Cut RTO on COD Orders: A Playbook for India

Cash on delivery still drives a large share of orders across Indian ecommerce, and it will for years. The problem is not COD itself. The problem is return to origin, the order that gets packed, shipped, attempted, refused, and sent all the way back. You pay forward and reverse freight, you lose the sale, and the product comes home tired. RTO is one of the few costs that scales with growth, so it deserves the same attention you give ad spend.

Most brands treat RTO as a courier problem. It is really an operations and data problem that shows up at the courier. The good news is that the biggest levers sit inside your own funnel and warehouse, not with the carrier.

Score the order before you ship it

Every order carries signals about how likely it is to come back. Treat those signals as a risk score, not a gut feel. The inputs that matter most in India are consistent across categories.

• Address completeness. Missing landmark, no building number, or a pin code that does not match the city is the single largest predictor of a failed delivery.
• Order value. Very high value COD baskets refuse more often, especially in tier two and tier three towns where the customer wants to inspect before paying.
• Repeat behaviour. A phone number or address with prior RTO history is a clear flag. Keep a suppression list and update it weekly.
• Buying pattern. Multiple units of the same size, or a rushed checkout at odd hours, correlate with impulse orders that get abandoned.

You do not need a data science team to start. A simple rule set that flags orders above a value threshold, in a watched pin code, or with an incomplete address will catch most of your risk. Refine it as you gather your own numbers.

Verify the risky ones, leave the rest alone

Once an order is flagged, verify intent before it enters the packing queue. The mistake brands make is verifying everyone, which adds friction and annoys good customers. Verify only the flagged share, usually ten to twenty percent of COD orders.

Verification can be an automated call, a WhatsApp confirmation, or an SMS with a one-tap confirm link. The message should be short and respectful. You are confirming the address and the intent to pay, nothing more. Orders that confirm move ahead. Orders that go silent after two attempts get held or cancelled before they cost you freight in both directions.

The economics are simple. Reverse freight, forward freight, and handling on a returned order often cost more than the profit on the order itself. Spending a few rupees to verify a risky order is cheap insurance against that loss.

Fix address quality at the source

A large slice of RTO is caused by addresses that were never deliverable. You can attack this at checkout and after it.

• At checkout, validate the pin code against the city and state, make the phone field mandatory and format-checked, and ask for a landmark on high-value orders. Small prompts here prevent large costs later.
• After checkout, run flagged addresses through a normalisation step so that the courier receives a clean, complete address. Couriers cannot deliver what they cannot find.

Share your worst pin codes with your operations team every month. Patterns repeat. The same clusters of undeliverable addresses tend to reappear, and once you know them you can route them to your most reliable carrier or add verification automatically.

Nudge the highest-risk baskets to prepaid

The cleanest way to remove RTO risk is to remove COD from the orders most likely to bounce. You do this with a carrot, not a stick. Offer a small discount, free shipping, or a coupon for choosing prepaid, and surface that offer only on flagged orders. A customer who would have refused a COD parcel often pays upfront when the saving is visible.

Keep COD available everywhere else. Removing it broadly will cost you more in lost orders than you save in returns. The goal is surgical, not sweeping.

Recover the near-misses

Not every RTO is a lost cause. When a first delivery attempt fails, the order is not dead yet. Set up a fast recovery loop. As soon as a failed attempt is logged, reach the customer, confirm they still want the order, and ask the courier to reattempt. A same-day response recovers a real share of parcels that would otherwise complete their journey back to your warehouse.

Track this as its own metric. Recovered orders per hundred failed attempts tells you whether your customer support and courier coordination are actually working.

Measure what matters

You cannot manage RTO without a weekly view. Track RTO rate by category, by pin code cluster, by courier, and by order value band. Watch net delivered revenue, not gross orders, because a shipped order that comes back is worse than an order you never took. When you see a spike, trace it to one of these dimensions and act on that specific cause. RTO rarely has one villain. It is a hundred small leaks, and this playbook is how you close them one band at a time.