Auditing Return Rates by Category: Where the Money Actually Goes

November 11, 2026

Split the return rate by category before you do anything else. A blended rate of 20 percent can hide a 45 percent rate on occasion wear and a 6 percent rate on basics, and the fix for each is different. The category split tells you where the money goes, and which SKUs to inspect first.

The blended rate is a lie

A store with a 20 percent return rate sounds like it has a returns problem. It might not. It might have an occasion wear problem, a footwear problem, and a basics business that barely returns anything. A blended conversion rate hides the truth in the same way, and the fix is the same: split the number before you act on it.

The split that matters is by category, then by SKU inside the category. Occasion wear, formal shoes, and event-driven categories run far above the store average, because they are the categories people buy to wear once. Basics run far below it, because nobody wardrobes a pack of socks. A dress with twice the return rate of similar items, and no complaints about fit, belongs on the inspection list, and it belongs there because the data says so, not because someone at the counter had a feeling about it.

The second split is by reason, and it is less useful than people think. The reason field is a survey of customer intent, and dishonest customers produce dishonest surveys. A sophisticated wardrober picks the most plausible benign reason, usually changed my mind or did not fit. The tell is not in the reason. It is in the pattern of everything around the return: the timing, the category, the account, and the condition of the item when it comes back.

How to run the audit

Start with product-level return-rate monitoring. Export returns by SKU for the last 90 days, and rank them by return rate, not by count. A SKU with a 45 percent return rate on 40 units is a bigger problem than a SKU with a 15 percent rate on 400 units, because the rate tells you the behavior and the count tells you the scale. Watch for SKUs whose rate has risen against the previous 30 days: a sudden rise is a signal, and it is the signal that catches a new wardrobing pattern before it becomes a habit.

Then add the timing. Returns clustered in the days after public holidays, bank holidays, wedding season, graduation, and long weekends are the pattern that matters. An item ordered a few days before an event and returned a few days after it is the shape of wardrobing, and it is visible in the data if you look for it. Automated monitoring typically watches returns inside a rolling 90 day window across locations, flagging accounts whose count or dollar volume passes a threshold, first with a warning to staff, then with denial at a higher threshold.

The third pass is the account. A customer with a 15 percent return rate concentrated in cocktail dresses is a very different signal than the same return rate spread across basics. A customer with 20 orders and 4 returns, where the returned items are the four highest-value pieces they bought, is almost certainly wardrobing. The pattern worth a flag is repetition: a third event category return, returned within 72 hours of a weekend delivery, from an account returning at triple your average rate.

What to do with what you find

The audit is not the punishment. It is the map. The SKUs with the highest return rates get the inspection first, the categories with the worst rates get the policy review, and the accounts with the clearest patterns get the conversation. The National Retail Federation put fraudulent returns at about $101 billion, with roughly 9 percent of returns fraudulent, and merchandise returns at $890 billion in 2024, about 16.9 percent of US retail sales. Those are big numbers, but they are not your numbers. Your numbers are in the category split, and they are the ones you can act on.

For the worst categories, the fixes are concrete: visible use tags that must be removed to wear the item, serialized and tamper evident, so removal ends return eligibility. Category specific windows, 14 to 30 days being standard, shorter for occasion wear. And photo upload per return reason, so the condition is recorded when the customer starts the return, not when the item arrives back at the warehouse.

And keep the audit running. A one-time audit finds the pattern that exists today. A monthly audit finds the pattern that starts next month, when the wardrobers who got flagged move to a new category. The auditing the true cost of free returns post covers what those returns cost once they are in the door. The auditing occasion wear returns after event weekends post covers the timing pattern in detail. This audit is the step that connects them: it tells you which categories to watch, and which SKUs to inspect first.

Frequently asked questions

What is a good return rate by category?

There is no single good rate, because categories differ so much. Basics run far below the store average, occasion wear and formal shoes run far above it. The useful number is the category rate against the store average: a dress with twice the return rate of similar items, and no fit complaints, belongs on the inspection list.

How do I find which SKUs have the highest return rates?

Export returns by SKU for the last 90 days and rank by return rate, not by count. Watch for SKUs whose rate has risen against the previous 30 days. A sudden rise is the signal that catches a new wardrobing pattern early.

Why can't I trust the return reason field?

Because the reason is chosen by the customer, and a dishonest customer picks the most plausible benign reason, usually changed my mind or did not fit. The tell is in the pattern around the return: timing, category, account history, and the item's condition.

How much do fraudulent returns cost retailers?

The National Retail Federation put fraudulent returns at about $101 billion, with roughly 9 percent of all returns fraudulent, against $890 billion of total returns in 2024, about 16.9 percent of US retail sales. Your store's number is in the category split, not the headline.