The Return Fraud Problem Every D2C Brand in India Is Quietly Living With

If you run a direct-to-consumer brand in India, you already know the number that keeps you up at night. It is not your CAC. It is not your ad spend. It is the return rate sitting somewhere between 20 and 40 percent, depending on your category, and getting worse every festive season.

Most founders and ops leads have made peace with returns as a cost of doing business. What far fewer have made peace with is the slice of those returns that are not honest at all. A shirt that comes back worn and smelling of perfume, claimed as "never used." A shoebox that arrives back with a completely different pair inside. A customer who insists the package never showed up, even though the tracking page says otherwise. This is not bad luck. It is return fraud, and in 2026 it has become one of the largest hidden costs on a D2C brand's P&L.

This piece is about what is actually happening inside Indian D2C returns right now, why the usual fixes do not touch the real problem, and what an evidence-first approach to delivery and returns actually looks like in practice.

The size of the problem is bigger than most brands assume

Globally, online return rates have settled at around 19 to 20 percent, more than double what physical retail sees. That alone is a heavy tax on margin. But the fraud layered on top of that base rate is what really erodes profitability.

Industry estimates from the National Retail Federation put fraudulent returns at roughly 9 to 15 percent of all returns processed, translating into losses north of 100 billion dollars a year globally. Shopify's own merchant data suggests that for every 100 dollars in returned merchandise, sellers lose about 10 dollars specifically to fraudulent activity. And the trend line is not flat. Abusive returns, tracked by fraud analytics firm Signifyd, climbed 64 percent between early 2024 and mid 2025, with apparel abuse alone rising 13 percent year over year.

In India, the picture is arguably worse because of two things unique to this market: heavy reliance on cash on delivery, and fashion and footwear categories that already run return rates between 25 and 35 percent, spiking toward 40 percent during big sale events. Add wardrobing and swap fraud on top of that base, and a meaningful chunk of what looks like a "normal" apparel return rate is actually customers using your inventory for free and sending it back.

Then there is RTO, return to origin, which is a related but distinct problem. When a COD order gets refused or never delivered, the brand eats both the forward and reverse shipping cost with zero revenue to show for it. Depending on category and courier, RTO can sit anywhere from 10 to nearly 40 percent of COD orders, and Indian D2C brands collectively lose thousands of crores a year to it. Return fraud and RTO are different problems, but they share the same root cause: nobody actually verifies what happened at the doorstep.

What return fraud actually looks like on the ground

It helps to name the specific behaviours, because "return fraud" as a phrase hides how mundane most of it looks.

Wardrobing. A customer buys a dress for a wedding, wears it once, steams out the crease, and returns it claiming it was never worn. Nearly half of retailers report seeing this regularly, and for occasion wear and going-out apparel it is close to routine.

Bracketing. A customer orders the same t-shirt in three sizes, keeps the one that fits, and returns the other two. On its own bracketing is not exactly fraud, it is a rational response to sellers who do not publish accurate size charts. But it is now practiced by well over half of online shoppers, and it quietly inflates return volume and reverse logistics cost across the board.

Swap fraud. The most damaging of the three. A customer receives a branded shirt worth 1,500 rupees, and returns a completely different, cheaper item, sometimes a cheap knockoff bought elsewhere, sometimes literally a t-shirt from their own cupboard. Without any record of what was actually delivered, the warehouse has no way to prove the swap happened.

Empty box and short shipment claims. The customer receives the parcel, opens it, and claims the box arrived empty or missing items, sometimes to get a partial refund without returning anything at all.

False damage claims. The product arrives in perfect condition, but the customer photographs it after deliberately damaging it, or in a growing number of cases in 2026, submits an AI-generated damage photo that never happened in the physical world.

Every one of these has the same defence available to the brand, and it is the same defence courts, marketplaces, and payment gateways all ask for: proof of what was actually delivered, compared against proof of what actually came back. Without that comparison, a brand's only options are to refund everyone (and eat the fraud) or dispute everyone (and lose genuine, loyal customers who did nothing wrong). Neither option is good business.

Why the existing toolkit does not solve this

Most D2C brands already have a logistics stack. Shiprocket or Delhivery for shipping, WhatsApp for customer updates, maybe an NDR (non-delivery report) tool to chase COD refusals, maybe a QC step at the warehouse when returns land back. On paper this looks like a complete pipeline. In practice it has a massive blind spot sitting right in the middle of it.

Courier partners are built to move parcels, not verify contents. A delivery agent has thirty seconds and forty more stops to make. Asking them to inspect, photograph, and log the condition of every item they hand over is unrealistic, and every brand that has tried to build a courier-side checklist has learned this the hard way, either through poor compliance or through couriers simply refusing to take on the liability.

Warehouse QC on the return side has the opposite problem. By the time a returned item lands back at the warehouse, days have passed. The team inspecting it has no idea what condition the item left the warehouse in, only what it looks like now. Even a sharp-eyed QC associate cannot tell you with certainty whether a stain was there at dispatch or happened afterward, because there is nothing to compare it against.

Photos at delivery, where they exist at all, tend to be a single blurry shot the courier snaps to prove drop-off, not evidence a brand can actually use to defend a refund decision. And customer-submitted return photos are, unsurprisingly, taken by the person with every incentive to make the item look pristine.

The result is that most Indian D2C brands are running their returns process on trust, backed by nothing. When a dispute happens, it comes down to the brand's word against the customer's, and brands tend to lose that argument because saying no to a refund, even a fraudulent one, carries real reputational risk on social media and marketplace ratings.

What an evidence chain actually changes

The fix is not more policy, and it is not stricter return windows that punish honest customers along with dishonest ones. The fix is building a verifiable record of the product at two fixed points: the moment it lands with the customer, and the moment it leaves the customer's hands again on return. If both of those are captured consistently, the comparison basically does the arguing for you.

This is the core idea behind Vefri. It does not ask couriers to become inspectors, and it does not ask your warehouse team to play detective on a returned item with no history. Instead it puts the verification step where it actually belongs, on the customer's own phone, at the two moments that matter.

Here is roughly how it plays out on a real order.

Step one, capture at delivery. The moment Shiprocket or Delhivery marks an order out for delivery, the customer gets a WhatsApp link. Before they even open the sealed package, they record a short guided video, ten seconds or so, walking through category-specific checkpoints, front of the product, any tags, the seal, whatever matters for that item type. This happens entirely inside a guided in-app camera, not a gallery upload, so there is no room to substitute an old photo or a stock image.

Step two, gate the return. If the customer later requests a return, the same guided capture flow triggers again, this time before the reverse pickup is even scheduled. No return AWB gets generated until the customer submits that return-side video. This alone filters out a chunk of casual fraud, because someone planning to send back a swapped item now has to film the actual item they are shipping, on camera, tied to that specific order ID.

Step three, compare and tier risk. This is where the two video captures actually get put to work. The system extracts frames from both videos, runs OCR and visual matching against the delivery record, and flags the order as low, medium, or high risk based on how well the return matches what was originally delivered. A pristine, matching return sails through. A mismatched print, a missing tag, or a completely different product gets flagged before it ever wastes a warehouse team's time.

Step four, warehouse QC and decision. The final call still sits with a human, your ops or QC team, but now they are looking at the item with full context: what left the warehouse, what the customer's own camera recorded at delivery, and what came back. Approving a refund or rejecting one with attached evidence is a completely different conversation than approving one on gut feel.

Notice what does not change in this flow. Couriers still just carry parcels, exactly as they do today, no new app, no training, no added liability for Shiprocket or Delhivery's delivery staff. Customers are not burdened with anything more complex than tapping a WhatsApp link and following on-screen prompts for ten seconds. The friction sits entirely on the side of the behaviour a brand actually wants to discourage.

What this looks like in the numbers

The brands piloting this kind of evidence chain are not seeing marginal improvements, they are seeing structural change in how disputes resolve. One apparel operations team described cutting disputed returns by more than two thirds within ninety days, simply because most "your word against mine" arguments stopped being arguments at all. The evidence made the decision for them.

Electronics and footwear sellers report a similar shift, and for good reason, these are the categories where swap fraud and wardrobing hurt the most in absolute rupee terms. A worn pair of sneakers or a swapped phone charger is expensive enough per unit that even a handful of fraudulent returns a month adds up to real money. Being able to show, frame by frame, that the pair returned does not match the pair shipped turns a slow, uncomfortable back-and-forth into a same-day decision.

There is a second, quieter benefit that gets less attention than the fraud-prevention headline: this same evidence chain protects genuine customers just as much as it protects the brand. When a warehouse mishandling error or a transit issue causes a legitimate complaint, the same delivery and return videos show exactly where things went wrong, and the brand can approve that refund immediately instead of dragging an innocent customer through a dispute process they did not deserve. Trust cuts both ways when there is actual proof involved instead of assumptions on either side.

The bigger shift this points to

Return fraud in Indian D2C is not going to shrink on its own. As COD volumes grow, as fashion and footwear categories keep expanding, and as fraud tactics get more sophisticated (AI-generated damage photos being the newest wrinkle), the brands that win are going to be the ones who stopped treating returns as an unavoidable cost centre and started treating them as a process that can actually be verified, end to end.

The old playbook of stricter return windows, non-refundable sale items, and hoping the warehouse team catches the obvious cases is not scaling. It punishes honest buyers with friction while leaving determined fraudsters mostly unbothered, since a stricter window does nothing to stop someone from wearing an item once inside that window and sending it back.

An evidence-first approach flips that. It adds almost no friction for honest customers, who are simply asked to record a ten-second video they were probably going to unbox on camera for social media anyway. It adds a lot of friction for the small group of repeat offenders who account for a disproportionate share of fraud losses. And it gives operations and QC teams something they have never really had in Indian D2C: an actual paper trail, on video, for every disputed order.

If you are running a D2C brand on Shopify with Shiprocket or Delhivery in your stack and returns are eating into margin in ways you cannot fully explain, the honest first step is simply looking at how many of your "resolved" disputes were resolved on evidence versus resolved on trust. For most brands, that number is uncomfortably close to zero. Building the evidence chain back into the process, from delivery capture through the return gate to the final warehouse decision, is what closes that gap.

Vefri was built specifically for this problem, connecting directly into Shopify, Shiprocket, Delhivery, and WhatsApp so brands can start capturing that evidence without ripping out any part of their existing stack. No custom courier app, no retraining delivery staff, just a verifiable record at the two points that actually matter: what left your warehouse, and what came back.

Questions ops teams usually ask before switching

Does this slow down delivery or annoy customers? Not in any way brands have reported so far. The capture link only asks for a short video before the package is opened, sent through WhatsApp, which most Indian shoppers already check within minutes of a delivery notification. It adds seconds, not friction that would make someone abandon a return or complain about the process.

What happens if a customer just refuses to record the video? The return gate holds the pickup until the video comes in, which in practice nudges compliance because customers who genuinely want their refund have every reason to complete a ten-second recording. The ones who push back hardest tend to be the ones with something to hide, which is itself a useful signal for the risk tier.

Do we need our courier partner to agree to anything? No. This is one of the more underrated parts of the setup. Because verification happens entirely on the customer's phone and inside your own warehouse, there is nothing for Shiprocket, Delhivery, or any other courier to integrate on their side beyond the order status webhook they already send. That removes an entire category of vendor negotiation that has killed similar initiatives in the past.

Is this only useful for high-fraud categories like fashion and electronics? Those categories see the fastest payback because average order values and return rates are both high, but the same logic holds for any category where "the product came back different from how it left" is even an occasional problem. Home decor, beauty, footwear, and accessories brands are all running some version of this today.

How long does it take to get set up? Since the integrations plug into tools most D2C brands already run, most teams are capturing evidence on live orders within days of connecting Shopify and their courier feed, not months into a drawn-out implementation.

Returns are not going away, and honestly they should not. A generous, low-friction return policy is a genuine competitive advantage for D2C brands competing against marketplaces. The goal was never to make returns harder. It was to make the ones worth questioning actually questionable, with proof instead of a guess, and let everyone else sail through the way a good return policy is supposed to work.