Fewer Failed Deliveries: A Founder's System for Cutting COD Returns
Every parcel that comes back costs twice. A simple system across order, confirmation, packing and courier keeps cash on delivery returns under control.

Every online seller in Bangladesh knows this moment. The courier report arrives, and a long list of parcels is marked "returned". The customer did not pick up the phone, changed their mind at the door, or said they never ordered at all.
Each of those parcels has already cost you. You paid for the product, the packing, the delivery charge and often the return charge. Your team spent time on the order. The stock sat in a courier hub for days instead of on your shelf. And the sale you counted on the dashboard never became money.
Most founders treat returns as weather — something that happens to the business. In my experience, they are closer to a leak in a pipe. You cannot stop every drop, but you can find where the water escapes and fix it. This article is the system I use to think about it.
Why returns hurt more than they look
A return rate is easy to read as a small number on a report. The real cost is larger, for three reasons.
First, a returned parcel costs you twice: once to send and once to bring back, plus the packing you cannot reuse and the product that sometimes comes back damaged.
Second, returns hide inside your sales figures. If the team is praised for orders placed, nobody feels the pain of orders that fail. The page looks busy, the ad reports look good, and the bank account tells a different story at the end of the month.
Third, returns damage your relationship with couriers. A seller with a high failure rate becomes a costly client, and that shows up over time in service, attention and terms.
So the first step is simple: measure delivered orders, not placed orders. Put the delivered number at the top of every sales report. When the team sees the gap between the two every day, the conversation changes on its own.
Where returns really come from
When I sit with a team and go through returned orders one by one, the reasons usually fall into a handful of groups. Rarely is it "bad customers". More often it is something we did, or did not do, earlier in the chain.
- The order was never real. Someone clicked, filled a form out of curiosity, or a competitor or prankster placed it. Nobody checked before it was packed.
- The customer was not sure. They ordered on impulse from an ad, then had a day or two to think. By the time the rider called, the feeling had gone.
- The promise did not match the product. The photo, size, colour or quality on the page was better than what arrived. The customer refused at the door, and they were right to.
- Delivery took too long. A parcel that arrives in two days meets a customer who still wants it. One that arrives in seven meets someone who has already bought elsewhere or spent the money.
- The customer could not be reached. Wrong number, wrong address, no landmark, phone switched off, or the rider called once and gave up.
- The money was not ready. Cash on delivery means the customer must have the full amount in hand at that moment. Some do not.
Each of these has a different fix. That is why a single rule — "call every customer" or "take advance payment" — never solves the whole problem. You need a small system that touches each stage.
The system: four checkpoints from order to doorstep
Think of the journey of an order as four checkpoints. At each one, someone owns a simple job.
1. At the order: ask for the right details
Most bad orders can be caught at the form. Ask for a full address with an area and a landmark, a phone number that is checked for the right number of digits, and one confirming choice such as size or colour. Make the price, delivery charge and expected delivery time clear before the customer submits.
Every extra field costs a little conversion. But a clear, honest order form filters out the people who were never going to pay, and that is worth more than a few extra form submissions.
2. At confirmation: call before you pack
This is the single most important checkpoint. No order is packed until a person has confirmed it, ideally by phone, within a few hours of the order being placed.
The call is not a formality. A good confirmation call does four things:
- Confirms the person actually placed the order and still wants it.
- Repeats the product, size, colour and total amount, including delivery charge.
- Checks the address and landmark, and the best time to deliver.
- Tells the customer when to expect the parcel and that the rider will call.
Give your confirmation team a short script and a few clear outcomes to record: confirmed, cancelled, unreachable after two attempts, or needs a call-back. Orders that cannot be confirmed are held, not shipped. It feels painful to hold an order. It is far more painful to pay for its return.
For first-time customers in far-off areas, or for high-value products, consider asking for a small advance or the delivery charge through bKash or Nagad. Explain it politely as a booking confirmation. Many genuine buyers will happily pay it; people who were never serious usually will not.
3. At packing: send exactly what was promised
Returns caused by "this is not what I saw" are the most avoidable of all. Two habits help here.
First, keep the page honest. Real photos, accurate sizes, a clear note where colour may vary on screen. A product that looks slightly less glamorous on the page but matches it in the box will earn you fewer refusals and more repeat buyers.
Second, add a check at packing. Someone other than the packer matches the item against the confirmed order — product, size, colour, quantity — before the parcel is sealed. A wrong size sent to a happy customer is a return you created yourself.
4. At the courier: speed and follow-up
Send confirmed orders out the same day, or the next morning at the latest. Every day an order waits is a day the customer's interest cools.
Then do not let the parcel disappear into the courier's system. Someone on your team should own in-transit follow-up: watching for parcels marked "customer unreachable" or "on hold", calling the customer directly, and arranging a fresh attempt before the courier sends it back. Many parcels that would have returned can be saved with one call from the seller.
Finally, review your courier choice by area. If one courier consistently performs better in a district, route orders there accordingly. Judge couriers on delivered rate and speed, not only on price.
Make one person own the number
A system without an owner slowly falls apart. Choose one person — in a small business, perhaps your operations lead — who owns the delivered rate and reports it every week.
Their weekly report can be very short:
- Orders placed, orders confirmed, orders shipped, orders delivered, orders returned.
- The top reasons for returns this week, from the courier notes and follow-up calls.
- Returns by product, by ad campaign and by area.
- One change the team will make next week.
The breakdown by product and campaign is where the real lessons live. Sometimes one product causes a large share of returns because the page oversells it. Sometimes one ad attracts curious clickers rather than buyers. You will not see that from a single total.
This links directly to how you pay and praise your team. If your sales staff earn incentives on orders taken, they will push for orders taken. Pay incentives on delivered orders and the whole team starts to care about confirmation, honesty and follow-up — because now it is their number too.
What a founder should not do
A few common reactions make the problem worse.
Do not blame the customer in public or argue with them on the page. A small share of people will always abuse cash on delivery; most returns still have a cause inside your process.
Do not chase a lower return rate by cutting out whole regions or customer groups without looking at the data. You may lose good buyers to solve a problem that was really a slow courier or a weak confirmation call.
And do not keep returned stock sitting in a corner. Check it, repack it or mark it damaged, and put it back into sellable stock within days. Returned goods that nobody counts are cash you have quietly written off.
What to do this week
- Add delivered orders to the top of your daily sales report, next to orders placed.
- Pull last month's returned orders and sort them into the reasons above. Find the biggest one.
- Write a one-page confirmation call script and make "no confirmation, no packing" a rule.
- Name one person who owns in-transit follow-up and the weekly delivered-rate report.
- Look at how sales incentives are paid, and plan the move to delivered orders.
Cutting returns is rarely about one clever trick. It is about many small checks done by the right people every day. For a structured review of your order-to-delivery process, see Business Consulting.
If you want a second pair of eyes on this for your business, book a free strategy call.