Treat Returns As a Structural Cost

If your returns rate has crept up over the past year and you can’t point to a single clear cause, you’re not imagining it. You’re watching a structural shift play out in your own numbers.

The Numbers Are Bigger Than Most Merchants Assume

A new eMarketer report puts real shape around what a lot of ecommerce operators have been sensing anecdotally. Ecommerce returns are on track to make up almost 48% of all US retail returns by 2029. Total returns across all of retail are projected to hit an estimated $951.36 billion.

Returns are becoming one of its defining costs, scaling right alongside the channel itself as online shopping keeps pulling share away from physical stores.

For years, returns got treated as a customer service line item: a cost of doing business, something to manage in the background while the real strategic conversations happened around acquisition, conversion, and retention. At this scale, returns are a margin lever, and increasingly, a strategic one.

It’s worth understanding why this shift is structural rather than cyclical. Physical retail has a built-in filter that ecommerce doesn’t: a customer holding a product, trying it on, or seeing it in person before buying. Online, that filter disappears. Shoppers routinely buy multiple sizes or colors of the same item with every intention of returning what doesn’t work, a practice retailers now widely refer to as bracketing. As online penetration keeps climbing across categories that used to be dominated by in-store purchases, apparel especially, that built-in filter keeps shrinking, and the return volume that used to get absorbed by the fitting room now shows up in your reverse logistics costs instead.

The Tension Most Merchants Are Already Feeling

Fees and shorter return windows can slow the bleeding on returns, but they come with real risk to conversion and loyalty.

Customers have been trained by a decade of ecommerce competition to expect generous, frictionless returns as a baseline. Tighten that experience at the wrong moment, in the wrong category, or without the data to back it up, and you’re potentially creating a conversion and loyalty problem to replace it.

The instinct when returns spike is to reach for the blunt lever nearest at hand: shorten the window, add a restocking fee, make the return flow slightly more annoying to complete. Those changes can move the aggregate return rate. What they don’t do is tell you why customers were returning in the first place, which means you’re paying the conversion and loyalty cost of friction without necessarily fixing the underlying problem driving the returns.

This is exactly why ecommerce returns strategy can’t be handled as a policy tweak made in isolation. A return window that gets shortened company-wide because returns are “too high” might be masking the fact that one specific product category, one specific vendor’s fit blocks, or one specific description gap is driving the bulk of the volume. Blunt policy changes fix the aggregate number while doing nothing about the actual cause, and sometimes make the customer experience worse for the categories that were never the problem in the first place.

Think about what that looks like in practice. If 70% of your returns are concentrated in one apparel category with a known fit inconsistency, and you respond by shortening your return window storewide, you’ve just added friction to every other category that was performing fine, while doing nothing to fix the fit issue actually driving the volume. The customers most likely to notice and object are the ones in categories that didn’t cause the problem.

Treat Your Ecommerce Return Policy Like a Margin Lever

Your return policy should be informed by your own return-rate and reason-code data, not by industry benchmarks or a gut sense that “returns feel high right now.”

That means a few things in practice:

Reason codes need to be granular enough to act on

“Didn’t fit” and “not as described” are two very different problems with two very different fixes. If your reason-code data is too coarse to distinguish them, you’re not actually diagnosing anything. A vague reason code doesn’t just make analysis harder, it actively hides where the money is going. “Didn’t fit” might point you toward a sizing chart problem you can fix in a week. “Not as described” might point you toward a photography or copywriting gap that’s costing you conversions even among customers who don’t return the item. Lumping both into a single “customer changed mind” bucket erases that distinction entirely.

Category-level analysis matters more than company-wide averages

A softgoods brand’s return problem and a hardgoods brand’s return problem rarely look the same, and even within one catalog, categories can behave very differently. Policy changes calibrated to an overall average risk overcorrecting in categories that didn’t need it. A footwear category with genuine fit variability across brands is a fundamentally different problem than a home goods category where most returns come from shipping damage. Treating both with the same blanket policy solves neither.

Reverse logistics costs need to be part of the conversation

The cost of a return is more than the refunded revenue. It’s restocking, inspection, repackaging, and in some cases, product that can’t be resold at full value at all. A returns strategy that only looks at the return rate and ignores what happens after the product comes back is looking at half the picture. Two products with identical return rates can have wildly different actual costs to the business, depending on whether that returned inventory goes back on the shelf at full price or gets marked down, liquidated, or written off entirely.

Why This Requires Cross-Functional Ownership

One reason ecommerce returns strategies tend to stay stuck at the policy-tweak level is that no single team owns the full picture. Customer service owns the return experience and the reason codes customers select, but rarely has visibility into the margin impact of those returns once inventory comes back. Merchandising and product teams own sizing, fit, and description accuracy, the root causes behind a large share of avoidable returns, but often aren’t looped into returns data at all unless something becomes a visible problem. Finance feels the margin impact most acutely but usually sees it as an aggregate number on a P&L, disconnected from the specific product or category driving it.

That fragmentation is exactly why blunt, storewide policy changes are so common. It’s far easier for one team to adjust a policy setting than it is to coordinate a cross-functional diagnostic effort involving reason-code data, category performance, and reverse logistics costs. But the fragmented approach is also why so many ecommerce returns initiatives quietly fail to move the number: they treat a systemic, cross-functional cost as if it were a single-team fix.

Where to Start

If your team hasn’t looked closely at return reason codes in a while, or if your current return policy was set based on a general sense of what felt reasonable rather than a look at your own data, that’s the place to start. Pull the reason-code breakdown, segment it by category, and see where the actual concentration of returns lives before deciding what, if anything, needs to change about the policy itself.

A practical sequence looks something like this. Start by auditing your reason codes themselves, before touching policy at all. If your current codes are too broad to distinguish a fit problem from a description problem, fix that first. You can’t diagnose what you can’t measure.

From there, segment the data by category rather than looking at a single company-wide return rate. This is usually where the real story shows up. A blended average of 22% often turns out to be a handful of categories running at 40% dragging down a much healthier baseline elsewhere.

Once you know where the volume is concentrated, separate root-cause fixes from policy fixes. A sizing chart problem gets solved by fixing the sizing chart, not by shortening the return window for every customer buying that product. A description accuracy problem gets solved by better photography and copy, not by a restocking fee. Policy changes should be reserved for the residual volume that root-cause fixes can’t touch.

Finally, loop in reverse logistics costs before finalizing any policy decision. A category with a high return rate but low reverse logistics cost, because returned product goes right back on the shelf, is a very different business problem than a category with a moderate return rate but high reverse logistics cost, because returned product is difficult or impossible to resell.

For a lot of merchants, that diagnostic work, untangling return reasons, mapping reverse logistics costs, and figuring out where policy changes would actually move the needle, ends up being a bigger lift than a quick settings change in admin. If that’s where you’re at, this is exactly the kind of work our Returns Reduction Sprint is built for.

Returns aren’t going away, and at this point, they aren’t shrinking either. The merchants who treat them as a data problem worth solving, instead of a cost to quietly absorb, are the ones who’ll come out ahead as the numbers keep climbing.