COD benchmarks: what to measure and what 'good' looks like

Benchmarks are useful for framing questions and dangerous for setting targets. A fashion store shipping ₹900 carts to tier-3 cities and an electronics store shipping ₹9,000 carts to metros do not share a meaningful RTO number, and comparing yourself to a published average will mostly tell you which category you are in.

What is worth borrowing is the definitions. Measured consistently, your own trend is the only benchmark that can actually be acted on.

Define these six precisely

COD share. COD orders divided by total orders. Segment by value band — a blended figure hides the fact that failure concentrates at the top.

COD RTO rate. Orders returned to origin divided by orders shipped COD, not orders placed. Cancelled-before-dispatch orders belong in a different metric; mixing them in makes your rules look better than they are.

Prepaid share. The one number that moves everything else. Track it weekly, and always alongside the discount cost that bought it.

Verification completion. OTPs verified over OTPs sent. This is an operational health metric, not a customer-quality metric — treat a drop as a system problem until proven otherwise.

Cost per failed order. Fully loaded: both freight legs, packaging, handling on both ends, and an amortised write-down for stock that cannot be resold as A-grade. Calculate it once a quarter.

Contribution per hundred sessions. The one that nets everything off. Discounts, fees, returns and recovered orders all land here, which makes it the only figure that can tell you whether a change was actually worth making.

Compare windows, not memories

Thirty days against the previous thirty. Always. COD behaviour swings with festivals, salary cycles, weather and your own promotional calendar, and a week-on-week read will have you chasing noise.

Two practical rules. Change one thing at a time, and give it a full window before you judge it. And when a number moves, check the operational metrics before the customer-behaviour ones — a fall in verification completion is far more often a broken template than a change in who is shopping with you.

What "good" actually means

A merchant whose COD RTO rate fell from 30% to 22% over a quarter is doing better than one sitting flat at 18%. Direction beats level, because direction is the part you control.