Return Reason Codes: The Data Layer That Separates Abuse From Bad Fit
Direct answer: Fix your reason codes before you fix your return rate. Replace vague options like 'did not like' with specific, mutually exclusive choices: too small, too large, not as pictured, quality defect, arrived late, and changed mind. Require the code at return initiation, not as an optional survey. Within a quarter, the data will show you which categories drive legitimate returns and which customers hide behind vague codes to mask abuse patterns.
Why raw return rates mislead
Apparel return rates vary wildly by category. Footwear and dresses run high because fit is uncertain; basics run low. A 50 percent return rate means something completely different for a denim brand than for a t-shirt brand. Brands that set a single company-wide threshold end up flagging their best customers in high-return categories while missing real abuse in low-return ones. The threshold has to be relative: flag accounts that sit several multiples above the median for their category and price band, not above an arbitrary company number.
Rate also ignores the kept items. A customer who orders ten, returns seven, and keeps three that fit perfectly is a profitable customer with a fit problem. A customer who orders ten, returns nine worn, and keeps one is a loss. The dashboard number is the same. The business reality is opposite. Any threshold that does not account for condition and keep-value will punish the wrong people.
The three signals that matter more than count
First, condition. Items returned with tags attached and no wear are shopping. Items returned worn, washed, stretched, or damaged are abuse, at almost any rate. Condition is the single most honest signal because it cannot be explained by sizing. Second, velocity. A customer who places and returns five orders in a week is behaving differently from one who does the same over five months, even at identical return rates. Burst patterns correlate with event-driven abuse like wardrobing far more than steady rates do.
Third, the reason-code pattern. Legitimate high returners cite fit and sizing. Abusive patterns show reason codes that do not match the merchandise: "defective" on items that arrive in perfect condition, "wrong item" on orders picked correctly, "did not arrive" on tracked deliveries. When the stated reason contradicts the warehouse evidence, the account deserves scrutiny regardless of its rate.
Enforcing in tiers
Start with the lightest touch that works. For accounts moderately above the norm, a quiet change: slower refund processing, exchanges emphasized over refunds, fit guidance inserted into their purchase path. Many serial returners are just uncertain shoppers, and guidance converts them into keepers. For accounts far outside the norm with clean condition, consider restocking fees disclosed at checkout, which change the economics without a confrontation.
Reserve hard enforcement for the clear cases: high rates plus poor condition, contradicting reason codes, or coordinated multi-account behavior. Warnings first, then restrictions, then bans, each documented with the evidence that triggered it. Keep the criteria behavior-based and private, and review the flagged cohort quarterly, because patterns drift and a threshold tuned in a heavy-return year will be too aggressive in a normal one. The goal is a return rate that reflects your real customers, not a war on the customers you have.