Assortment optimization for small-format stores

Assortment optimization for small-format stores means choosing the smallest set of items that still covers what shoppers come in for. In a large supermarket, an extra item is one more slot on a long aisle. In a corner store with one shelf for a category, every item has to earn its space, and a poor choice pushes out a better one.

The data work isn't complicated. It's mostly a matter of asking the right question of each item: if this disappeared, would shoppers buy something else here, or leave without it?

Why is small-format different?

Space, mainly. A small store might have one cooler door for a beverage subcategory, or a few feet of shelf for household cleaners. That limits the number of items and the facings per item. A slow seller that would be harmless in a big store actively costs sales in a small one, because it occupies space a faster item could fill.

Small stores also tend to serve quick trips. Shoppers often want a specific thing, now. If the item they came for isn't there, a substitute has to be obvious and nearby, or the sale is lost. Trade groups such as NACS, the convenience store industry association, publish background on how small-format retail operates.

Which metrics matter most?

Three do most of the work.

Velocity per store carrying shows how fast an item sells where it's available. It's the fairest comparison between items with different distribution.

Contribution to the category shows how much of the category's sales an item accounts for. A mid-velocity item that's widely carried can matter more to the total than a fast item in a handful of stores.

Transferable demand estimates how much of an item's sales would move to other items if it were removed. Items with close substitutes in the set have high transferability and are safer to cut. Items with no close substitute, such as a specific size or a distinct flavor, have low transferability. Removing them loses sales outright.

The third is the hardest to measure. The cleanest way is to look at stores where an item was delisted or went out of stock for a while, and see what happened to the rest of the category there.

A fourth consideration, which the data can't fully answer, is the store owner's view. Owners of independent stores know their regulars. An item that looks weak on paper may be one a few loyal customers ask for by name.

How do you handle differences between stores?

Group stores that behave alike. Store size, location type and the mix of categories they sell are common grouping factors. An assortment that fits a store near a transit stop may not fit one in a residential block.

Avoid building groups on assumptions about who shops there. Group on what the store sells, which the data shows directly. The goal is a small number of assortment versions, each fitting a cluster of similar stores, instead of one list for all or a custom list for each.

What would you do with this on Monday?

Pick one category. Rank its items by velocity per store carrying and by contribution. Flag the bottom few on both lists. For each flagged item, ask whether it has a close substitute in the set.

Items that rank low and have close substitutes are the first candidates to remove or reduce. Test the change in a few stores before rolling it out, and compare category sales in those stores against similar stores that didn't change. Category management practice treats this as an ongoing cycle, and NRS Insights reporting can help you see how the category moves around your test.