Start with the number we can defend
According to the National Retail Federation and Happy Returns, retailers expect 19.3% of online sales to be returned in 2025. Across retail overall, they estimate 15.8% of annual sales will be returned, totaling $849.9 billion in merchandise value. That is the clean starting point for merchants: roughly one in five online sales still comes back.
That does not mean one in five online sales is avoidable. It does mean the volume is large enough that even modest prevention has financial weight.
The avoidable slice is real, even if it is not one neat percentage
Some returns are unavoidable. If an item arrives damaged, the wrong product ships, or a legitimate defect shows up, the merchant should resolve it quickly and fairly.
But many returns are driven by softer causes: fit uncertainty, expectation mismatch, multiple sizes ordered for comparison, or simple hesitation after delivery. In apparel, PowerReviews found that 39% of consumers said they returned an item because they did not like the fit, while 28% said the item did not look the way they expected. Those are not policy problems. They are expectation and confidence problems.
That distinction matters because a fit-driven or preference-driven return is not the same operationally as a defect claim. One often needs a refund or replacement. The other may need clearer information, a better exchange path, or a reason to keep the item.
What the margin math looks like without hand-waving
Take a simple example. Imagine a store doing €1,000,000 in annual online revenue. If its return behavior roughly follows the 2025 NRF benchmark, then about €193,000 of sales value is at risk of coming back.
Now apply a conservative intervention assumption: the merchant prevents or resolves just 10% of that return-intent value before it becomes a normal return. That is €19,300 in gross merchandise value preserved.
If the merchant used a controlled incentive averaging 10% of order value to save those orders, the discount cost would be roughly €1,930. That still leaves most of the merchandise value intact, before counting any avoided handling, restocking, or support cost.
This is scenario math, not a benchmark. But it makes the right point: merchants do not need a dramatic save rate for pre-return intervention to be worth testing.
Why charging for returns is not the same as solving them
Retailers have become more aggressive about return policy controls, but friction alone is a risky lever. NRF and Happy Returns found that 82% of consumers say free returns are an important consideration when shopping online. In the 2024 version of the same research, 67% said a negative returns experience would discourage them from shopping with a retailer again.
So yes, merchants need discipline. But simply making returns more painful can push the problem upstream into conversion and trust. Good prevention does something different: it separates legitimate claims from low-conviction returns and gives each one the right path.
The opportunity is in the window before the RMA
Most return tools are built for speed after the customer has already decided to send something back. That matters, but it is late in the process. By that point, the merchant is mostly optimizing cost containment.
The more interesting question comes earlier: what happens in the moment between "I might return this" and "I have started a return"?
That is where merchants can still change the outcome. Order verification, structured reason capture, item-level selection, exchange options, and tightly controlled keep offers all belong in that window. If the reason is damage, route it into the normal return path. If the reason is a softer preference case, give the customer a better option before the warehouse inherits the problem.
A better framing for merchants
The most defensible argument is not "we lose 19% of revenue to avoidable returns." The data does not support that claim cleanly enough.
The better argument is this: nearly one in five online sales is still expected to be returned in 2025, and a meaningful share of those returns is driven by reasons merchants can influence before reverse logistics begins.
That is enough to justify a pre-return workflow. Not because every return should be stopped, but because some of them clearly should not have become returns in the first place.
Sources
This article uses published data from the National Retail Federation, Happy Returns, Narvar, and PowerReviews. For merchants who want to review the underlying numbers directly: