How to offer cash on delivery without bleeding cash
Understand what cash on delivery really costs an Indian seller and how to tier it by order value pincode and buyer history instead of all or nothing.
The order you win and the cash you wait for
Cash on delivery is still how a large share of Indian online buyers prefer to pay and switching it off removes orders immediately. The problem is that it moves your risk from the payment gateway to your own balance sheet. You pack the unit, you pay the forward freight and then you wait to find out whether a stranger opens the door.
Two things go wrong. The parcel is refused and comes back as an RTO with both freight legs billed to you. Or it is delivered and the cash reaches you a week or two later than the stock left. Neither shows up as a loss on an order-level report. Both show up in your bank balance.
What COD costs and what it never provides
COD does not get you paid at dispatch and no arrangement changes that. Couriers charge a per-shipment COD handling fee plus a percentage of the collected value and they remit on a cycle rather than on delivery. Treat the gap between dispatch and remittance as working capital you have lent to your own customers.
The RTO cost is the bigger number. A refused parcel bills forward freight, return freight and often the COD fee while producing no revenue. Add the monthly labour of reconciling courier remittance statements against orders. Sequencing matters too. Under GST the liability on a supply is generally tied to the invoice and the supply rather than to the date cash arrives, so tax can fall due on an order whose money is still with the courier. A qualified professional should confirm how this applies to your registration and billing practice.
Who should not offer it at all
If your working capital cannot fund three weeks of dispatched but unpaid orders then COD will damage you faster than losing those orders would. Run the number before you run the offer. Take your expected monthly COD revenue and set aside three weeks of it as money you cannot touch. If that figure breaks your supplier payments then the honest answer is prepaid only.
High-ticket sellers should also stay out. An RTO on a twenty thousand rupee unit wipes the margin on several delivered orders and the fraud exposure is real because nothing is verified before dispatch. Perishable goods, fragile goods and made-to-order items have the same problem in a sharper form. A refused custom-printed unit is not inventory. It is scrap.
This is one case where the cheaper decision is to serve fewer buyers. A seller with thin margins in a high-return category who offers unrestricted COD to compete on checkout convenience is buying volume with cash they do not have.
Tiering instead of switching it off
COD is not a yes or no setting. The useful version is conditional. Set an order value ceiling above which COD disappears and choose that ceiling from your own RTO data rather than from a competitor. Suppress COD for first-time buyers in the pincodes where your refusal rate is worst and keep it for repeat buyers everywhere.
Add verification before dispatch. An OTP or a call confirmation on COD orders above a threshold removes casual orders before you pay freight on them. Validate the phone number and flag incomplete addresses rather than dispatching and hoping. Repeat refusers should be blocked by number.
Then make prepaid the easier path. A small COD handling fee shown honestly at checkout and a matching prepaid discount shifts a share of buyers without insulting anyone. A UPI prompt at the order confirmation step converts some COD orders after the fact. Partial advance on large orders is the middle option where the category allows it.
What to measure and what to distrust
The metric that matters is contribution margin per delivered COD order after the COD fee, the RTO freight allocated across delivered orders and the cost of the cash being late. Break it down by pincode and by order value band. That is the view that tells you where to set the ceiling.
The metric that flatters is COD share of orders and headline conversion rate. Both improve the moment you offer COD to everybody and both are silent about RTO. A seller optimising for conversion alone will happily take a thousand orders that lose money on two hundred refusals.
The mistake that empties the account
The common mistake is reading the courier remittance statement as revenue and buying stock against it. Remittances arrive net of fees and net of nothing else. The RTO freight on parcels that never delivered is billed separately and often later. A seller who orders inventory against gross remittance is financing their own refusals with supplier money.
It is expensive because the error compounds with growth. Double your COD volume and you double the float and the RTO bill at the same time while the revenue line looks excellent. Businesses rarely fail at this quietly. They fail in a festive season when volume and refusals peak together.
Revisit the rules every few months
COD thresholds set two years ago are probably wrong now. UPI adoption keeps moving buyer behaviour and the pincodes that refused parcels last year are not always the worst this year. Treat your ceiling, your blocked pincode list and your prepaid incentive as quarterly settings rather than as policy.
This guide explains how something works. It is not legal, tax or financial advice, and rules change. Confirm anything regulatory with a qualified professional before you act on it.