The SKU concentration rule: a handful of products drive most of your bracketing

September 26, 2026

Most apparel brands treat bracketing as a customer problem. The data usually says something different: it is a product problem, concentrated in a small slice of the catalog. When you rank SKUs by bracketed orders, the curve is brutal. A handful of products, often fewer than one in twenty, accounts for the majority of multi-size orders and the return cost that comes with them.

This matters because it changes the fix. If bracketing were spread evenly, you would need policy changes and scoring across the board. When it is concentrated, you can fix most of the damage by fixing a short list of products: their size information, their fit guidance, or in some cases the products themselves.

How to run the ranking

Pull twelve months of orders and returns. For each SKU, count orders that contained two or more sizes of that SKU, then divide by total orders of the SKU. That is the bracketing rate per product. Sort the catalog from highest to lowest and look at the cumulative share.

The pattern we see across brands is remarkably consistent. The top 5 percent of SKUs by bracketing rate typically account for 40 to 60 percent of all bracketed orders. The top 10 percent accounts for roughly two thirds. The long tail of the catalog, the hundreds of products that sell steadily and fit predictably, barely brackets at all.

Two refinements make the ranking honest. First, require a minimum order volume before a SKU qualifies. A product with four orders, two of them bracketed, is noise. Second, look at bracketing rate alongside absolute volume. A product with a modest bracketing rate but enormous volume can still cost more than a high-rate niche item. Rank by rate to find the fit problems; rank by volume to find the dollar problems. The overlap of those two lists is your priority queue.

What the top of the list usually looks like

The products at the top of the bracketing ranking share recognizable traits. New silhouettes with no established fit reputation bracket heavily, because customers have no history to calibrate against. Products with inconsistent grading across sizes, where the jump from medium to large is bigger than the jump from small to medium, invite multi-size ordering as insurance. Categories with subjective fit, like relaxed denim or oversized outerwear, bracket because the size chart cannot describe drape.

One more common entry: the bestseller that outgrew its fit content. A product page launched two years ago with three fit photos and a generic size chart now drives half the brand's bracketing, because nobody refreshed the guidance as the product scaled. The product did not get worse. The information around it just stopped keeping up.

What to fix, product by product

Work the list from the top and give each SKU one of four treatments. The cheapest is better fit content: garment measurements, model try-on notes with real heights and sizes, and customer reviews that mention fit. This works when the product actually fits fine and the problem is uncertainty.

The second treatment is fit guidance keyed to the product's actual keep data. If the data shows that customers who order a medium keep the large 70 percent of the time, say so on the page. Fit recommendations earn trust fastest when they come from observed behavior rather than a generic chart.

The third is a targeted nudge at checkout or in the cart: when a customer adds two sizes of a high-bracketing SKU, a prompt suggesting the historically kept size can cut the second unit from the order. This is where product-level scoring pays for itself, because the intervention is surgical instead of a blanket policy change.

The fourth, reserved for the worst offenders, is a product decision. Some products are graded so inconsistently that no amount of guidance fixes the bracketing. If a SKU sits at the top of the ranking quarter after quarter despite better content, the fit itself is the problem, and the margin lost to its returns should be part of the line-planning conversation.

Why this beats customer-level work alone

Customer scoring and product ranking are complements, not substitutes. Scoring tells you who to watch. The SKU ranking tells you where the behavior starts. Fixing a fit-problem product reduces bracketing across every customer who buys it, including honest shoppers who were bracketing defensively because they could not trust the size information.

There is also a sequencing argument. Product fixes are one-time projects with compounding returns: you fix the page once and every future order benefits. Customer-level enforcement is ongoing work. Do the product work first, then score what remains. Most brands find the remaining bracketing, the kind driven by habitual behavior rather than fit uncertainty, is far easier to manage once the product noise is out of the data.

Run the ranking quarterly. Products move: new launches enter at the top, fixes push old offenders down, and seasonal categories cycle through. The list is a living priority queue, and the brands that treat it that way spend their fit budget where the bracketing actually lives.