An empty box return is a return fraud tactic in which a shopper initiates a return, ships back a package that contains nothing (or a weight substitute such as paper, rocks, or a cheaper item), and collects the refund while keeping the original merchandise. The return looks complete to every automated system along the way: a label is generated, a carrier scan confirms movement, and the package arrives at the warehouse. The loss is only visible when someone opens the box.

For merchants, empty box returns combine physical loss, refund loss, and processing cost in a single transaction. For fraud and operations teams, they are the clearest example of why a carrier scan is not proof of a return.

Wyllo classifies empty box returns as a post purchase abuse pattern that typically follows a legitimate, approved order. Because the payment was good and the identity was real, checkout fraud tools never see it. It frequently appears alongside Fake Tracking ID (FTID) schemes, refund abuse, and product swaps.

By the numbers: Retailers in the NRF and Happy Returns 2025 Retail Returns Landscape named empty boxes as the second most common type of return fraud they encounter (65%), behind overstated quantity (71%) and ahead of counterfeit items (64%). The same research found that 9% of all returns were fraudulent against $849.9 billion in returned merchandise. Read the Happy Returns and Wyllo return fraud report

Ecosystem Overview

Empty box returns exploit the handoffs between systems that each confirm one thing and assume the rest:

• Returns management platforms that generate labels and approve returns based on order data
• Carrier tracking systems that confirm a package moved, not what it contained
• Warehouse intake and inspection workflows that run after the refund decision in many setups
• Customer support tools where “where is my refund” pressure accelerates approval

The weak point is timing. When a refund is released on label creation, on first carrier scan, or on receipt before inspection, the empty box has already done its job. When the merchant responds by moving every return to refund upon inspection, the cost shifts to honest customers instead.

Within the Wyllo platform, Return Fraud and Abuse Prevention scores the return request before the refund, using order history, delivery outcomes, return patterns, and refund behavior across the customer’s full lifecycle. CX Support puts that signal in front of the agent handling the ticket.

Common Empty Box Return Patterns

The true empty box. The shopper ships the original packaging with nothing inside. Weight checks at intake catch this version easily, which is why experienced bad actors rarely use it more than once.

The weighted box. Paper, packing material, or a low value item is added so the package weight matches the original shipment. This defeats weight checks and requires the box to be opened.

The partial return. One item from a multi item order is returned while the refund is claimed for all of them, closely related to order missing some items claims.

The scan and refund play. An empty envelope is shipped only to generate a carrier scan that triggers an automated refund, overlapping with FTID tactics.

Empty box methods are traded in fraud communities as step by step playbooks, often with merchant specific notes on which refund policies release money fastest.

Supporting Insight

Empty box returns are among the fraud types that physical inspection is best at catching, which is exactly why they have driven so many merchants toward blanket inspection holds. The tradeoff is steep. Research on return policies shows that 71% of consumers are less likely to shop with a retailer again after a poor returns experience, and 81% review the returns policy before they buy (NRF and Happy Returns 2025). A policy designed to stop the fraudulent 9% lands on the honest 91%.

The behavioral view tells a different story. The Happy Returns and Wyllo report found that roughly 1% of shoppers are marked as risky returners and that they are 85 times more likely to be flagged than the general population. The risk is concentrated and identifiable, so the control can be too.

How Merchants Mitigate Empty Box Returns

Operational controls include weight and dimension matching at intake, photo documentation of opened returns, and routing high value or high risk returns through verification at the point of drop off, where item verification reduces fraud by at least 85%.

Policy and workflow controls focus on refund timing that flexes by risk rather than by rule: instant refunds for trusted customers, refund on receipt for a watch tier, and refund upon inspection only for the small population whose behavior has earned it.

Behavioral intelligence connects the return request to everything the merchant already knows about the shopper: prior returns, refund ratio, claim history, account linkages, and delivery outcomes. An empty box is rarely a first offense. The pattern is visible before the package ships.

Summary

Empty box returns turn a merchant’s own refund automation against it, collecting refunds for merchandise that was never sent back. Inspection catches the box; it does not catch the pattern, and applied to everyone it punishes the customers who were never the problem.

The durable fix is intent aware decisioning on the return request itself, so that friction lands only where it is earned.

Frequently Asked Questions

How common are empty box returns?

Retailers rank them among the most common return fraud types they see. In the NRF and Happy Returns 2025 research, 65% of retailers named empty boxes as a fraud type they encounter, second only to overstated quantity claims.

Does refund upon inspection stop empty box returns?

It stops the refund on that return, but it applies a two week hold to every honest customer as well, and it does nothing about the shopper’s next attempt. A targeted hold based on risk signals catches the same fraud at a fraction of the cost to the rest of the customer base.

How can a merchant tell an empty box return is coming?

By looking at the shopper, not the package. High refund ratios, repeated “item not received” or damaged on arrival claims, linked accounts, and a return cadence that repeats within weeks are the signals that precede an empty box, and they are visible when the return is requested.