Why COD Parcels Come Back, and What Actually Cuts Returns
Three out of four online orders in Bangladesh are paid in cash at the door. PCMI's 2024 market data puts cash on delivery at 75% of e-commerce payment volume, and DHL reports that more than 90% of Bangladeshi buyers still prefer it. A cash order is not a payment. It is a promise, and promises get broken at industrial scale.
Pathao Courier, which moves over 250,000 parcels a day, publishes the honest numbers: a typical merchant delivers only 75 to 85 percent of what they ship. Their own worked example prices the damage. A failed parcel costs roughly Tk 195 in forward and return charges, and a store doing 100 orders a day at a 20% return rate burns about Tk 117,000 a month on parcels that come back.
The returns are not a logistics problem
The most useful finding in the regional data is about intent, not couriers. GoKwik, which scores COD risk across 180 million Indian shoppers, attributes 60 to 70 percent of returns to low buying intent, and only 20 to 25 percent to genuine delivery failures. The midnight impulse order, the customer who ordered from three shops at once, the phone that never picks up: these are conversations that ended too early, not parcels that got lost.
The audited gap is stark. Meesho's IPO filings show 75.5% delivery success on COD orders against 97.8% on prepaid. Same products, same couriers, same customers. The 22-point difference is what an unconfirmed promise costs.
What measurably moves the number
The published intervention data forms a clear hierarchy. Automated pre-dispatch confirmation calls alone cut returns 20 to 30 percent, the strongest single intervention on record. A partial advance through mobile money cuts them by up to 55 percent, because a customer with Tk 100 already committed answers the rider's call. Bangladeshi sellers know this instinctively, which is why the confirm-everything phone ritual is near universal in f-commerce.
The ritual just does not scale. At fifty orders a day it needs a dedicated caller; at two hundred it needs a shift plan. Most stores quietly stop calling, and the return rate climbs back toward the industry band.
Automating the ritual instead of abandoning it
This is precisely the loop Omnistra runs. Every incoming order is scored against delivery history and order shape, risky ones go into a confirmation queue, and an Omnistra voice agent calls within minutes of checkout, in the customer's own language: confirms intent, reads the address back, and requests a bKash advance where your policy calls for it. Clean orders ship without friction. The economics of the confirmation call stop being the reason it does not happen.
One honest caveat: Bangladesh publishes no official national RTO statistic, so the 20 to 30 percent figure you will hear quoted is a practitioner estimate triangulated from Pathao's success-rate band and India's audited numbers. The direction is beyond doubt. The exact number for your store is something you should measure, and cut.
