Short answer: most returns in Bangladeshi f-commerce are decided before the parcel leaves. Confirm every cash-on-delivery order by reply, OTP or call; check the buyer's phone number against courier delivery history; hold stock only for confirmed orders; price so a return doesn't wipe out the margin; and put returned parcels back on sale within a day. Shops that do these consistently usually see their return-to-origin (RTO) rate fall by a third or more. Nothing below needs a developer.
A returned COD parcel costs the outward delivery charge, and with some couriers a return charge too (Pathao's rate page states half the delivery cost outside Dhaka metro; Steadfast's terms say it does not charge for return parcels; both checked 22 October 2026), the packaging, the days the item was off sale, and the chance it comes back damaged. On a ৳1,200 item with ৳350 margin, one return eats the profit of the next two or three sales. Industry commentary puts f-commerce RTO anywhere between 10 % and 30 %, and some vendors quote higher; your own number is the only one that matters, so start by measuring it (step 7).
A chat order is a request, not a sale. Before you pack:
If you use PrimeCommerce, a cash-on-delivery order's units are held while it waits, you mark it confirmed, and an order whose hold runs out before it is sent is cancelled and its units go back on sale.
Couriers know, for every phone number, how many parcels were delivered and how many were refused across all the shops they serve. Sellers in Bangladesh check this in two ways:
A practical rule: under roughly 70 % delivered, or more than two refusals, treat the order as hold and call; don't auto-cancel, because a changed address or a one-off bad experience looks the same as fraud in the numbers.
"Unreachable" and "wrong address" are the biggest failed-delivery reasons outside Dhaka. Ask for: district, upazila/thana, area or road, a landmark, and a second phone number. Couriers like Pathao and RedX need the area selected from their list; a free-text address that doesn't match their coverage sits in a hub for days and then comes back.
If 1 in 5 parcels comes back, your delivery cost per delivered parcel is not ৳100, it's ৳100 + (0.2 × ৳200) ÷ 0.8 = ৳175. Build that into the price or the delivery charge rather than discovering it in the month-end numbers. Charging a small advance (৳100–200 via bKash/Nagad) on high-ticket or out-of-Dhaka orders is common and, in most sellers' experience, cuts refusals sharply, at the cost of some lost orders. Test it on one category for two weeks.
Orders that ship the same or next day return less; by day four the buyer may have bought elsewhere or forgotten. Send the tracking code and the courier's name by Messenger or SMS, and a second message on the day it's out for delivery. A buyer expecting a call from a Steadfast rider answers it.
When the courier reports delivery failed, don't wait for the automatic return. Call the buyer that day: reschedule or confirm the return. One retry with a confirmed time succeeds often enough to be worth it; two retries rarely do.
The hidden RTO cost is stock that sits in a corner. When a parcel comes back: open it, grade it (unopened / opened but fine / damaged), and put the good ones back in inventory the same day. If you track stock in a sheet, add a "returned, resaleable" column; if you use PrimeCommerce, the Returns screen receives the parcel into quarantine with its condition, and what you restock goes back on sale from there.
RTO rate = parcels returned ÷ parcels dispatched, per month, per courier and per district. Keep it on a wall. Most shops that start measuring find one courier or two districts account for most of the problem — a routing change, not a strategy change.
Related: Delivery & charges · Returns & refunds · Payment methods
Last updated 22 October 2026.
There is no official figure. Sellers and logistics commentators commonly cite 10–30 % for COD-heavy f-commerce; well-run shops with confirmation and address checks report single digits to low teens. Measure your own.
In most sellers' experience, yes, substantially, because it filters out impulse and fake orders. It also reduces total orders. Try it on expensive items or out-of-Dhaka orders first.
They read your own courier account's data, which is your data to see, but they are not official APIs and can stop working at any time. Treat them as one signal among several, and keep your own records.
It differs by district and by shop. Track your RTO per courier for a month and route by what your numbers say rather than by reputation.
Yes. A shop may decline an order. Be careful with blanket bans: ask once by phone before blocking a number for good.
Sellers on PrimeHat use PrimeCommerce, which holds a cash-on-delivery order's stock while it waits to be confirmed, books the parcel with your courier from the order, and takes returned parcels through quarantine and back into stock from a Returns screen. See Sell on PrimeHat and PrimeCommerce.