For all the talk of DTC's "return to fundamentals" since the 2023 reset, the line that actually decides whether a digital-native brand makes money has barely moved. Returns — and the cascading costs they trigger — remain the largest unsolved problem in direct-to-consumer P&Ls, and most of the spreadsheets we've reviewed still understate the damage.

Operators we spoke to across apparel, footwear, and home goods describe return rates that sit somewhere between 18% and 30% on average, with peaks well above 40% in categories like women's denim and occasion dresses. The headline number is bad enough. The full cost stack underneath it is worse.

The cost stack is wider than the refund

Most finance teams still model returns as "refund value plus shipping." That's the visible tip. The real cost includes inbound logistics, inspection labor, repackaging, the markdown taken on resaleable units that re-enter inventory in B-grade condition, the write-off on units that can't be resold at all, and — increasingly — the carrying cost of inventory that gets stuck in returns processing for four to six weeks before it's available to sell again.

When you stack those layers, the all-in cost of a returned unit lands closer to 25-40% of the order value, depending on category. For a brand running 22% returns at a 30% all-in cost, that's roughly seven points of margin gone before you've touched CAC. For apparel brands at the high end of the curve, the math gets ugly fast.

The brands that survived the funding drought figured this out somewhere in 2023-2024. The ones still flailing are usually the ones where the CFO and the head of e-commerce are still arguing about whose number returns actually is.

What's working, and what isn't

The interventions that genuinely move the needle are unglamorous. According to retailers managing this transition, the highest-ROI moves cluster in three areas:

  • Pre-purchase fit and expectation tools. Size advisors that actually use post-purchase data (not just the pixel data on the page) have produced measurable drops in size-related returns — operators report 15-25% reductions on the SKUs where the tool is properly tuned. The catch: it only works if you feed it return reason data with enough granularity to distinguish "ran small" from "fabric weight felt cheap."
  • Charging for returns, intelligently. The Zara-style flat fee got most of the press, but the more interesting work is segment-based: free returns for high-LTV customers, paid returns for serial returners, with the segmentation actually enforced at checkout. Brands doing this well report 8-15% drops in return volume without measurable hits to acquisition. The brands doing it badly just annoyed their best customers and lost them.
  • Pushing returns to physical drop-off. Every operator we spoke to who runs hybrid (DTC plus retail partners or owned stores) is aggressively routing returns to physical locations. The unit economics on a store-handled return are roughly half the cost of a mail-back, and the in-store conversion on returns-into-exchanges runs 20-35% in most categories.

What isn't working, despite the pitch decks: AI-powered "return prevention" overlays that pop up at checkout asking customers to reconsider. The data we've seen suggests they reduce gross orders by more than they reduce returns. They also annoy people.

The resale and B-grade question

The second-order conversation, and the one that's quietly become more interesting, is what to do with returned inventory that can't go back to A-grade stock. Three years ago the default was a liquidator at 15-20 cents on the dollar. Now, the brands with the right infrastructure are running their own B-grade channels — outlet sites, "imperfect" SKUs, or partnerships with platforms in the resale space.

The margins on a self-run B-grade channel can recover 40-55% of the original wholesale value, versus the 15-20% from liquidation. The infrastructure cost is real, though, and the cannibalization risk on full-price is non-trivial. Most brands under $50M in revenue can't justify it standalone; the ones that crack it usually do it through a third-party operator that handles the channel for a take rate.

The honest read

Returns are not going to be "solved" by software. They're a structural feature of buying clothes and shoes on the internet, and any operator pitching you a 50% reduction is selling something. What the better-run DTC P&Ls have done is accept the structural floor, model it honestly, and squeeze the controllable layers — pre-purchase fit, returner segmentation, drop-off routing, and B-grade recovery.

The brands that did this work in 2023-2024 are the ones with margins now. The ones that didn't are the ones being rolled up.