Cashfree vs Razorpay for a ₹3 Crore D2C Brand: Settlement Timing, Not Fees, Decides It

Comparing Razorpay and Cashfree on MDR percentages is the wrong starting point for a D2C brand doing ₹25 lakh a month. The working-capital cost of a one-day settlement shift tends to dwarf the fee delta either sales team will negotiate, and that gap is where most of the real decision actually lives.

16 October 2026 · 9 min read

For a brand doing ₹25 lakh in monthly GMV, a single day's shift in settlement cadence is worth more in working capital than the entire MDR negotiation anyone bothered to have with either sales team. That's the actual arithmetic, and it almost never shows up in the comparison deck the founder arrives with.

Quick back-of-envelope so we're all looking at the same numbers. ₹25 lakh a month settling one day later, held at a working-capital cost of roughly 11% p.a. (which is where most of our D2C clients sit for their short-term facility — give or take, depending on who their banker is and whether they've bothered to shop around), works out to something in the ballpark of ₹7,500 to ₹8,000 a month of float cost per extra day. The MDR delta people fight over between Razorpay and Cashfree is around 0.1% — call it ₹2,500 on the same ₹25 lakh. The settlement-day cost is roughly three times the fee argument. And that's for one day. If you're on T+2 versus T+1, double it.

None of this is in the pricing-page comparison. Neither sales team is incentivised to raise it.

Why we used to default to Razorpay for every D2C build

Mostly inertia, if we're being honest. Razorpay's WooCommerce plugin has been around longer, their checkout UX was the one most Indian buyers had muscle memory for, and their docs — through roughly 2020-2022 — were simply better written than anyone else's in the space. When a client asked "what gateway", the answer was Razorpay unless there was a specific reason otherwise. That answer was right often enough that we stopped thinking about it.

The specific reason otherwise, back then, was usually international card acceptance (Cashfree used to be more forgiving on some BIN ranges — this may have changed, we haven't tested recently) or a client who'd already signed up for Cashfree because their CA recommended it. We'd integrate whatever was in front of us.

What shifted our default wasn't a feature. It was watching the finance side of three or four D2C builds in a row and realising the questions the founder was asking at month-end weren't about the gateway at all. They were about when the money hit the current account, and whether the Razorpay payouts report matched the Tally bank statement without manual intervention. The gateway choice was downstream of a question nobody had framed properly at the integration stage.

Why we now ask about settlement cadence before anything else

The question we ask first now, before anything about checkout UX or EMI support or whether the client wants a hosted page: what's your current settlement cadence, and what's the real cost of pulling that forward by a day?

Both gateways offer instant settlement as a surcharge product. Razorpay's instant settlement — last we looked, somewhere in the 0.15% to 0.25% range depending on the slab and whether you're on the base plan or a negotiated one — makes sense when you're genuinely cash-strapped and the alternative is a bank OD at 12-14%. Cashfree's instant settlement sits in roughly the same band (we think it's slightly cheaper for larger volumes, but the published rate card and the actual negotiated rate drift apart once you cross ₹50 lakh monthly, so take any specific number with a pinch of salt).

The point at which the surcharge stops being worth it is somewhere around the moment your working-capital cost drops below the instant-settlement fee annualised. For a brand with a cash reserve that covers two months of operating costs, paying 0.2% to get the money today instead of in two days is paying 36% p.a. on a two-day loan. Nobody frames it that way, but that's the math. For a brand that's actually tight on cash and would otherwise be dipping into an OD at 13%, the instant-settlement fee is a bargain. Context matters.

On 11 August last year, a client's Cashfree settlement for the previous day's sales got delayed past the RBI maintenance window — those Saturday-night-into-Sunday-morning NEFT/RTGS downtime windows that everyone forgets about until a settlement sits in limbo. The money did land, eventually, but it landed on Monday morning rather than the expected cadence. Nobody on the client's team noticed immediately because the Cashfree dashboard still showed the settlement as "initiated". Riya caught it during the Tuesday review — she was reconciling the previous week's bank statement against the gateway payout report and the dates didn't line up by one row. One extra day. On their volume, call it ₹9,000 of float cost, which is a rounding error — but the reconciliation effort to figure out why the dates had drifted ate about four hours of her time, plus a back-and-forth with Cashfree support that stretched across two days.

Razorpay handles the same scenario differently. In our experience their reversed/delayed settlements show up as a distinct row in the payouts report with a clear status flag, which Tally picks up cleanly if your import mapping is set up for it. Cashfree, at least on the reports we've looked at recently, folds reversed settlements back into the next successful settlement line with a reference ID that you have to click through to understand. That's a WooCommerce+Tally reconciliation problem specifically — if your accountant is matching settlement amounts to bank credits line by line, Cashfree's representation needs an extra step of manual tagging. (We see this less often on Shopify + Zoho Books stacks, where the matching logic is looser to begin with — may not bite the same way there.) Not a dealbreaker, but not nothing either — depends how line-by-line your accountant actually works.

(Aside: we've seen one client's CA just give up on matching gateway reports to the bank statement and reconcile only at the monthly aggregate level. Technically defensible. Ask us how we feel about it.)

The webhook side is where we've had a specific operational preference, and it's narrower than people realise. Cashfree's payment_success webhook fires in a different order relative to the order-creation flow than Razorpay's does — specifically, if your WooCommerce fulfilment hook is listening on the payment-success event to trigger the Shiprocket push or whatever your downstream is, the timing assumptions you built against Razorpay don't transfer cleanly. We've had to add a short delay-and-recheck pattern on Cashfree integrations to avoid firing fulfilment before the order status has fully committed. It's a ten-line fix once you know. The first time you hit it in production, it looks like a race condition in your own code, which is where the two days of debugging goes.

Razorpay's webhook race against the order_status_completed hook is its own separate mess — we've written about that elsewhere — but at least the ordering is predictable once you've learned it.

Where the migration from one to the other actually hurts

The migration itself, if the client has decided they want to switch, is roughly two weeks of engineering plus one reconciliation month of overlap where both gateways are live and the finance team is watching two sets of reports. That's a conservative estimate; we've done it faster when the catalogue is small and the client's accountant is sharp, and we've done it slower when the subscription-billing side had to be rebuilt against the other gateway's recurring-payments API.

Breaking down the two weeks, roughly: three days for the gateway integration swap on the WooCommerce side (plugin install, API key management, test transactions, verifying webhook endpoints, updating the checkout UX if the hosted flow is different); two days for refund-flow testing including partial refunds and the credit-note trigger; two days for subscription/recurring migration if applicable — this one is the wildcard, can balloon easily; two days for reconciliation report format changes on the Tally side; the rest for the inevitable Shiprocket webhook adjustments, the thank-you page analytics tracking that always breaks, and the client's accountant asking the same three questions four different ways.

The reconciliation-month overlap is the part founders underestimate. You can't just flip the switch and point all new orders at the new gateway, because you have 30-45 days of refund tail from the old gateway's transactions. During that window, refunds have to be processed against the original gateway, which means your finance team is running two reconciliation processes against two different report formats against one Tally ledger. We used to run this as a hard cutover; we stopped after the second time an accountant threatened to quit.

Which is why we rarely recommend switching unless settlement cadence is genuinely the binding constraint. If a brand is comfortably cash-positive, has a working relationship with their current gateway's support (that part matters more than people admit), and the reconciliation flow is stable, the ₹2,500 a month MDR saving does not pay for the migration inside any reasonable payback window. If the brand is growing fast and settlement float is pinching, the calculation flips — but only if the alternative gateway's cadence is materially better for the volume slab the brand is actually in, not the slab the sales team is promising they'll move you to next quarter.

One honest admission before we close. We've been going back and forth internally on how to model settlement float in the P&L forecast for a brand that's growing month-on-month at 15%+ — Mira and the finance side want a steady-state working-capital-cost line; the build team keeps pointing out that steady state is exactly the state these brands aren't in. A brand that's doing ₹25 lakh this month and ₹35 lakh in four months has a float cost that's compounding in a way the usual formula doesn't capture, and the gateway choice interacts with that in a way we haven't modelled cleanly. If someone reading this has solved it in a spreadsheet that doesn't require three analysts to maintain, send it over. We've got a tab open that's been half-finished for a while.

Ready to build something that works?

Get a free website mockup for your business. No commitment, no fluff — just what it could look like.

WhatsApp replies in seconds.

WhatsApp Call Free mockup
WhatsApp us · replies in seconds