Bracketing: when buying five sizes is normal and when it's abuse

There's a customer in your Shopify admin right now who just placed an order for three dresses. Same style. Sizes 6, 8, and 10. They know they're only keeping one. The other two are coming back.

This is bracketing. Amazon and Zappos trained a generation of apparel shoppers to do it, and it migrated into DTC about five years ago. It's not fraud. It's not even unreasonable. The reason people bracket is that online fit is genuinely hard and your size chart was probably written for models with a different body than theirs.

But bracketing at scale, done by enough customers, has the same effect on your P&L as wardrobing. Three returns for every net sale. Forward shipping, return shipping, warehouse labor, repackaging, occasional item damage, and eventually an item retired from stock because it's been in too many envelopes.

The question for any apparel store above a certain volume isn't whether bracketing happens. It does. The question is how to tell normal bracketing apart from the shape of it that costs you real money.

The profile of healthy bracketing

Before drawing a line, it helps to name what the healthy version looks like, because you don't want to accidentally train policy against it.

A customer buys multiples of a new style they haven't tried before, keeps one, returns the rest within a reasonable window, and re-orders the size that fit when the next drop arrives. Their return rate stabilizes around 40% to 60% on first purchases and drops into the single digits once they know their size in your fit. Their average order value is healthy because they bought multiples. Their lifetime value is above average because once they find their size they become repeat buyers.

This customer isn't a problem. They're a signal that your size chart needs more information, not that your return policy needs tightening.

The profile of abusive bracketing

The shape that costs you money is different in four specific ways.

1. The return rate doesn't decline. A bracketer who's genuinely learning your fit should need fewer brackets per order over time. A customer who's still returning 60% of their items on order number eight isn't learning. They're either treating your return policy as a try-at-home service indefinitely, or they're returning items regardless of fit because they've decided they don't want them.

2. The items returned are unrelated to the items kept. Healthy bracketing keeps the size that fit. Abusive bracketing keeps the item they wanted and returns the items they didn't. Different colors, different styles, higher-price versions of items they never intended to keep. They're using bracketing as a free swap.

3. The bracket spans non-fit variables. A customer ordering the same dress in 6, 8, and 10 is bracketing size. A customer ordering the same dress in black, navy, and burgundy is bracketing color preference, a different category, because color is a decision they could have made before ordering. You didn't cause the indecision. You just absorbed the cost of it.

4. The return condition is inconsistent with "didn't fit." Items that come back with signs of wear or extended use aren't bracket returns. They're wardrobing using the language of bracketing. The customer claims they bracketed because a simple "this doesn't fit" is socially cleaner than "I wore this to brunch."

The economics, specifically

Before getting to what to do about it, be honest about the numbers, because the instinct to crack down can cost more than the problem it solves.

Say you sell apparel at an average order value of $150 with 55% gross margin. A healthy bracketer orders three items, keeps one, returns two. Your revenue: $50 kept at 55% margin is $27.50 gross. Your costs: two items shipped out and back (call it $18 total in shipping if you absorb both ways), warehouse handling (around $6), and a small probability of item damage ($1 expected). Net contribution around $2.50.

That's bad. Not catastrophic (you're still profitable), but the margin on the customer is a tenth of what it would be if they'd ordered the right size the first time.

Now say you tighten the return window, or introduce a restocking fee, or limit the number of returns per order. Your healthy bracketer feels friction. Some portion stops ordering at all. The ones who do order, order fewer items at once, which means they're more likely to get the wrong size on their first purchase, and they're now a one-item customer with a cancelled order instead of a three-item customer with a net-one sale.

The math on policy changes in bracketing is rarely as favorable as the initial pitch suggests. Every friction point costs you some percentage of the customers you wanted to keep.

What actually works

The intervention that survives real-world testing is precision bracketing limits. Not a ban, not a fee, not a window reduction. A specific cap on the number of size variants of the same SKU a customer can order in a single purchase, scaled to the customer's history.

• First-time customers: allow up to three size variants of the same item. They don't know your fit yet. You want their business.
• Repeat customers with a stable size history: limit to two variants, with a soft warning at checkout. "You've ordered size 8 in five previous purchases. Need three sizes this time?"
• Customers with a high-return-rate pattern over six months: single-variant only at checkout, with a customer-service override available on request.

This policy is invisible to 95% of your customers (they weren't bracketing anyway, or they're within the limits). It slightly frustrates the 4% who are healthy bracketers but have developed the habit of ordering four sizes out of caution. And it meaningfully constrains the 1% whose behavior is genuinely abusive, because those customers build up a return-rate signature that the system recognizes.

The second intervention, complementary to the first, is better fit information at the point of sale. Bracketing is caused by fit uncertainty. Reduce the uncertainty:

• Customer reviews that include the reviewer's measurements and which size they chose
• A "runs small / large / true" indicator pulled from your actual return data, per SKU
• A fit-finder quiz that asks for measurements once and remembers them
• Clear photos of the same item on models of different body types, with sizes labeled

None of this eliminates bracketing. It reduces the noise floor, which means the customers who are still bracketing heavily are more likely to be the ones you'd actually want to limit.

The takeaway

Bracketing is legitimate customer behavior in response to a real problem (online fit uncertainty) that apparel retailers have spent twenty years not fully solving. Treating all bracketing as abuse punishes customers for a gap in your product experience, and the backlash costs more than the bracketing.

The minority of bracketing that is abusive has a distinctive shape. Persistent high return rates, unrelated-variable bracketing, asymmetric returns relative to fit, wear signs on returned items. That minority is catchable with the same kind of customer-level pattern analysis that catches wardrobing. Once you can identify them, the right response is usually a precision cap rather than a broad policy change.

The stores that handle bracketing well do two things at once. They reduce the cause (better fit information) and they scope the enforcement (per-customer variant limits). The stores that handle it badly pick one crude lever and pull it, and spend the next quarter wondering why their conversion rate dropped.