Six COD rules that actually reduce RTO

Rules are the bluntest instrument in COD, which is why they should be introduced one at a time and measured individually. Switch on five at once and you will have no idea which one moved your numbers, or which one is quietly costing you good orders.

Here is the order we would introduce them in.

1. Verify every COD order

Not a rule so much as the floor everything else stands on. Before you start deciding which orders to refuse, stop the ones that were never real. Do this first, run it for a full thirty days, and only then look at the residue.

How to tell it worked: COD order count falls, delivered order count holds or rises, RTO rate drops within one shipping cycle.

2. Cap COD above a cart value

The single highest-return rule for most stores. Failure rate tends to rise with cart value while the cost of each failure rises proportionally — the two effects compound.

Find the threshold from your own data: bucket the last ninety days of COD orders by value and look for where the RTO rate turns upward. It is usually sharper than merchants expect. Set the cap just below that inflection, and offer prepaid or a deposit above it rather than refusing the sale.

How to tell it worked: RTO rate in the capped band collapses; total orders in that band fall somewhat; contribution from the band rises.

3. Block your worst pincodes

Every catalogue has a short tail of postal codes with failure rates several times the mean. They are usually a mix of genuinely difficult delivery geography and specific local patterns.

Two cautions. Require volume before you judge — three failures out of four orders is noise, not a pattern; wait for twenty-five or more orders before blocking. And revisit the list quarterly, because courier coverage changes and a pincode that failed last year may be fine now.

How to tell it worked: overall RTO falls by more than the revenue you gave up in those pincodes.

4. Exclude the categories that fail

Fit-dependent apparel, fragile goods, and anything with a high subjective-satisfaction component fail at rates that have nothing to do with the customer's intent. If one collection carries an RTO rate twice your average, exclude it from COD before you start blocking people.

How to tell it worked: the excluded collection's contribution per unit rises, even at lower unit volume.

5. Add a COD handling fee

A small fee — ₹30 to ₹50 — does two things. It recovers part of the expected cost, and more importantly it makes prepaid visibly cheaper at the moment of choice without you having to discount anything.

Surface it as a clearly labelled, separate charge. Never bury it in the product price, and never remove the cheapest delivery option to enforce it: buyers must always keep the cheapest choice available to them.

How to tell it worked: prepaid share rises within two weeks; COD order count barely moves.

6. Offer a deposit instead of a refusal

The last rule to add, and the one that recovers orders the other five gave away. Wherever a cap or exclusion would refuse a COD order outright, offer to take 25–30% upfront instead. You keep the sale, and the customer's own money now argues for accepting the parcel.

How to tell it worked: recovered orders in the capped bands, at an RTO rate close to prepaid.

Reading the results

Do not evaluate any of these in under thirty days, and always compare against the previous thirty rather than against a vague memory of "before." COD behaviour moves with festivals, salary cycles and weather. A rolling window-over-window comparison is the only honest read, which is why the dashboard defaults to exactly that.