Part II — Payments

Chapter 13 · Merchant Acquiring and Payment Gateways

Every chapter so far paid from the customer's side of the counter. This one crosses it: what does it take to accept a payment? The acquiring chain and its licences, the thin-margin economics Chapter 8 kept promising, settlement cycles priced as working capital, chargebacks against UPI's no-dispute asymmetry — and the pricing exercise that makes a merchant of you.

📖 Chapter 13 of 35 ⏱ ~60 min read 🎓 Session 16 · CILO 2 Core 70% Stable 30% Volatile 0% 📝 Quiz · 30 MCQs

🎯 Learning Objectives

1. The Merchant's Problem

"Accepting payments" sounds like one thing. It is at least six: signing the merchant and verifying her (KYC for merchants), equipping her (POS terminal, QR code, payment link), authorising each transaction on some rail, moving the money to her account (settlement), matching every rupee to an invoice (reconciliation), and handling the failures (refunds, disputes). Chapter 6 met the licence; this chapter meets the job.

The chain has two professional classes. An acquirer is traditionally a bank that enrols merchants and settles funds to them. A payment aggregator (PA) — the non-bank version, Razorpay and its peers — does the merchant-facing work at scale: onboarding, software, support, pricing, while merchant funds sit in escrow with a scheduled bank under RBI authorisation, with the aggregator never free to use them. The pure payment gateway (PG) sells software only — routing, checkout, reconciliation — and touches no funds; the moment funds flow through the firm's accounts, it is a PA in law, whatever the website says. That is Chapter 6's funds-flow rule, now in its natural habitat.

📝
The merchant's eye view: she rarely knows or cares which class serves her. She knows: sign-up time, the fee on each instrument, when money lands, and who answers the phone when a payment fails. Every product decision in this chapter is one of those four answers wearing a job title.

2. The Economics of Acquiring

Recall Chapter 12's split: of an MDR, the interchange goes to the issuer and the scheme fee to the network, and the acquirer keeps the remainder — typically the thinnest slice. Acquiring is therefore a volume business built on basis points: the aggregator's real margins come from scale on software, cross-sell and float-adjacent services, not from the payment fee itself.

UPI sharpened the arithmetic twice over. At zero MDR, serving a merchant's UPI flow earned the aggregator nothing while still costing technology and support — Chapter 8's cross-subsidy, now seen from the merchant's side of the counter. The October 2026 framework changed the large-ticket end (0.4 per cent above ₹2,000, per Chapter 12) but left the everyday flow free. So the industry did what Chapter 8 predicted: it turned acceptance into customer acquisition — the merchant onboarded for payments becomes the buyer of lending, payroll, inventory and accounting software. The gateway's fee schedule is a marketing document; the business model lives one product behind it.

Merchants, for their part, see three pricing shapes. Percentage pricing (MDR passed through or marked up): cheap for small tickets, expensive for large ones. Flat per-transaction pricing: predictable, great for large tickets, punishing for small ones — a flat ₹8 on a ₹100 sale is an 8 per cent fee. Hybrid and subscription pricing: a monthly platform fee plus lower variable rates, which the SaaS era made common. The hands-on exercise makes you compute all three against a real ticket mix — the punchline is that no plan wins universally; the ticket distribution decides.

📋 Stable content — Reviewed: October 2026 · Next review: April 2027

3. Settlement Cycles as Working Capital

When a customer pays, the merchant's account does not receive money; it receives a promise on a schedule. Card and UPI merchant settlements run broadly on T+1 — funds land the next working day — and the PA framework's discipline holds throughout: merchant money sits in escrow with a scheduled bank from capture to payout, on prescribed timelines, never available to the aggregator. Instant or same-day settlement products exist for merchants who will pay for speed.

The schedule is not bureaucracy; it is a loan. Every day of settlement delay is a day the merchant finances her own sales. Price it the Chapter 8 way: a merchant settling ₹1,16,000 a day on T+1 carries one day of sales as working capital; at a 12 per cent overdraft rate that float costs about ₹38 a day. An instant-settlement upgrade priced at 0.05 per cent of settled value would cost ₹58 a day — more than the float it frees at 12 per cent, and less the moment her true cost of funds crosses roughly 18 per cent. Settlement speed, like everything else in this part, is a priced product — and the right answer depends on the merchant's cost of money, not on any slogan about speed.

🌎
Real World: Small merchants feel the cycle hardest: thin margins, no overdraft, daily purchases. The payments industry's honest answer to financial inclusion at the till was the QR code — acceptance at zero fixed cost — and its honest tension is that the merchants least able to finance T+1 are the ones least able to pay for T+0.

4. Chargebacks and the UPI Asymmetry

A chargeback is the card system's dispute machinery. The cardholder contests a charge with her issuer; the issuer reverses the payment through the network; the acquirer debits the merchant (amount plus a fee); and the merchant may answer with evidence — delivery proof, signature, terms — in representment, which the issuer accepts or rejects. Reason codes sort disputes into fraud ("never made this payment") and service ("made it, but the goods never came"), and the burden of proof sits mostly with the merchant.

Now the asymmetry this chapter exists to surface. Cards carry dispute rights and cost MDR; UPI carries no chargeback and costs (almost) nothing. An authorised UPI payment is account-to-account and irrevocable by default — Chapter 9's design consequence. For a merchant, the trade is exact: accept cards and buy dispute protection plus credit capability at 1.5–2.5 per cent; accept UPI and get near-free settlement with no recourse if the customer claims failure. For customers the same asymmetry runs the other way — which is precisely why Chapter 14 (fraud, mule accounts and the liability rules) is the payments story's necessary next chapter. Refunds, note, are different from disputes: a refund is the merchant voluntarily reversing its own capture, on either rail, through the acquirer's systems.

5. The Gateway Stack: Routing, Retries, Tokens

The gateway's product is one integration, every rail: cards, UPI, netbanking, wallets, EMI options, pay-later — behind a single checkout and a single reconciliation file. On top of that base sit the features that actually differentiate gateways.

Smart routing: when a transaction can travel by several rails or several acquiring banks, the gateway chooses by success rate, cost and latency — sending each attempt where it is most likely to succeed. Retries and cascades: a failed card attempt can be re-routed automatically before the customer ever sees an error; in a market where payment failures once killed checkouts, success-rate engineering became the core competency. Reconciliation: matching every settlement line to every order across rails — the unglamorous product that decides whether the merchant's books close. And tokenisation: under RBI's card-on-file framework (from 2022), card details stored for repeat payments must be replaced by tokens — the gateway holds the vault, the merchant never touches raw card numbers, and subscription commerce kept working.

💡
Pro Tip: When a gateway pitches "99 per cent success rates", ask two questions: measured on whose traffic, and inclusive of retries or not? Success rate is the most gamed metric in Indian payments — the honest version is "attempts that became money, divided by attempts", stated per rail and per issuer.

6. Beyond the Till

Acquiring is becoming invisible, and that is its future. The QR sticker became the soundbox — the little speaker that confirms payments by voice — which is a merchant SaaS subscription wearing a payments costume, and for many small merchants the first software they ever paid for. Checkout is absorbing credit: EMI options, buy-now- pay-later and credit-on-UPI ride the same acquiring rails, which is Chapter 19's subject arriving through the merchant's till. And payments are dissolving into platforms — the marketplace, the booking app, the ERP — where acceptance is an API call, which is Chapter 34's embedded world.

Notice what stays constant through every one of those shifts: someone onboards the merchant, someone moves the money on a schedule, someone owns the failure. The licences of Chapter 6, the economics of Chapter 8, the MDR of Chapter 12 and the settlement maths of this chapter travel together into every future version of the till. And the risk side — the fraud that rides payment rails, the mule accounts that drain them, and the liability rules that decide who eats a loss — is important enough that Part II gives it a full chapter next.

🛠 Hands-On: The Gateway Pricing Exercise

This is the session's classroom activity, made self-paced. A merchant's monthly profile: 3,000 transactions — 2,400 UPI (all tickets below ₹2,000, averaging ₹700), 400 debit (averaging ₹2,500), 200 credit (averaging ₹4,000). Three pricing plans are on the table. Compute, compare, advise — then price the schedule and the dispute.

Steps

  1. Volumes first. Compute UPI, debit and credit monthly value, and total value.
  2. Plan A — pure MDR pass-through. UPI ₹0 (all below the line); debit 0.9 per cent; credit 2 per cent. Monthly cost?
  3. Plan B — hybrid. ₹1,499 monthly platform fee + ₹3 per transaction + MDR pass-through on cards. Monthly cost?
  4. Plan C — UPI-only. Zero fees; but assume one fifth of card customers abandon rather than switch payment method. At a 25 per cent gross margin, what does "free" cost?
  5. Settlement check. The merchant's daily sales ≈ total ÷ 30. Her overdraft rate is 12 per cent. Instant settlement is offered at 0.05 per cent of settled value. Should she buy it? At what cost of funds would the answer flip?
  6. The dispute. A ₹8,000 credit-card sale is charged back as fraud. The merchant has delivery proof. Trace the money and the process, and state the possible net outcomes.
  7. Advise. In 120 words: which plan, which settlement speed, and what the advice reveals about "cheapest fees" as a way to choose payments.
View Solution / Walkthrough

1. UPI: 2,400 × ₹700 = ₹16,80,000; debit: 400 × ₹2,500 = ₹10,00,000; credit: 200 × ₹4,000 = ₹8,00,000. Total ₹34,80,000 a month; daily ≈ ₹1,16,000.

2. Plan A: ₹0 + 0.9% × 10,00,000 (= ₹9,000) + 2% × 8,00,000 (= ₹16,000) = ₹25,000 — an all-in take of 0.72 per cent of GMV, entirely earned on the 17 per cent of transactions that are cards.

3. Plan B: 1,499 + (3,000 × 3 = 9,000) + 25,000 = ₹35,499. The platform fee buys the software — reconciliation, support, soundbox — and here it is not worth ₹10,499 a month unless those tools save more.

4. Plan C: fees ₹0; lost sales = 20% of card value = ₹3,60,000; lost contribution at 25 per cent margin = ₹90,000. "Free" is the most expensive plan on the table — payment choice is revenue, not only cost.

5. Float freed by instant settlement: one day's sales ≈ ₹1,16,000; worth 1,16,000 × 12%/365 ≈ ₹38/day. Cost: 0.05% × 1,16,000 = ₹58/day. At a 12 per cent cost of funds, do not buy. The flip point: 58 × 365 / 1,16,000 ≈ 18 per cent — above that cost of funds, instant settlement pays for itself.

6. Cardholder disputes → issuer chargeback → acquirer debits the merchant ₹8,000 + fee → merchant submits delivery proof in representment → issuer decides. Outcomes: win (₹8,000 returned, fee often retained), lose (₹8,000 gone — goods shipped, money returned), or partial. On UPI none of this exists — which is exactly the asymmetry of Section 4.

7. Strong advice: Plan A now; revisit B when the software saves real reconciliation hours; never C at this card mix; skip instant settlement at 12 per cent funds cost. The reveal: "cheapest fees" ignores what payments earn — card acceptance is a revenue product, settlement speed is a financing product, and the fee schedule is only one of the three dials.

Key Takeaways

1

Acceptance is a six-part service — onboarding, devices, authorisation, settlement, reconciliation, disputes — split between bank acquirers, authorised PAs and pure software gateways.

2

Acquiring runs on basis points: after interchange and scheme fees, the thinnest slice remains — so UPI acceptance became customer acquisition for everything sold behind the till.

3

Settlement is a loan: T+1 means the merchant finances a day of sales. Whether instant settlement is worth 0.05 per cent depends entirely on her cost of funds.

4

The dispute asymmetry is exact: cards buy chargeback protection at 1.5–2.5 per cent; UPI is near-free with no recourse on authorised payments.

5

No pricing plan wins universally — the ticket distribution decides. Percentage suits small tickets, flat suits large, and "free" can cost a fifth of your card revenue.

Test Your Understanding

Choose one best answer per question. A calculator is allowed for the arithmetic items; the merchant profile of the hands-on is the data for the computation questions. Aim for 24/30 before moving to Chapter 14.

1. In acquiring, the traditional bank-side role is the:

2. Under the PA framework, a non-bank aggregator's merchant funds must:

3. A software firm becomes a payment aggregator in law the moment:

4. Why is acquiring called a thin-margin volume business? (Hand arithmetic allowed.)

5. With UPI free at the everyday ticket, aggregators serve it because:

6. A merchant settling on T+1 effectively:

7. Daily sales ₹1,16,000; overdraft 12 per cent; instant settlement priced at 0.05 per cent of value. Monthly-wise, the service: (Calculator allowed.)

8. The instant-settlement decision flips in the merchant's favour when her cost of funds exceeds about:

9. A chargeback is initiated by:

10. In representment, the merchant:

11. The dispute asymmetry between cards and UPI is that:

12. A refund differs from a chargeback because a refund is:

13. Smart routing in a gateway chooses:

14. The retry-and-cascade feature exists because:

15. Under RBI's card tokenisation rules (from 2022), a merchant storing cards for repeat sales must:

16. The gateway feature that most decides whether a merchant's accounts close is:

17. For the hands-on merchant, Plan A (MDR pass-through) costs monthly: (Calculator allowed.)

18. Under Plan C (UPI-only, one fifth of card customers abandoning, 25 per cent margin), "free" costs: (Calculator allowed.)

19. Flat per-transaction pricing is a poor fit when:

20. Percentage (MDR-based) pricing suits a merchant whose:

21. The payments soundbox is best read as:

22. Between capture and payout, the obligation to safeguard merchant money sits with:

23. The honest version of a "99 per cent success rate" claim is:

24. EMI and buy-now-pay-later options at checkout are, structurally:

25. When a ₹8,000 chargeback lands, the merchant's immediate position is:

26. The payments industry's main inclusion answer for small merchants was:

27. The gateway's foundational product promise is:

28. What the pricing exercise chiefly reveals is that:

29. The chapter's one-line summary of acquiring is that it:

30. Part II's necessary next chapter, on everything this one deferred, covers: