OTT and SaaS MDR Reconciliation Playbook for Indian Subscription Businesses

Last updated: 23 June 2026 — A seven-step monthly playbook for OTT and SaaS subscription merchants in India, calibrated to the Income-tax Act 2025 framework live since 1 April 2026 and the zero-MDR regime current under the Payment and Settlement Systems Act 2007.

An OTT subscription business in Mumbai processes ₹15 crore of monthly gross merchandise value across UPI AutoPay, cards, RuPay credit on UPI, net banking, and a thin wallet tail. The CFO has a contracted blended rate of 1.55 percent on cards, a contracted platform fee on UPI bank-account volume, and a recurring add-on schedule that is supposed to apply only to specific product flags. The blended monthly number on the settlement summary lands within touching distance of expectations every month, so for a year the finance team treats the gateway statement as truth. The OTT and SaaS MDR reconciliation playbook is the discipline that takes three hours each month, ingests the per-transaction file rather than the summary, runs an eight-pattern leakage check against the per-transaction classification, and surfaces the basis points that the blended view conceals. On this book, the first quarter of the discipline identifies ₹6 lakh of leakage and recovers ₹3.8 lakh of it within the dispute window.

Quick reference: the seven-step monthly close

Why is the published headline the wrong reconciliation baseline?

A finance team that anchors monthly reconciliation to the published 1.95 or 2 percent gateway rate is reconciling against a rate card designed for a sub-five-lakh-monthly merchant who never negotiated. For an OTT or SaaS business processing crores per month, the contracted rate sits in the merchant agreement schedule and is materially lower on cards — commonly 1.4 to 1.6 percent at ₹1 crore plus monthly card volume — with specific addenda for premium, international, EMI, and recurring product flags. Reconciliation that uses the published number under-detects leakage because it accepts a rate the contract no longer requires. The playbook anchors every per-network effective-rate comparison to the live merchant-agreement schedule, refreshed each time the gateway updates a slab or a product is enabled.

The published rate also conflates instrument categories. A single 1.95 percent headline hides that bank-account UPI is zero network MDR while Amex is around three percent. A UPI-heavy method mix on a flat blended quote overpays relative to a method-mix-weighted true cost, while a card-heavy mix can under-pay short term and absorb a silent reclassification later. The contracted baseline is per-instrument per-network, not blended.

What does step one — ingesting the gateway settlement — actually require?

The per-transaction settlement export, not the daily summary, is the ingestion target. Razorpay, PayU, Cashfree, and PhonePe each expose a per-transaction settlement file with payment-level identifiers, instrument codes, network codes, BIN strings or BIN ranges, fee column, GST column, and a product-flag column where applicable. The unified table is keyed on settlement_id and payment_id, with the gateway name carried as a column so a multi-gateway OTT or SaaS book lands in one place. The orchestration layer — Juspay where present — is a separate fee line, not a percentage MDR, and rides as a per-transaction fixed-fee column in the same file rather than inside the MDR percentage.

The ingestion has to retain the settlement currency, the gross transaction value, the fee value, the GST value, the net credit value, the settlement date, and any chargeback or dispute identifier. Daily summaries discard the BIN string and the product flag and are therefore unfit for the per-network and per-pattern checks that follow.

What does step two — classify by network and instrument — produce?

Every transaction lands in a classification table with five tags. Instrument is the rail — UPI, debit card, credit card, prepaid card, net banking, wallet, EMI. Network is the brand within the rail — UPI bank account, RuPay credit on UPI, PPI on UPI, Visa, Mastercard, RuPay debit, RuPay credit, Amex, Diners. BIN tier is the issuer-product class — consumer, premium or rewards, commercial or corporate, signature or infinite — read from a BIN-to-tier mapping table sourced from the acquirer. Scope is domestic or international, read from the issuer-country segment of the BIN. Product flag is the gateway product the transaction was routed through — Subscriptions, Recurring, EMI, Instant Settlement, International.

The classification has to separate three UPI lines that many gateways display under a single UPI heading: bank-account UPI at zero network MDR, RuPay credit on UPI at roughly two percent above ₹2,000, and PPI or wallet on UPI at 0.5 to 1.1 percent above ₹2,000. Without that separation, a UPI-heavy OTT mix quietly accumulates 2-plus percent on the credit-on-UPI slice while the finance team believes UPI is free.

How is per-network effective rate computed and what does it expose?

Per-network effective rate is fees divided by network volume in the period. For each network — bank-account UPI, RuPay credit on UPI, PPI on UPI, Visa consumer credit, Mastercard consumer credit, Amex, Diners, RuPay debit, Visa or Mastercard debit, net banking, wallets, EMI, international — the playbook computes total fees as the sum of the fee column for that network and total volume as the sum of gross transaction value, then divides. The result is the network’s actual effective rate for the month.

Compared against the contracted slab for that network, the gap is the leakage. The discipline also tracks legitimate gateway platform fee separately from network MDR, because a 1 to 2 percent platform fee on a UPI bank-account transaction is contractually billable while a network MDR line on the same transaction is not. The two appear in different fields on the settlement file in some exports and conflated into a single fee column in others — the classification work in step two has to surface which is which before step three is meaningful.

What are the eight leakage patterns the playbook flags every month?

The merchant-fee taxonomy carries eight high-confidence patterns. Each maps to a per-transaction detection rule and a rupee-quantified recovery line.

Pattern one — non-zero network MDR on zero-MDR instruments. Bank-account UPI P2M and RuPay debit P2M are statutorily zero network MDR under Section 10A of the Payment and Settlement Systems Act 2007 and Section 269SU of the Income-tax Act read with Rule 119AA. Any positive network-MDR line on those instruments is a flag. Platform fee is a separate line and is legitimate where contracted.

Pattern two — premium and signature cards routed without contract basis. Premium, rewards, signature, and infinite cards carry higher interchange and gateways may apply a non-qualified surcharge or route them into the three-percent premium slab. The flag fires when consumer-tier BINs are billed above the contracted standard rate.

Pattern three — domestic BIN charged at international rate. BIN mis-classification routes a domestically issued card through a cross-border assessment and a forex layer. The flag matches issuer country from the BIN against the declared scope.

Pattern four — commercial or corporate cards in the wrong slab. Commercial cards carry higher interchange. The common direction of leakage is consumer cards routed into the three-percent commercial slab. The flag compares slab to BIN tier.

Pattern five — Amex or Diners hidden inside a blended flat rate. A single 2 percent blended quote conceals that Amex and Diners cost 2.95 to 3.5 percent. The flag computes Amex and Diners effective rate separately and surfaces the spread.

Pattern six — flat-rate concealment of per-network cost differences. A method-mix-weighted expected cost compared to the flat rate actually billed exposes the gap. For a UPI-heavy OTT mix this can be material in either direction.

Pattern seven — MDR retained on refunds and chargebacks with dispute fees stacked. MDR is non-refundable industry-wide. On refunded orders the original MDR is retained; on chargebacks a ₹200 to ₹750 dispute fee is stacked on top. The flag totals MDR retained on refunded and charged-back transactions per month.

Pattern eight — recurring add-on and eNACH rejection fees without contract basis. Subscription add-ons of around 0.99 percent per recurring transaction stack on base MDR; eNACH mandate-rejection fees of about ₹15 plus 18 percent GST per failed debit accumulate on retries. The flag fires when the add-on or the rejection fee appears on a transaction whose contract schedule does not include the matching line.

The eight patterns cover the same per-transaction file. A single monthly run returns all eight flag lists.

How do steps six and seven — chase recovery and brief the board — close the loop?

Recovery is built per gateway and per pattern. The dispute pack carries transaction identifiers, rupee exposure per flagged transaction, the contract clause or regulatory citation that justifies the dispute, and the gateway’s dispute window in days. Patterns one, three, four, five, and eight typically have the strongest in-window recovery probability because the citation is contractual or statutory; patterns two and seven often require renegotiation rather than dispute because they are within the gateway’s contractual posture but not within the merchant’s intent. Pattern six is always a renegotiation, not a dispute.

The board view is one page. Contracted blended rate, actual effective rate, basis-point gap, rupee gap, dispute pipeline, recovery realised, and the ninety-day rolling trend per network and per pattern. Over three months the trend tells the board whether a renegotiation is holding, whether a gateway reclassification has happened silently, or whether the method mix has shifted into a higher-cost cell faster than the contract anticipated.

Worked example: ₹15 crore OTT book, three hours a month, ₹3.8 lakh recovered

A Mumbai-based OTT subscription business runs a ₹15 crore monthly GMV book across Razorpay and PayU with a small Cashfree tail. The method mix is roughly 55 percent UPI bank-account, 7 percent RuPay credit on UPI, 4 percent PPI on UPI, 22 percent Visa or Mastercard consumer credit, 4 percent Amex, 3 percent debit cards, 3 percent net banking, and 2 percent EMI. The contracted blended rate on cards is 1.55 percent; the UPI platform fee is contracted at a low single-digit basis-point figure separately from the zero network MDR; the recurring add-on is contracted only on Razorpay’s Subscriptions product.

The monthly close runs in three hours of analyst time once the contract-rate table and the BIN mapping are stable. The first quarter surfaces three structural leakage lines.

PPI on UPI billed against UPI volume — ₹2.1 lakh per quarter. The Razorpay settlement export displayed PPI-on-UPI transactions under a single UPI heading and the team had been treating the entire UPI line as zero. The per-network effective rate for the PPI-on-UPI segment came in at 1.1 percent on the above-₹2,000 slice, applied legitimately as interchange under the NPCI March 2023 wallet-interoperability circular, but billed against transactions the finance team thought were zero-cost. The variance was structural — the cost is real — but the disclosure was opaque. The renegotiation outcome was a separate line in the settlement export for PPI-on-UPI volume.

Amex hidden inside a blended flat rate — ₹1.7 lakh per quarter. The Amex effective rate landed at 2.95 percent on Razorpay and 3.0 percent on PayU. Both rates are within published gateway slabs for Amex. The leakage was not the rate itself; it was that the team’s monthly working anchored to the 1.55 percent contracted card blended rate and therefore implicitly cross-subsidised the Amex slice. The fix was twofold: surface Amex as a separate per-network line in the management accounts so the unit economics are visible, and renegotiate forward to a sub-2.95 percent Amex addendum, in exchange for guaranteed Amex volume.

Recurring add-on billed without contract basis — ₹2.2 lakh per quarter. A subset of one-time card-on-file transactions had been routed through Razorpay’s Subscriptions product as a side effect of a checkout-flow change. The 0.99 percent recurring add-on was applied to those transactions even though the customer journey was a single charge, not a recurring schedule. The contract schedule covered the add-on only for genuinely recurring SKUs. The flag fired in pattern eight; the transactions were collected by checkout-flow segment; the dispute went in within the gateway’s window with the checkout-flow change date and the contractual scope of the add-on.

Total first-quarter leakage identified was ₹6 lakh. The disputes on the PPI-on-UPI structural variance and the recurring-add-on misapplication landed ₹3.8 lakh of in-window recovery. The Amex line went into the contract review file for the next quarterly renegotiation rather than into a dispute. Analyst time was 9 hours across the three months — three hours per monthly close — once the classification tables were stable. The 90-day trend went on the board sheet and the basis-point gap between contracted and actual effective rate closed by 8 basis points by quarter end.

How does the GST and TDS overlay fit into the playbook?

GST is unchanged. At 18 percent it applies to the MDR and the gateway platform fee components, never to the transaction value, and it appears as a separate line on the monthly tax invoice. The reconciliation discipline tracks the per-transaction GST totals from the settlement file and reconciles them against the aggregator’s monthly tax invoice before GSTR-3B filing, so that input tax credit lands cleanly in the merchant’s GSTR-2B.

The Income-tax Act 2025 framework live since 1 April 2026 carries TDS through Section 393, Section 394, and Section 413 with payment codes 1001-1092. The legacy 194O regime sits inside the new framework as Section 393(1) Sl. 8(v) at payment code 1035, at 0.1 percent — not the original 1 percent that ran from October 2020 to September 2024. This is the TDS the e-commerce operator deducts on payments to the e-commerce participant, on gross transaction value, distinct from the MDR or platform fee. The OTT and SaaS playbook books all three as separate lines per transaction: gross sale (with code 1035 at 0.1 percent where applicable), MDR or platform fee (with 18 percent GST as a separate line), and net credit to bank.

How does the playbook tie into the broader merchant-fees discipline?

The seven-step monthly close is the operating cadence. The eight-pattern leakage taxonomy is the detection layer. The contracted-rate table is the baseline. The 90-day trend is the board view. The disputes are the in-window recoveries; the renegotiations are the forward rate. For a deeper read on any one pattern — flat-rate concealment, the Razorpay 0.99 percent subscription add-on, the eNACH rejection-fee stack, the Amex and Diners hidden cost, the RuPay credit on UPI mislabel, or the refund and chargeback MDR retention — the merchant-fees cluster carries pattern-by-pattern playbooks that the monthly close calls into.

Continue reading

• Subscription SaaS MDR economics: AutoPay vs cards vs eNACH
• Razorpay MDR reconciliation: published 2% vs negotiated 1.4-1.6%
• Flat-rate MDR concealing per-network cost differences
• Amex and Diners hidden inside blended MDR
• Recurring add-on and eNACH mandate-rejection fees stacking
• MDR charged on zero-MDR UPI and RuPay debit
• Cluster hub: Merchant fees and MDR reconciliation
• Money page: Payment gateway reconciliation

External authority: National Payments Corporation of India — the operational source for UPI and RuPay network rules, the zero-MDR mandate on bank-account UPI P2M and RuPay debit P2M, the RuPay-credit-on-UPI interchange schedule operative since October 2022, and the March 2023 wallet-interoperability circular that frames the PPI-on-UPI interchange cell.