Skip to main content

UPI MDR Is Delayed Until After Diwali. Don't Let 0.4% Rewrite Your COD Policy.

A balance weighing a prepaid phone payment with a 0.4% fee against a returned cash-on-delivery parcel

India’s announced UPI MDR framework is waiting for a new implementation date. For ecommerce merchants, the useful question is not whether 0.4% sounds expensive. It is whether prepaid payments are actually cheaper than the expected cost of COD failure.

India’s proposed UPI MDR was supposed to start on October 15. NPCI has pushed implementation until after Diwali, with the new date to be notified separately. (business-standard.com)

That pause gives commerce teams time to do something more useful than add a prepaid discount or hide COD: calculate the real marginal cost of each payment route.

A ₹3,000 eligible UPI payment would carry ₹12 in MDR under the announced framework. A failed COD shipment can cost much more—but only after probability, delivery performance and conversion are taken into account.

First, what is actually changing?

The announced framework introduces a 0.4% Merchant Discount Rate on eligible Person-to-Merchant (P2M) UPI transactions above ₹2,000.

The fee is capped at ₹300 per transaction for payments of ₹75,000 and above. Transactions up to ₹2,000 remain outside the new MDR framework. Small merchants under the relevant P2PM classification also retain zero-MDR treatment.

The official FAQ states that more than 95% of P2M UPI transaction volume is below ₹2,000. It also distinguishes direct bank-account-to-merchant UPI payments from credit-linked UPI products. (financialservices.gov.in)

There are two details merchants should not overlook.

First, the original October 15 implementation date has been superseded by NPCI’s October 13 postponement. As of October 31, the new date remains unannounced.

Second, the official FAQ says merchants cannot pass the MDR on to consumers as a UPI surcharge. The customer-facing price is not supposed to change because of this fee. (financialservices.gov.in)

This is not a reason to assume that every UPI transaction will become more expensive. It is a reason to identify which of your own transactions are eligible.

Why ₹12 and a COD return are not comparable costs

A merchant may look at the announced MDR and ask whether it is worth encouraging prepaid payments.

That is the wrong comparison if the merchant is looking only at the percentage fee.

The cost of accepting a prepaid payment includes the applicable MDR, any separately contracted payment gateway charges, and any incentive offered to encourage the buyer to pay online.

The cost of COD includes more than collecting cash. A failed shipment can involve forward freight, reverse freight, handling, customer support and inventory being unavailable while the parcel is in transit.

Amazon Shipping India’s NDR explainer cites a ₹180–₹240 cost for unresolved NDRs that escalate to RTO, based on Pragma Analysis of 142 Indian D2C brands in 2024. That is a published estimate, not a universal cost for every merchant. (shipping.amazon.in)

The distinction matters because an RTO cost is an expected cost. It does not apply to every COD order.

COD/RTO benchmarks are not interchangeable

Unicommerce’s India D2C Report 2026 reports a 58% COD return rate during the festive quarter in its platform data. (unicommerce.com) Shipway’s FY2025 ShipNotes report reported nearly 26% returns on non-prepaid orders across shipments processed through its platform.

These numbers should not be averaged into a national benchmark.

The underlying datasets, time periods, merchant populations and definitions may differ. A merchant should establish its own baseline using its own orders, dispatches and delivery outcomes.

The useful question is not whether your RTO rate is above a vendor’s published benchmark. It is how much each failed order costs your business.

Run the actual merchant math

A payment decision should compare the expected cost of the COD route with the expected cost of the prepaid route.

A simplified model is:

Expected COD cost per order
=
COD collection/remittance fee
+
(RTO probability × RTO cost)

For prepaid payments:

Expected prepaid cost per order
=
Eligible UPI MDR
+
Payment gateway charges
+
Prepaid incentive
+
Expected prepaid RTO cost
+
Expected contribution lost through checkout friction

The exact inputs depend on the merchant’s payment contracts, courier costs, delivery outcomes and checkout behaviour.

The important point is that the MDR is only one component of the prepaid cost.

A ₹3,000 example

For an eligible ₹3,000 UPI payment:

₹3,000 × 0.4% = ₹12

Now compare that ₹12 with an illustrative direct RTO logistics cost of ₹180–₹240.

The break-even RTO probability, if comparing only the MDR with direct RTO logistics cost, is:

₹12 ÷ ₹240 = 5%

to

₹12 ÷ ₹180 = 6.67%

This is a simplified illustration, not a universal merchant threshold.

It says that if the expected direct logistics cost of a COD failure is materially greater than the MDR, prepaid can have an economic advantage before considering other costs.

But add a prepaid discount, gateway fees, the risk of prepaid RTO and lost checkout conversion, and the break-even point changes.

A merchant with a low COD RTO rate and strong COD conversion may find that a blanket prepaid incentive costs more than it saves.

A merchant with a high failure rate in a particular courier or postal segment may reach the opposite conclusion.

The calculation should determine the policy—not the headline fee.

Don’t change the entire checkout based on one percentage

The next mistake is to treat COD as a single risk category.

A COD order from a repeat customer with a successful delivery history is not necessarily equivalent to a first-time order with incomplete address information.

A high-RTO postal area does not automatically mean that every buyer in that area is unreliable. It may reflect courier coverage, delivery delays, address quality or serviceability.

A useful policy should consider several dimensions:

  • Order value and expected contribution
  • Customer’s previous delivery outcomes
  • Address completeness and serviceability
  • Courier and postal-zone performance
  • Product category and delivery characteristics
  • Delivery promise and actual delivery time
  • Whether the buyer has already confirmed the order

The intervention should match the problem.

An incomplete address may need correction. A customer who is unavailable may need a delivery reminder. An order with a higher risk of non-delivery may justify verification or a prepaid incentive.

A blanket COD block is a much broader intervention than any of those.

Our view is that merchants should preserve the ability to accept COD for legitimate buyers while applying more selective controls where the economics justify them.

COD risk does not end at checkout

The payment decision is only one part of the order lifecycle.

A merchant can verify a buyer’s intent before dispatch and still experience a failed delivery. An OTP or WhatsApp confirmation does not independently establish that the address is serviceable, the courier will attempt delivery correctly or the buyer will be available on the delivery date.

The operating model needs to continue beyond checkout.

A practical sequence is:

  1. Checkout records the order, payment method and relevant order attributes.
  2. The risk policy determines whether the order passes through, requires verification or needs review.
  3. The OMS applies the appropriate hold or release state.
  4. The courier provides shipment and delivery status events.
  5. A Non-Delivery Report (NDR) triggers a recovery decision where appropriate.
  6. Finance reconciles the final payment and logistics costs against the order outcome.

An NDR is not the same as an RTO. It is a signal that a delivery attempt failed and that further action may still be possible. Shiprocket’s documentation describes options such as contacting the buyer, requesting a reattempt or initiating an RTO. (support.shiprocket.in)

That distinction is operationally important.

If a courier reports that a customer was unavailable, the response may be to coordinate another attempt. If the address is incomplete, the response may be to correct it. If the customer has refused the order, the next action may be different.

A risk score alone cannot resolve all of those situations.

What merchants should do before the new MDR date is announced

The postponement creates a useful period for analysis.

A commerce team can use its recent order history to:

  • Identify the share of UPI payments that would be eligible for MDR.
  • Calculate actual COD RTO rates by courier, postal zone and order value.
  • Establish the direct and indirect cost of a failed delivery.
  • Review payment gateway contracts and settlement statements.
  • Model the cost of prepaid incentives and checkout friction.
  • Compare expected contribution per delivered order, not just payment fees.
  • Test selective policies against a control group.
  • Keep checkout rules reversible until the actual implementation date and billing treatment are confirmed.

The existing Flo Pilot & RTO Measurement Framework provides a more detailed measurement model, including intervention rate, verification rate, false positives and net delivered order rate.

The article’s purpose is different: establish the economic decision that should come before the intervention.

The decision is about contribution margin, not payment preference

UPI MDR does not make COD inherently bad.

It makes payment acceptance cost more explicit for eligible transactions. COD can still be valuable when it preserves conversion among legitimate buyers. Prepaid can be valuable when it reduces expected delivery failure or improves cash-flow and collection economics.

Neither conclusion should be assumed across every customer, order value, product or courier.

The better decision is to calculate the expected cost of each route, measure the impact of any intervention on conversion, and use the results to determine where a different payment policy is justified.

For merchants evaluating COD risk software, Tailoredd’s Flo product page currently lists the product as Pilot / Early Access. The public RTO measurement framework describes how merchants can evaluate performance using their own data.

There are no universal RTO-reduction percentages to rely on. The decision should be based on the merchant’s own economics.

Frequently asked questions

Is the 0.4% UPI MDR live as of October 31, 2026?

No. NPCI postponed implementation until after Diwali, with the new date to be notified separately. The original October 15 date is no longer the implementation date. (business-standard.com)

Does the announced MDR apply to all UPI payments?

No. The announced 0.4% MDR applies to eligible P2M UPI transactions above ₹2,000, with a ₹300 cap for transactions of ₹75,000 and above. The official FAQ also identifies exemptions and separate treatment for other transaction types. (financialservices.gov.in)

Should ecommerce merchants push COD buyers toward prepaid payments?

Only where the expected cost of COD failure and collection is greater than the cost of the prepaid route, after accounting for MDR, gateway charges, incentives and lost conversion.

Can merchants add the UPI MDR as a checkout surcharge?

The official FAQ says merchants cannot pass the MDR on to UPI customers. The customer-facing price should not be increased because of this fee. (financialservices.gov.in)

tailoredd Product Office
Written by

tailoredd Product Office

Product Strategy and Commerce Operations

The tailoredd Product Office publishes guidance for product managers and heads of commerce on roadmap prioritization, operating models, and platform-fit decisions.

View Profile →