Return Fraud on Shopee & Lazada: Seller Playbook

A practical taxonomy of return fraud on Shopee and Lazada: the tell for each scam pattern, the evidence that wins the dispute, and how to catch repeat offenders.

AI-assisted content notice. This article was drafted with AI assistance and reviewed by a human editor at Hail Pilot before publishing.

Why Return Fraud Quietly Drains Your Margin

Every fraudulent return costs you three things at once, and only one of them is obvious. The product is the visible loss. The return shipping is the second. The third, the one most sellers never reconcile, is the marketplace transaction fee that stays deducted even after the sale is unwound.

On Shopee Singapore the transaction fee runs about 3.27% (a 3% base plus 9% GST on the fee, as of 2026 — check current rates). On Shopee Malaysia it is around 3.78% (3.5% plus 8% SST). When a refund is forced through on a genuine order, the escrow reversal usually handles the fee cleanly. When it is forced through on a fraudulent return, you have shipped a real product out, paid to bring a worthless parcel back, and in many cases eaten the fee on a sale that produced no revenue. A single mid-value order lost this way can wipe the margin on ten clean ones.

The quiet part is the compounding. A buyer who succeeds once learns your shop is soft, and the same tactic returns under a fresh order. Left unrecognised, return fraud stops being an incident and becomes a monthly line item. The fix is not to fight every return, most are honest, but to recognise the small share that follow a scam pattern and to have the counter-evidence ready before the dispute clock starts.

The Return-Fraud Patterns, and the Evidence That Beats Each

Almost all return fraud collapses into five patterns, and each leaves a specific tell. Learn the tell and you know which piece of evidence to reach for. The table maps them; the notes below add the operator detail.

Pattern	The tell	The evidence that beats it
Empty-box / partial-box	Return parcel weight far below the item; buyer opens the return the day it is delivered back to you	Outbound weigh-in slip + inbound weigh-in on receipt; packing video showing item sealed in
Item-not-received (INR) on a delivered parcel	Tracking shows "delivered" with a scan, buyer claims nothing came	Courier proof of delivery: scan timestamp, GPS/geotag, recipient signature or photo
Wrong / worn item returned (wardrobing)	Returned SKU differs from what shipped, or item shows use, missing tags, worn soles	Serial/IMEI or SKU photo at dispatch; unboxing/packing video; condition photos
Damaged-in-transit blame-shifting	Buyer claims arrived broken but the fragile item was under-declared or unbranded packaging	Packing video showing protective wrap; weighed parcel; courier handling terms
Coordinated / repeat abuse	Same buyer or address files returns across multiple orders; new account, high-value target	Your own risk history across orders; dispute notes tied to the case file

Empty-box and partial-box returns

The buyer files "item not as described" or a return request, ships back a box that weighs almost nothing (or contains a brick, a water bottle, an old cable), and times the claim so the refund releases before you inspect the parcel. The tell is weight. An item that left your warehouse at 480g does not come back at 60g.

The evidence that beats it is a weigh-in at both ends. Photograph or slip the outbound parcel weight at dispatch, and weigh the return the moment it lands, on camera. Pair that with a packing video that shows the actual item sealed into the box. Against a documented outbound-versus-inbound weight gap, "the seller sent me an empty box" does not survive review.

Item-not-received (INR) scam on a delivered parcel

Tracking says delivered, with a courier scan and often a drop-photo, and the buyer opens an item-not-received claim anyway. This is the item not received scam in its purest form: it bets that you cannot produce delivery proof faster than the platform's auto-refund fires.

You beat it with the courier's proof of delivery. The scan timestamp, the geotag or GPS point, and the recipient signature or doorstep photo are the record that decides these cases. Save them per order the day they post, because retrieving them from a courier portal two weeks later, mid-dispute, is where most of these losses actually happen. We go deep on this in the proof-of-delivery guide for INR chargebacks.

Wrong or worn item returned (wardrobing)

Two variants share one tell. In the swap, the returned item is a different SKU, a knock-off, or an older damaged unit the buyer already owned. In wardrobing, the exact item comes back used: dress worn once with the tag tucked in, shoes with scuffed soles, electronics with a different serial.

The counter-evidence is identity plus condition, logged at dispatch. For anything with a serial or IMEI, record it against the order before you ship, then a returned unit with a different serial is proven fraud, not opinion. For apparel and unserialised goods, a packing video and clear condition photos at send-out give you the "shipped new, came back worn" comparison. A returned SKU that does not match your dispatch photo is a wrong item returned claim you can win outright.

Damaged-in-transit blame-shifting

The buyer claims the item arrived broken and wants a full refund plus keep the goods. Sometimes it is real. The fraud version has a tell: the packaging was thin, the fragile item was under-declared, or the damage pattern does not match transit stress.

Your evidence is the packing video showing protective wrapping and the item intact at seal, plus the weighed, correctly declared parcel. That shifts the question from "did the seller pack carelessly" to "did the courier mishandle a properly packed parcel," which routes the liability to the logistics partner where it belongs.

Coordinated and repeat abuse

The costliest pattern is the buyer, or address, or device, that files a return on order after order. One refund looks like bad luck. Three across a quarter, all high-value, often on a young account, is a pattern. The tell only appears when you can see across your own orders rather than case by case. The counter-evidence here is your own history, which is where risk scoring earns its place. Our deep dive on detecting serial refund abusers covers the signals worth watching.

The Evidence That Actually Wins

Disputes are won by the evidence you captured before the return, not the argument you write after it. Five artifacts settle the large majority of return-fraud cases, and all five are cheap to produce at dispatch and expensive to reconstruct later.

A packing or unboxing video, filmed unbroken from empty box to sealed label, showing the correct item going in.
The outbound parcel weight, photographed or slipped, so an empty-box return contradicts a recorded number.
Courier proof of delivery: scan timestamp, geotag, signature or drop-photo, saved per order.
Serial or IMEI logged against the order for any item that carries one.
Timestamps on everything, because a complaint raised after 14 days of silence reads very differently from one raised on delivery day.

The artifacts matter, but organisation is what converts them into a win. A dispute is decided in a short window, and the reviewer weighs what you attach, not what you have filed loosely across a phone camera roll, a courier portal, and three chat threads. If it takes you two days to find the packing video, you have already lost the parcels where the auto-refund fired on day one.

This is the specific job Hail Pilot's evidence engine does: when you connect your Shopee, Lazada, Amazon.sg, or CSV data, it pulls the order, the chat log, the tracking, and your uploaded photos into one case file per dispute, so the proof is assembled the moment a return looks wrong rather than the night before a deadline. It does not fight the dispute for you; it makes sure the evidence is in one place, timestamped, and ready to submit. The dispute copilot can then draft the reply text for you to review before you send it.

How Disputing Differs on Shopee vs Lazada

Both platforms run a return/refund window and both lean toward the buyer on auto-decisions, but where your evidence goes and how fast you must move differ. Treat them as two separate playbooks rather than one.

On Shopee, buyer return and refund requests surface in the order's dispute flow, and you respond inside the return/refund case with your evidence attached to that order. The recurring trap is speed: many refunds are granted automatically, so the useful move is to have delivery proof and packing evidence ready to submit the same day a claim opens, not to assemble it afterward. For the evidence structure Shopee reviewers actually reward, see the Shopee dispute evidence guide.

On Lazada, the dispute and reverse-logistics flow has its own timelines and its own evidence-upload points, and the return often moves through Lazada's return process before the refund settles. The winning submissions are specific and platform-cited rather than a general plea. We wrote a full walkthrough on winning a Lazada dispute in Singapore that covers where the evidence goes and the windows to watch.

The common thread across both: platform windows are short and they do not pause for your weekend. The seller who keeps proof organised per order responds inside the window; the seller who reconstructs it after the fact misses the window and absorbs the loss. Exact windows and portal layouts change, so confirm the current rules in each platform's seller help centre before you rely on a specific number.

Spotting the Repeat Offender in Your Own Orders

The highest-value move against return fraud is catching the repeat offender inside your own order history, and doing it without punishing honest buyers. Most return fraud that scales is not a one-off stranger; it is a small number of buyers who found the tactic works and reuse it.

Pattern detection across your own orders is what surfaces them: the same buyer, address, or device filing returns at a rate far above your baseline, clustered on high-value items or young accounts. Consider a shop that sells mid-range electronics and sees three "arrived damaged, keep the item" claims in one month, all routing to the same delivery address under two different buyer names. No single claim proves fraud. The cluster does, and it tells you which future orders from that address deserve a packing video and a signed delivery before anything ships.

Two principles keep this fair and safe. First, cold-start neutrality: a buyer with no history is unknown and neutral, not guilty. New customers are your growth, and treating a first-time buyer as a fraudster is how you lose good revenue chasing a rare bad actor. Risk scoring raises a flag on a demonstrated pattern; it does not pre-judge a blank slate. Second, privacy: any cross-seller signal Hail Pilot uses is built on HMAC-hashed tokens, never raw personal data, and the platform is PDPA-aware by design. You are pattern-matching behaviour on your own orders, not maintaining a blacklist of people.

That distinction is the whole point. The goal is not to make returns hard for everyone, it is to make the specific fraud pattern expensive for the specific buyer running it, while every honest return stays frictionless.

Frequently Asked Questions

What counts as return fraud on Shopee or Lazada? Return fraud is any return or refund a buyer claims but is not entitled to: shipping back an empty or partial box, returning a used or wrong item, or claiming a delivered parcel never arrived. It is distinct from a genuine return, where the buyer sends back the actual item in the described condition. The line is evidence, which is why capturing proof at dispatch matters.

How do I prove an empty box scam? Weigh the outbound parcel at dispatch and record it, then weigh the return on camera when it lands. A documented gap between the shipped weight and the returned weight is hard evidence that the item was removed. Pair it with a packing video showing the correct item sealed into the box, and the empty-box claim does not survive review.

A buyer says the parcel never arrived but tracking shows delivered. What do I do? This is the item not received scam. Pull the courier's proof of delivery for that order: the delivery scan timestamp, the geotag or GPS point, and the recipient signature or drop-photo. Submit that inside the platform's dispute flow before the auto-refund releases. Save delivery proof per order the day it posts, because retrieving it mid-dispute is where most of these cases are lost.

Can I stop a refund once it has been auto-approved? Often you can only recover it by disputing after the fact, which is harder than blocking it before it clears. The reliable defence is speed: have your delivery proof and packing evidence attached to the order the same day the claim opens, so your response lands inside the review window rather than after the money has moved.

What is wardrobing and how do I fight it? Wardrobing is when a buyer uses an item then returns it as unwanted: a dress worn once, shoes with scuffed soles, electronics with a swapped serial. Beat it with identity and condition logged at dispatch. Record the serial or IMEI against the order for anything that carries one, and film a packing video plus condition photos so a used or mismatched return contradicts a documented shipped-new record.

Does chasing return fraud mean treating every buyer as a suspect? No, and it should not. A buyer with no history is treated as unknown and neutral, not guilty. Risk scoring flags a demonstrated pattern of abuse across your own orders; it does not pre-judge first-time customers. The aim is to make one specific fraud tactic expensive for one specific repeat buyer while honest returns stay easy.

Is the marketplace transaction fee refunded when a fraudulent return is reversed? Not always. On a clean order the escrow reversal usually handles the fee, but a forced or fraudulent refund can leave you having shipped a product, paid return shipping, and still absorbed the transaction fee, roughly 3.27% on Shopee Singapore or 3.78% on Shopee Malaysia as of 2026 (check current rates). That uncounted fee is why reconciling fees per order matters.

How do I catch a repeat return-fraud buyer? Look across your own order history for clusters: the same buyer, address, or device filing returns far above your normal rate, often on high-value items or young accounts. One claim is noise; a cluster is a pattern that tells you which future orders deserve a packing video and a signed delivery. Hail Pilot scores this across your connected channels using hashed tokens, never raw personal data.

Stop Losing Fraudulent Returns Alone

You cannot argue your way out of a return-fraud dispute after the fact; you can only submit the evidence you captured before it. Hail Pilot connects your Shopee, Lazada, Amazon.sg, and CSV orders, organises the packing video, tracking, and chat log into one case file per dispute, and flags the repeat-return pattern across your own orders while keeping unknown buyers neutral. See how the evidence engine and dispute copilot work on the features page, and read the platform-specific playbooks for Shopee and Lazada.

Stop eating the product, the shipping, and the fee on returns you could win. Start your free trial.

By Charlie Lee — Founder, Hail Pilot. Reviewed 2026-07-07.

Return Fraud on Shopee & Lazada: How Sellers Fight Back