courier | Pick the best courier per market, not the cheapest | The heaviest back-end factor; in Greece, changing courier alone cut returns |
| Warehouse | Ship from a warehouse close to the market | Delivery in 24-48 hours raises acceptance |
| Address check | Verify every address before it ships | No more parcels that go out and come back for a missing house number |
| Duplicates and phones | Catch duplicate orders and wrong numbers before shipping | Fewer parcels that could never be delivered |
| Customer messages | SMS, WhatsApp or email after the order and before delivery | The customer is reassured and is home on the day |
| Price rounding | Round the COD amount | The customer can pay with exact cash |
Step 1 — How do we pick products and prices that don't come back?
Everything starts with what we call emotional closing. If your return rate is high, it usually means your marketing does not convince the customer strongly enough that your product solves their problem. A customer who is truly convinced does not refuse at the door. They email you to ask where their parcel is.
That is why problem-solution products and gifts come back less, and compulsive products come back more: the impulse fades before the courier arrives. Price is friction at the door. The customer has to hand over the full amount in cash. In our stores, the more expensive products come back more often. When we A/B tested a product at 30 against 25, the 25 version had a much lower return rate. If your marketing cannot justify a high price, lower it or make the marketing stronger.
For how we find products in the first place, see our product research method for COD.
Step 2 — How do we build an ad funnel that lowers returns?
Perceived value comes from your ads and your funnel, nothing else. By funnel we mean everything from the Meta ad to the website, including any page in between.
What changed our numbers: sending traffic first to an advertorial (a listicle or a blog-style page) that is coherent with the ad and explains what the product does. We split-tested it by market, with separate SKUs, and the return rate dropped. The customer arrives on the product page better educated and more convinced, which brings us back to emotional closing.
The second front-end lever is ad comments, and few people have a real process for it. Picture a customer who has just ordered and has not received the parcel yet. They see your ad again, open the comments, read negative ones, and refuse at the door. Our process:
- A virtual assistant checks the comments on every ad, in every market, and deletes the bad ones.
- On the Facebook page, we set keyword filters so certain words are hidden automatically.
How we structure the ad accounts themselves is in our COD media buying structure.
Step 3 — Which courier and warehouse should you use?
The courier is the heaviest back-end factor we have found. We tested a lot of delivery companies and the differences are huge. In Greece, we lowered the return rate just by changing courier.
Our old mistake: always choosing the cheapest courier in each market. A courier 1.5 euros cheaper per parcel ends up far more expensive if it adds 5 points of returns, because that is 5 points of margin. We compare them in COD couriers in Europe.
Then the warehouse: always ship from one close to the markets you sell in. If you can deliver in 24 to 48 hours, you increase the chance the parcel is accepted.
Step 4 — How do we check every order before it ships?
If nobody checks your addresses before shipping, you are losing a lot of money: an order without a house number means a courier who cannot find the place, and a trip you pay twice.
What every order needs before it leaves:
- Address: complete, with a house number and enough detail for the courier.
- Phone: a number the courier can actually call.
- Duplicates: the same customer ordering the same thing twice.
- A message to the customer: right after the order, to confirm it, and before delivery, so they know the parcel is coming and are home.
We started with two to three virtual assistants checking every order by hand. At 300 to 400 orders a day, it stopped working: mistakes, missed orders, no way to scale.
Step 5 — How do we reduce the returns we can't fully control?
Some returns are structural, and the data we see confirms it:
- The courier marks the parcel as returned without really trying. We see it a lot in Spain and Portugal.
- The customer is not home, because nobody told them the parcel was arriving that day. Messages before delivery reduce this.
- The customer does not have the exact amount, and the courier almost never carries change. Round your COD prices, especially when you sell in several currencies.
How do you measure your return rate per market?
Never look at one global number: each market has its own courier, prices and customers.
| What to track | How | Why |
|---|---|---|
| Return rate per market | Returned parcels divided by parcels that reached an outcome (delivered or returned) | Parcels still in transit would flatter the rate |
| Before and after each change | Same market, same product, one lever changed at a time | Otherwise you can't tell what worked |
| Split tests | Separate SKUs per variant (price, funnel) in the same market | That is how we measured the price and advertorial effects |
| Margin, not orders | Delivered and paid revenue minus ad spend, shipping out and shipping back | A big order day with high returns can lose money |
If every lever moves by about 5%, a realistic target is a return rate 5 to 10 points lower in each market. Not zero, and not from one change.
Common mistakes
- Looking for the one factor that will cut returns by 10 points. It does not exist.
- Choosing the cheapest courier in every market by default.
- Selling compulsive products, or pricing too high for the strength of your marketing.
- Leaving negative comments under ads that are still running.
- Shipping without checking addresses, phones and duplicates, or without messaging the customer.
Your next steps
- Pull your return rate per market for the last full month and compare it with your margin.
- Pick your worst market and audit its courier: is it there because it is the cheapest?
- Put a daily comment-moderation routine in place on every running ad, plus keyword filters on your page.
- Test an advertorial between your ad and your product page on one market, with a separate SKU.
- Make sure every order is checked (address, phone, duplicate) and every customer gets a message before the parcel ships.
We break this down on camera on our YouTube channel: How to Reduce Your Return Rate in Cash on Delivery (Full Breakdown).
COD Confirm
We built COD Confirm because checking orders by hand with virtual assistants stopped working at 300 to 400 orders a day. Every address gets a score out of 100, every customer receives an SMS in the language of the market, customers whose address needs a check get a link to confirm or correct it, and our call center calls those who don't answer, up to 3 times. Duplicate orders are held for a check before anything ships, which leaves us free to work on the front end. See pricing