Somewhere in your first few hundred orders, the COD question stops being theoretical. The dashboard says sales are up. The bank account disagrees, because half the money is riding around India in courier bags. The RTO report is ugly, your agency swears switching COD off will kill conversion, and the remittance lands whenever the courier's calendar says so. You have three options: enable COD everywhere, kill it completely, or run it as a managed mix with rules. Most founders pick one of the first two, on emotion, after one bad month.
This guide is the third option, with numbers. What a COD order actually costs next to a prepaid order for the same product. When COD earns its keep and when it quietly eats the margin. Which prepaid conversion tactics genuinely work, with realistic expectations attached. And what a healthy mix looks like at each stage of the business. Ravikant Tyagi has run this trade-off from the operations chair, as Distribution Head at Eureka Forbes and Supply Chain and Operations Leader at Atomberg, where payment mix and cash cycles decided how much stock the business could afford next month.
One boundary before we start. This is not an RTO reduction playbook; that lives in our complete guide to reducing RTO in COD orders, which covers verification, NDR management and courier allocation in depth. This guide sits one level above: how much COD you should be running at all, and how to engineer the mix.
COD is not a payment method. It is a customer-acquisition lever with a logistics bill attached. On a ₹999 order, a COD sale costs roughly ₹70 more than the same prepaid sale once you count the collection fee, the RTO probability and a 10 to 15 day wait for your own money. That penalty is worth paying only where COD brings genuinely incremental orders and the margin absorbs it. Under ₹500, run UPI-first with COD behind verification and a token advance. Above ₹1,000, run managed COD with partial advances and pin-code rules. Healthy blended COD share: 55 to 70 percent at launch, 45 to 55 percent while growing, 30 to 45 percent at scale. Move the mix with sized incentives and surgical rules. Killing COD outright usually costs more than it saves.
What a COD order really costs vs prepaid: the full table
Quick definitions so the table reads cleanly. A prepaid order is paid at checkout by UPI, card or netbanking, and the gateway settles the money to your bank in one or two working days. A COD (cash on delivery) order is paid at the door: the courier collects the cash and the shipping aggregator later transfers it to you. Remittance is that batched payout, and the standard cycle runs D+7 or D+8, meaning seven to eight days after delivery, not after the order. RTO (return to origin) is a shipped order that is refused or undeliverable and travels back at your cost.
Here is the honest comparison on a ₹999 order. Assumptions: gateway fee of 2 percent plus GST (Razorpay's standard rate; our Razorpay vs Cashfree breakdown has the exact slabs), COD collection fee of ₹45 (couriers charge a flat ₹40 to ₹50 or 1.5 to 2 percent of order value, whichever is higher), forward shipping ₹80, reverse shipping ₹70, packaging and repack labour ₹40 written off per return. Prepaid RTO at 3 percent, COD RTO at 25 percent, the middle of the 20 to 30 percent band Shiprocket reports for COD orders.
| Cost head | Prepaid order (₹999) | COD order (₹999) |
|---|---|---|
| Payment or collection fee | ₹24 (2% + GST) | ₹45 (flat fee or percentage, whichever higher) |
| RTO probability | 2-5% | 20-30% |
| Cash burned per RTO (two-way freight + packaging + repack) | ₹190 | ₹190 |
| Expected RTO provision per shipped order | ₹6 (at 3%) | ₹48 (at 25%) |
| Verification and NDR follow-up | Nil | ₹5-10 in tools and time |
| Order-to-bank time | 1-2 days after payment | 10-15 days after the order |
| Working-capital cost of the wait (at 18% p.a.) | About ₹1 | About ₹5 |
| Extra cost vs prepaid, per shipped order | Baseline | ₹70-75 |
Two notes on that table. The ₹190 burn per RTO excludes the ad money that acquired the order, which is gone either way; our RTO playbook prices the full burn at ₹340 once CAC is counted. And the bottom row is mostly fixed in rupees, which is why the same penalty that a ₹1,499 product shrugs off will bury a ₹399 one.
Read the ₹70 against your margin, not your revenue. A ₹999 product carrying ₹350 of contribution before payment costs loses a fifth of its profit to the COD penalty alone. According to the Margin Waterfall™ framework, every deduction layer gets priced per order before you decide where orders may come from. COD does not add a new layer to the waterfall; it widens two existing ones, collection fee and RTO provision, then delays whatever survives by two weeks. If you have not built this sheet for your product yet, start with our guide to D2C unit economics in India.
Margin Waterfall™: selling price minus COGS, packaging, shipping, payment or collection fees, RTO loss, then CAC. Run it twice, once per payment method, and the COD question answers itself in rupees: the gap between the two bottom lines is the most you should ever spend converting a buyer to prepaid.
Is COD really 65 percent of Indian ecommerce?
You will hear it in every logistics sales call: COD is 60 to 65 percent of Indian ecommerce. The number is true in aggregate and misleading for you specifically.
Current network data tells a sharper story. Across Shiprocket's network, roughly 40 percent of orders ride COD and 60 percent are prepaid. During Diwali 2025, GoKwik reported prepaid crossing half of D2C orders for the first time: 52.4 percent prepaid against 47.6 percent COD, up from 37.4 percent prepaid a year earlier, almost entirely on UPI. And in the other direction, GoKwik's COD data shows cash-led orders climbing toward 90 percent as you move past Tier 3 towns.
All three numbers are correct. They describe different Indias. A metro skincare brand with a ₹1,400 AOV and Instagram traffic lives in the thirties. A ₹499 fashion brand running Meta ads into Tier 3 pin codes lives in the seventies. The takeaway that matters: your COD share is not a market constant you inherit. It is an output of your category, your AOV, your traffic geography and your checkout design. Which means it is a number you can engineer. The rest of this guide is the engineering.
COD or prepaid by AOV and category
The decision changes shape with order value, because the COD penalty is roughly fixed in rupees while your margin scales with AOV.
Under ₹500, the math is brutal. An impulse product at ₹399 might carry ₹120 of contribution per order; the ₹70 penalty takes more than half of it, and a single RTO erases the profit of four or five delivered orders. Yet this price band sells to the most COD-loyal audience in the country. The answer is not killing COD. It is making COD prove intent: OTP verification on every order, a ₹50 UPI token adjusted against the bill, and a checkout built to make prepaid the path of least resistance.
Above ₹1,000, the purchase is considered, not impulsive. The buyer compared options, read reviews and chose deliberately, so refusal rates drop and the penalty shrinks to 5 percent of a ₹1,400 order. Margin can carry that. The risk shifts to concentration: one refused ₹2,500 parcel locks more capital than five refused ₹399 ones. Manage it with partial advances on high-value orders and stricter rules for first-time buyers.
| AOV band | Typical categories | COD stance | Guardrails |
|---|---|---|---|
| Under ₹500 | Snacks, accessories, trial-size skincare | UPI-first, COD restricted | OTP on every COD order, ₹50 token advance, COD off in red pin codes |
| ₹500-1,500 | Skincare, casual fashion, home basics | COD on, actively managed | WhatsApp confirmation, flat ₹50 prepaid nudge, pin-code rules |
| ₹1,500-3,000 | Footwear, ethnic wear, small appliances | COD selective | ₹100-200 partial advance, prepaid-only above ₹2,500 for first-time buyers |
| Above ₹3,000 | Jewellery, premium electronics, furniture | Mostly prepaid | COD for repeat customers only; offer EMI and BNPL instead |
Layer category on top of AOV. Fashion and footwear sit at the top of every RTO table because sizing doubt survives until the parcel is opened; food and consumables sit near the bottom. If you sell in a high-RTO category, run one band stricter than your AOV suggests. Personalized and fragile products should skip COD entirely; the reverse leg destroys them and there is nothing left to restock.
If AOV is under ₹500 → UPI-first checkout, COD behind OTP plus a ₹50 token, prepaid discount capped at ₹30. If AOV is ₹500 to ₹1,500 → COD on with WhatsApp verification, pin-code rules and a flat ₹50 prepaid nudge. If AOV is above ₹1,500 → partial advance of ₹100 to ₹200 on every COD order, prepaid-only above ₹2,500 for first-time buyers. If blended RTO stays above 30 percent after verification and rules → restrict COD to repeat customers and green pin codes only. If COD share is already under 30 percent → stop squeezing; further restriction refuses more good orders than bad ones.
Prepaid conversion tactics that actually work
Every tactic below works. None of them doubles your prepaid share in a week. Budget a quarter of testing to move the mix 10 to 20 percentage points, and hold each tactic against one number: its cost must stay below your per-order COD penalty.
A prepaid discount sized to the penalty
Offer flat ₹50 or 5 percent off for paying online. The sizing rule is mechanical: if COD costs you about ₹70 extra per order, anything up to ₹60 spent moving that order to prepaid is profit. On sub-₹1,000 AOVs a flat ₹50 usually beats a percentage because it reads bigger at checkout. Expect 8 to 15 percentage points of mix shift over a couple of months, and watch total prepaid orders rather than coupon redemptions, so you are not subsidizing buyers who would have prepaid anyway.
UPI-first checkout
Put UPI intent at the top of the payment page, cards next, COD last behind one more tap. The macro trend does half the work here: UPI is what pushed prepaid past half of GoKwik's Diwali volume, and marketplace-trained buyers already trust it. The mirror tactic, a visible ₹40 COD handling fee, has the same economics as a discount but harsher conversion optics; test it on a slice of traffic before rolling it out.
Partial COD advance
The sharpest middle path. Collect ₹50 to ₹100 by UPI at checkout with the balance payable at the door. The buyer keeps doorstep safety; you get a working payment handle and real commitment, and refusals collapse on token-paid orders. A useful side effect: with the UPI screen already open, a slice of buyers simply pays the full amount. Run it on orders above your median AOV first.
COD verification
OTP or WhatsApp confirmation before dispatch converts nobody to prepaid directly. What it does is remove the fake and half-hearted COD orders that inflate your RTO provision, which shrinks the penalty and buys your other tactics time to work. The full setup, including NDR discipline, is in the RTO playbook linked above.
Post-order payment links
Between order and dispatch, send a WhatsApp payment link: pay now, skip the doorstep cash hunt, occasionally sweetened with priority dispatch. Converting even mid single digits of COD orders is worth the automation, because every conversion refunds the entire penalty on that order.
Where to disable COD: pin-code and order-value rules
Blanket COD decisions are lazy; surgical ones protect conversion and margin at the same time. Restrict or disable COD in exactly four places:
- Red pin codes. After 60 to 90 days of orders, compute RTO per pin code and cut COD wherever it runs above 30 percent. Courier risk engines like Delhivery's RTO Predictor can pre-screen orders while your own ledger matures, and your first-party data wins once it exists.
- First-time high-value orders. A first order above ₹2,000 to ₹2,500 on full COD is your riskiest single parcel. Require a partial advance or prepaid. Repeat customers with clean delivery history keep COD at any value.
- Repeat refusers. Two RTOs against one phone number is a policy, not a coincidence. Block the number for COD; prepaid stays open.
- Fragile, bulky and personalized products. Punishing reverse freight, zero restock value. These never ride COD.
Everything else stays on. The point of rules is to keep COD alive for the large majority of COD buyers who genuinely accept their parcels, while the minority stops billing you for two-way tourism.
The cash-flow math of COD at scale
Margin decides whether you are profitable. Cash decides whether you are alive in the meantime. Walk through a brand doing ₹15 lakh a month: 1,500 orders at ₹999 AOV, 60 percent COD.
The prepaid ₹6 lakh behaves like a normal business: gateways settle T+1 or T+2 and the cash lands within days. The COD ₹9 lakh takes the scenic route. Delivery takes 3 to 7 days. The courier then holds the collected cash, and your aggregator remits on a D+7 or D+8 cycle in two or three weekly batches. Faster remittance exists, but it is sold as a paid add-on, which tells you what the default is: Shiprocket's Early COD sells 2-day remittance, and Delhivery advertises 48-hour COD transfers for its direct clients. Net effect, order-to-bank on COD lands at 10 to 15 days.
Now the float. ₹9 lakh of monthly COD is ₹30,000 a day. With a 12-day pipeline, about ₹3.6 lakh of your money is permanently in transit. Not lost, just never available: a rolling interest-free loan to your logistics chain. On top of that, a quarter of those COD shipments never convert at all, so roughly ₹2.25 lakh of dashboard sales comes back as stock plus two-way freight invoices.
The part founders miss is that float scales with growth. Add 20 percent more COD volume next month and roughly another ₹70,000 of working capital leaves the building, exactly when the bigger inventory PO is due. This is how a brand doing ₹15 lakh a month with healthy paper margins misses a supplier payment. If growth is inventory-constrained, price early-COD fees against the gross margin of the stock you cannot buy while you wait; the fee often wins. Remittance speed also varies by aggregator, one of the deciding factors in our Shiprocket vs NimbusPost vs Delhivery comparison.
I learned to read businesses through cash conversion cycles long before I saw a D2C dashboard. At Eureka Forbes and Atomberg, two channels could show identical margin on paper and one would still starve you, because its cash came back twenty days later. COD is that channel inside your own store. When I look at a D2C P&L now, my first question is not ROAS. It is order-to-bank time by payment method. Founders who know that number to the day are running the business. Founders who don't are being run by their courier's remittance calendar.
What a healthy COD share looks like by stage
There is no universal right number. There are right numbers by stage.
Launch, the first 90 days: 55 to 70 percent COD is normal. A new brand has no trust, and COD is how strangers try you, especially outside metros. Keep it on, verify every order from day one, and treat the phase as paid data collection: you are buying the pin-code ledger and refusal patterns your later rules depend on. Kill COD here and you also strangle the order volume your ad account needs to learn. Just make sure the penalty is priced into the product from day one; our guide on how to price a product in India shows where it sits in the build-up.
Growing, ₹1 to ₹5 lakh a month: manage toward 45 to 55 percent. This is where the five conversion tactics run, and where repeat purchases quietly help. A customer who received one COD parcel from you prepays the second order far more readily; deliver well once and the mix shifts on its own. Hold blended RTO under 20 percent while the share moves.
Scale, ₹5 lakh a month and beyond: work toward 30 to 45 percent blended. At this volume the float and the RTO provision are each lakhs per month, so payment-mix work pays like a good hire. COD stops being a default and becomes a targeted tool: on for new customers in proven pin codes, advance-backed at high values, off where the ledger says off. The trap sits at the far end. Push blended COD below 25 to 30 percent in a Tier 2-3 heavy category and you are usually refusing more good orders than bad ones; the milestone plan in our roadmap to ₹5 lakh a month treats payment mix as a lever, not a trophy.
A fashion founder doing 60 orders a day sees one ugly remittance month, RTO at 32 percent, and kills COD overnight. Orders drop to about 38 a day within the week, because half his buyers sit in Tier 2 and 3 pin codes where COD is how strangers trust a new brand. Meta, fed less conversion volume, optimizes worse and CAC creeps up. He trades roughly ₹6 lakh of monthly revenue to avoid perhaps ₹1.2 lakh of RTO losses, panics, and switches COD back on with no verification, no token, no pin-code rules, inheriting the old RTO rate plus a crater in that month's cash. The failure was never COD. It was COD without guardrails, and the kill switch was the one control that cost more than the problem.
Your COD mix checklist
- Compute COD share of orders and revenue for the last 60 days, plus RTO split by payment method.
- Price your per-order COD penalty: fee delta, plus RTO-probability delta times two-way freight, plus float cost.
- Reorder checkout: UPI on top, COD last, delivery dates stated honestly on the payment page.
- Launch one prepaid incentive sized below the penalty, flat ₹50 or 5 percent, and measure mix shift for two weeks.
- Turn on COD verification, and a ₹50-100 partial advance for orders above your median AOV.
- Cap first-time buyers: partial advance or prepaid-only above ₹2,000-2,500.
- Build the pin-code ledger and cut COD where 60-90 day RTO exceeds 30 percent.
- Map your remittance calendar, compute float days, and price early COD against your next inventory PO.
- Review monthly: COD share, RTO by method, penalty per order, float days.
Your next action today
Pull the last 60 days of orders and compute three numbers: COD share of orders, RTO rate split by payment method, and your per-order COD penalty using the table above with your own freight and fees. Those three numbers place you on this guide's map and point to the one guardrail to ship this week: verification if you have none, a token advance if you sell above ₹1,000, a sized prepaid discount if the penalty embarrassed you. One guardrail this week beats a payments strategy document next month.
If you'd like the complete execution system, calculators, SOPs, templates and operating frameworks behind this process, continue inside D2C Acquisition.Lab.
