How Top D2C Brands Recover Revenue After Checkout

Almost every rupee of marketing effort in Indian D2C is spent getting someone to press "Place order." Then the spending stops, the attention stops, and a startling share of those orders quietly fail to turn into money.

It's worth sitting with the arithmetic for a second. A brand doing ₹1 crore a month in placed orders, with a 25% RTO rate and another few percent lost to cancellations and returns, is not a ₹1 crore brand. It's a ₹70-something-lakh brand that pays freight on the difference. And unlike the top of the funnel — where improving conversion means fighting Meta's auction — everything downstream of checkout is yours to fix. No auction, no competitor bidding against you. Just operational slack that most brands never look at.

The brands that do look at it tend to think about it the same way: there is a second funnel after checkout, and it deserves the same treatment as the first. Order placed → order confirmed → order dispatched → order delivered → order kept. Every arrow in that chain has a drop-off rate. Every drop-off rate can be measured, and most can be moved.

The second funnel, drawn honestly

Here's roughly what the post-checkout funnel looks like for a typical COD-heavy Indian brand. Your numbers will differ — that's rather the point of measuring them — but the shape is remarkably consistent:

• Order placed → confirmed: 10–20% never should have shipped. Impulse orders, duplicate orders, wrong address, someone else using the phone. Uncaught, these become RTOs at full freight cost.
• Confirmed → dispatched: small leakage, but real — address failures, pin codes you don't serve, inventory that wasn't actually there.
• Dispatched → delivered: the big one. First-attempt delivery failure runs high in India, and each failed attempt is a coin flip on whether the parcel comes home.
• Delivered → kept: returns and exchanges, where the revenue is technically booked and then quietly reversed.

Notice what all four have in common. Every one of those drop-offs is a moment where somebody needed a decision from the customer — and the customer wasn't asked, or was asked in a channel they ignored. That's the thesis of the whole post-checkout playbook: the leaks are decision points, not process failures.

1.Confirm the order before you pay to ship it

The first and largest recovery point is also the least glamorous. Before a COD order goes to the courier, verify that the person actually wants it.

What brands underestimate is how much of this is not fraud or malice. It's a customer who ordered two sizes to see which fits and forgot to cancel one. It's a teenager who ordered from a parent's phone. It's someone who bought at 1am and by morning has thought better of it. None of those people will refuse to talk to you — they just aren't going to proactively email you to cancel. Being asked is enough.

The mechanics matter more than the idea, and we've covered them at length in the Shopify RTO setup guide and the confirmation script that actually works in Hindi. But the summary is: ask early (within an hour or two of the order, while intent is fresh), ask in the customer's language, and let people cancel gracefully. A cancelled order costs you nothing. An RTO costs you two-way freight plus three weeks of stuck inventory.

2.Fix the address before the courier finds it broken

Indian addresses are written for humans who already know the area. "Near Shiv Mandir, blue gate, opposite the medical store" is a perfectly good address if you live there and a coin flip if you don't. Add a missing flat number, an unreachable phone, or a pin code that maps to a different neighbourhood, and you have a delivery that will fail for reasons entirely unrelated to whether the customer wants the product.

The recovery here is cheap and boring: a pre-dispatch touch that confirms the address is complete, grabs a landmark, and asks when someone will actually be home. Brands that do this well fold it into the same confirmation conversation as step one — one call, two jobs. There's no reason to make the customer pick up twice.

The second half of this is the part most brands skip: write the corrections back. A landmark collected on a call and left in a call log helps nobody. It needs to land on the shipping label.

3.Treat every failed delivery as a save, not a status

This is the recovery point with the widest gap between what brands could do and what they actually do. A courier attempts delivery, fails, and marks it NDR. In most brands, what happens next is nothing — or an SMS. The courier tries again, fails again, and the parcel starts its journey home.

What makes this so painful is that the order was already confirmed. The customer wanted it. You paid to acquire them, paid to pick and pack, paid to ship it across the country — and lost it in the last kilometre because nobody asked "when should we come back?"

The brands that recover here have one thing in common: they respond to NDR within hours, not days, and they respond by talking to the customer rather than notifying them. The window is short. A second attempt scheduled for a time the customer chose succeeds far more often than a second attempt scheduled by the courier's route optimiser. We wrote up the whole NDR sequence separately in what happens after a customer doesn't answer your delivery call.

4.Recover the carts worth recovering — at full price

Abandoned cart recovery is technically pre-checkout, but it belongs in this conversation because it's the same discipline: a customer with demonstrated intent stalled at a decision point.

The mistake nearly everyone makes is the reflexive discount. Fire a 10%-off coupon at every abandoned cart and you'll book some recoveries — and you'll also teach your most engaged customers that abandoning is the cheat code. Worse, you'll never learn why the cart died. Payment failure and second-guessing the size look identical in your analytics and need completely different responses.

Better brands segment. Low-value carts get an automated nudge, because the economics don't support anything more. High-value carts — the ₹3,000+ ones — get an actual conversation, which recovers at full price and tells you what broke. If your payment gateway is silently failing on a particular bank's UPI flow, you will find out from three phone calls long before you find it in a dashboard.

5.Turn the return into an exchange

The last leak is the one that looks like a completed sale right up until it isn't. A customer receives the order, something is off, and they file a return.

Two things are recoverable here. The first is the obvious one: a sizeable share of returns are size or variant issues, and a customer who'd happily take a different size will often take a refund instead simply because the return button is easier to find than the exchange option. Reaching out and offering the swap converts a chunk of those.

The second is less obvious and probably worth more. A customer in the middle of a return is deciding whether your brand is worth a second order. If the pickup drags for ten days and they have to chase you for the refund, you have lost far more than one sale. Handling it quickly — confirming the reason, locking a pickup slot, being clear about the refund timeline — is retention work disguised as logistics work.

The measurement that makes this real

None of the above is worth much as a list of good intentions. What separates brands that recover revenue from brands that talk about it is that the former put numbers on the second funnel and look at them weekly. At minimum:

• Confirmation rate — of COD orders reached, what share confirm? And separately: what share do you actually reach?
• Dispatch-to-delivery rate, split by courier and by region. The averages hide everything interesting.
• RTO rate, split by confirmed vs. unconfirmed orders. If these two numbers aren't very different, your confirmation process isn't working.
• NDR resolution rate — of failed first attempts, what share eventually deliver?
• Return rate and, crucially, exchange-to-refund ratio.

Five numbers. Most brands can't produce three of them on demand, which is precisely why the money sits there uncollected.

Why voice keeps showing up in this playbook

You'll have noticed that four of the five recovery points end with "talk to the customer." That isn't a channel preference; it's what the moments require. Each one needs a decision — confirm or cancel, which day, which size, is this address right — and decisions are what WhatsApp and SMS are worst at. They're excellent for delivering information and nearly useless for extracting an answer from someone who isn't inclined to type one. We laid out where that line sits in calls vs. WhatsApp.

The reason the post-checkout funnel went unfixed for so long isn't that nobody understood it. It's that the fix required a room full of tele-callers, and that maths only worked for brands large enough to run a call centre. That constraint is what actually changed. An agent that holds a real conversation in Hindi or Tamil makes the same operational playbook available to a brand doing 800 orders a month.

The playbook itself is old. Good retailers have always called their customers. What's new is that you no longer need forty people to do it.