Pincode-level COD policy
RTO varies from 18% to 35% by city, and some pincodes return most of what you ship them. When restricting COD by geography makes sense, and what it costs.
RTO is not evenly spread across the map. Shipway's FY25 data puts the lowest-RTO city (Vadodara) at 18% and the highest (Patna) at 35%. GoKwik's network data goes sharper: it has reported state-level RTO touching 34% for Bihar, and specific pincodes, including clusters in Assam, Rajasthan and Bihar, where around 72% of COD orders came back.
A brand shipping COD nationally at a flat policy is averaging across places with completely different economics. In some pincodes, the expected cost of a COD order exceeds its margin before it ships.
The arithmetic is simple. At a 35% RTO rate with a ₹200 all-in cost per failed order, every hundred COD orders shipped to that geography carries about ₹7,000 of expected failure cost. If the contribution margin per order is below ₹70, the brand loses money on the average order there, before counting any other cost. Above that margin, the orders are profitable but fragile.
Faced with that, brands have three options short of doing nothing: require prepaid in the worst pincodes, add verification steps only there, or route those pincodes to the courier with the best local delivery record. Most checkout and shipping tools support pincode-level rules for exactly this reason.
Three cautions before you start cutting pincodes.
You are cutting real demand along with the risk. Tier 2 and 3 cities now account for the majority of Indian ecommerce order volume, and COD is their default payment mode. A prepaid-only policy in tier 3 does not convert those customers; it hands them to a competitor who will take their COD order.
Geography is a proxy, and proxies punish the innocent. A 35% RTO pincode still contains 65 genuine customers in every hundred. Blunt restrictions tax good customers for their neighbours' behaviour. Person-level signals (order history, confirmation response, past refusals) are fairer and usually more precise than geography alone.
The worst numbers are partly self-inflicted. Slow lanes, weak local courier franchises and informal addressing all concentrate in the same places that show high RTO. Some of a pincode's failure rate is your logistics, not its customers, and a courier switch can change the number without touching demand.
Tier the response. Green pincodes (your own data, rolling 90 days): COD open. Amber: COD with confirmation required. Red: prepaid only, or COD above a verification bar. Review monthly, because pincodes move between tiers as your courier mix and verification improve.
And measure the cost of each restriction: orders lost in red pincodes are real revenue forgone, so the policy needs the same honest accounting as the RTO it prevents.