How to Reduce RTO for COD Orders

Every return-to-origin parcel costs you forward shipping, return shipping, packaging, and the working capital locked in stock that travelled for nothing. You earn zero revenue on it. For a brand running on Cash on Delivery in India, RTO is usually the single largest avoidable cost line in the business — and most brands attack it with one blunt fix instead of matching each fix to the cause it addresses.

Cash on Delivery accounts for the majority of online orders in India — payment processor Cashfree puts the figure at roughly 60 to 65 percent — and COD orders carry materially higher refusal rates than prepaid. Logistics platform Shipway has reported that around a quarter of Indian COD orders come back as RTO. Whatever your exact number, the mechanism is the same: COD removes financial commitment at the moment of purchase, so the cost of abandoning the order falls entirely on you.

The instinct is to turn COD off. That works, in the sense that amputation cures a broken finger. COD drives most of your order volume; removing it solves the RTO line by destroying the revenue line. The useful question is narrower: which COD orders are worth shipping, and how early can you tell?

The five real causes of COD RTO

Before choosing fixes, separate the causes. Brands that apply a single intervention across all five see partial improvement and conclude RTO is unfixable.

1. Low purchase intent. The customer ordered on impulse — often from a social ad — with no strong commitment, and declines at the door. This is the largest bucket for most D2C brands and the cheapest to fix, because it can be caught before you spend a rupee on shipping.

2. Bad address or phone data. Incomplete addresses, typos, wrong pincodes, and unreachable numbers. The courier cannot deliver, attempts twice, and returns the parcel. Nothing about the customer’s intent was wrong; your data capture was.

3. Delivery delay. Intent decays. A customer who was enthusiastic on day one is indifferent by day nine, and may have bought the same item elsewhere. Slow dispatch converts good orders into RTO.

4. Expectation mismatch. The product arrives and is not what the customer pictured — wrong size, different colour in person, smaller than it looked. They refuse at the door rather than deal with a return. This is a product-page problem masquerading as a logistics problem.

5. Deliberate fraud and abuse. Repeat offenders, fake numbers, competitor mischief, bulk orders placed with no intent to pay. Smaller than most merchants assume, but concentrated — a handful of numbers and pincodes often account for a disproportionate share.

Stage one: fixes at checkout

The cheapest RTO to prevent is the one that never becomes an order in the first place, or that converts to prepaid before it ships.

Make prepaid the more attractive choice. Not by hiding COD, but by pricing the difference honestly. A COD order genuinely costs you more to fulfil — handling fee, collection charges, higher failure risk. Charging a COD fee that reflects that cost, and offering a prepaid discount of similar magnitude, shifts a meaningful slice of your mix without blocking anyone. Customers who are willing to pay upfront are, almost by definition, the ones who intend to receive the parcel.

Validate the address while the customer is still on the page. Pincode validation, phone-number format checks, and prefilled addresses for returning customers remove an entire RTO cause at source. Fixing an address after dispatch is expensive; fixing it during checkout is free.

Apply COD rules rather than a blanket switch. COD availability can be conditional on order value, product category, customer history, or pincode. High-value orders carry more absolute risk, so a partial advance or prepaid-only rule above a threshold is reasonable. Categories with high expectation mismatch — apparel and footwear in particular — justify tighter rules than categories where the product is exactly what it says it is.

Collect a partial advance. Taking a small portion of the order value upfront is the single sharpest intent filter available, because it costs the customer something real while leaving most of the convenience of COD intact. How much to collect is its own question, covered in How Much COD Advance to Collect.

Stage two: fixes before dispatch

Once the order exists, you have a window before you spend money on it. Use it.

Verify the order. An OTP, a confirmation link over WhatsApp or SMS, or an automated call establishes two things at once: the phone number is real and reachable, and the customer is willing to spend ten seconds confirming they want the parcel. Orders that fail verification are the orders you should not ship.

The channel matters more than merchants expect. WhatsApp confirmation generally sees higher response than SMS in India, because SMS is crowded with promotional traffic and often ignored. An automated voice call reaches customers who ignore text entirely. Running one channel and treating non-response as refusal will cost you genuine orders — sequence the channels instead.

Set a confirmation window, not an indefinite wait. Holding orders for verification adds to delivery time, which is itself an RTO cause. A tight window — hours, not days — followed by a clear default action keeps the cure from becoming the disease.

Score and triage rather than verify everything. At low volume you can confirm every COD order. Past a few hundred a day you cannot, and you should not want to: verifying a repeat customer with three successful deliveries is wasted effort and added friction. Route orders by risk — order value, pincode history, customer history, product category — and send only the uncertain ones into the verification queue. Your operations team should be handling exceptions, not every order.

Use your own refusal history. The most predictive dataset you have is your own. Phone numbers that have refused before, pincodes with chronic failure rates, order-value bands where your refusals cluster. A blocklist built from your own data outperforms any generic risk score, and it costs nothing but the discipline to maintain it.

Stage three: fixes in transit

The parcel has shipped. Most brands stop paying attention here, which is why so much recoverable RTO is lost.

Ship fast. Dispatch time is the lever with the best ratio of effort to effect, and most brands are slower than they think. Measure the hours from order placed to courier pickup, not from order placed to label created. Every day of delay widens the gap between the intent that produced the order and the indifference that refuses it.

Tell the customer what is happening. A dispatch notification, an out-for-delivery message with the expected window, and a reachable support channel all raise the odds someone is home and expecting the parcel. A customer who knows ₹1,400 is arriving on Thursday prepares for it; one who has forgotten they ordered does not.

Handle NDRs within hours. A non-delivery report is not a verdict, it is a second chance with an expiry date. Most couriers make two or three attempts. If the first failure is handled the same day — call the customer, reconfirm the address, arrange a window — a large share of those parcels still deliver. Brands that batch NDR handling into a weekly review lose almost all of them. This is the highest-yield operational habit on this list and the one most often left undone.

Choose couriers by lane, not by average rate. Courier performance varies enormously by region in India. A carrier that is excellent in metros may be poor in tier-3 pincodes, and your blended delivery rate hides that. Allocating by lane performance rather than by headline price often beats every other intervention in specific geographies.

The fix nobody classifies as RTO work

Expectation mismatch is an RTO cause that lives entirely on your product page. If a meaningful share of your refusals are door-step rejections of a product the customer did receive and look at, no amount of verification will help — the order was genuine and the intent was real. What failed was the description.

Honest size charts with real measurements, photographs in ordinary light rather than studio light, scale references, and plain statements of material and finish all reduce refusals. So does saying what the product is not. The pattern to look for: refusals concentrated in specific SKUs rather than spread across the catalogue. That is a listing problem, and it is cheap to fix.

How to know whether any of this worked

RTO work goes wrong when it is measured as a single blended percentage, because that number moves for reasons unrelated to your fixes — seasonality, channel mix, a new product launch, one bad courier lane.

Measure instead:

RTO rate split by payment method. Prepaid and COD are different businesses. Blending them hides whether your COD work is landing, and makes a shift in payment mix look like an improvement in order quality.

Prepaid share of orders. If your COD-to-prepaid work is succeeding, this rises. It is a leading indicator; RTO improvement follows it.

Verification response rate by channel. Tells you whether non-response means refusal or means your messages are not being seen.

Hours from order to pickup. The dispatch-speed lever, measured honestly.

NDR recovery rate. Of parcels that failed a first attempt, what share eventually delivered. Most brands have never calculated this and are shocked by how low it is.

RTO by pincode, category, and order-value band. Blended RTO tells you there is a problem. Segmented RTO tells you where it is.

Change one thing at a time, give it enough volume to mean something, and compare segments rather than totals. A fix that cuts RTO four points in one category and does nothing in two others is still a good fix — but only if you can see it.

Where to start

If you are starting from no system at all, the order of operations that tends to produce the fastest return:

First, verify COD orders before dispatch and stop shipping the ones that fail. Second, measure and cut your dispatch time. Third, handle NDRs the same day they appear. Fourth, add COD fees and prepaid incentives to shift the mix. Fifth, segment your data and apply targeted rules where your own numbers say the risk sits.

The first three cost almost nothing and address the largest buckets. The last two are where sustained improvement comes from, and they depend on having the data from the first three.

COD King handles the pre-dispatch layer of this — OTP and WhatsApp verification, COD fees and rules, partial payments, prepaid conversion, and the pincode and order-level data to segment by. The dispatch-speed and NDR work is yours, and no app will do it for you.