Return rate is the share of an order book that comes back. The arithmetic is simple — returns divided by a matching denominator over the same period — but the denominator is where the metric stops being one number and becomes three. Returns over orders, units returned over units shipped, and returned dollars over gross revenue each answer a different operating question.
Reading the rate three ways
The three formulations sort by what they diagnose. Order-level (returns / orders) is the fulfillment and sizing lens — shipping errors, fit problems, expectation mismatches. Unit-level (units returned / units shipped) is the merchandising lens, weighted toward SKUs that cluster in multi-unit baskets. Revenue-weighted (returned dollars / gross revenue) is the margin lens: 5% on the priciest SKUs is a different P&L event from 5% on the cheapest. Most ops dashboards default to order-level; the revenue-weighted read is the one a CFO needs.
Three reads from the same returns
A brand runs 1,000 orders, 1,200 units shipped, and $200K gross revenue in a period. 60 orders come back, containing 80 units worth $20K. Order-level: 6.0%. Unit-level: 6.7%. Revenue-weighted: 10%. “Our return rate is X%” turns on which one is being quoted.
The window and its tail
The return window is the second decision. Policy windows commonly run 30, 60, or 90 days, and customer behavior lags the policy — measuring at a window shorter than the policy truncates the late tail. A brand with a 90-day policy measuring at 30 days reports a different metric than one with a 30-day policy doing the same, even when the headline matches.
Category recovery differs in direction more than in fixed benchmarks. Apparel decisions extend over weeks as customers reconsider sizing or move past a season; furniture and durables resolve faster, because the customer either keeps the item or initiates a heavy reverse-logistics move at much higher per-return cost.
Not refund rate, not defect rate
Return rate is not refund rate. Brands that issue store credit, exchanges, or replacements have a return rate higher than their refund rate, because refund rate is the cash-out subset; conflating the two understates margin impact at brands with strong store-credit programs. Nor is it defect rate — damaged-product returns feed defect rate, but most DTC returns are fit, taste, or expectation mismatches.
Why the line earns its place
Return rate is the gap between gross revenue (what marketing reports) and net revenue (what funds the business). Most growth dashboards read against gross demand, so CAC and ROAS overstate efficiency at any brand with a non-trivial return rate — acquisition cost is being paid to ship product that comes back. The fix is to run contribution margin, payback period, and the full gross-to-net waterfall on returns-adjusted revenue. The operator move underneath is to instrument returns by reason code and SKU and route the highest-return SKUs back to merchandising and product copy. Better pre-purchase information is the consistent reduction lever; stricter policy suppresses reported returns by transferring the cost to chargebacks, repeat rate, and brand trust.