AwardFlagship wins the 2026 Hilldun Business Innovation Award Read the announcement
Loading...
Inventory Management

How To Group Stores For Assortment Planning

How to Group Stores for Assortment Planning

Two stores each sold $240,000 of footwear last spring. The first would happily have traded a wall of occasion shoes for more walking sandals. At the second, customers kept asking whether the dressier styles came in another color. On a departmental spreadsheet the two are twins, and a buyer working from a store ranking alone would probably send both the same range. On a Saturday afternoon, they are running different businesses.

Store clustering gives an assortment planner a manageable way to act on those differences. A cluster is a group of locations that can usefully share a buying decision, and “usefully” does a lot of work there. A map can show a beautiful pattern of coastal and inland stores. If the pattern changes neither the styles a merchant buys nor the space given to them, it has very little to contribute at line review.

Start With The Rail That Will Actually Change

Imagine buying a spring footwear range for 24 stores, with room for eight sandal styles in each. Six are established lines the brand wants everywhere, which leaves the buyer discretion over the last two spaces. One possible range adds two technical walking sandals; another adds two dress styles. Now the analytical question has a shape: can the team identify stores where one use of those two spaces is likely to earn more full-price margin than the other?

Go back to the two $240,000 stores. Suppose each recorded 1,000 full-price sandal sales over comparable weeks when both families were well represented. At one store, 620 pairs were walking sandals and 380 were dress styles. At the other, the split was 310 and 690. That puts their walking shares at 62 and 31 percent, a difference worth investigating even if winter boots or a single premium brand accounts for most of each location’s annual footwear turnover. Clustering for a particular category keeps successful winter business from dictating what goes on a sandal fixture in May.

The share is only a starting point. Find out whether one location sold walking sandals at a higher average price, whether it had twice the space for them and whether its most popular dress size spent a month unavailable. Units help explain customer choices. Margin dollars help determine whether those choices justify a space. Selling weeks, presentation capacity and size availability tell you how much opportunity customers really had to choose. Looking at all five measures may change where the buyer places either store.

Watch For The Range Teaching You Its Own History

Last spring’s merchants supplied the merchandise that generated last spring’s data. A store that received two dress styles and six walking styles had more ways to register walking demand. The unbought dress sandal left no transaction behind. Historical mix can therefore preserve years of conservative allocation under an impressive new cluster name. Start comparisons with periods and products available in several locations. Where comparable exposure is poor, give the store an explicit tryout before assigning it a permanent identity.

How to Group Stores for Assortment Planning

The web channel offers clues, provided its role is understood. A high rate of local pickup on dress sandals can justify investigating unmet demand on the floor. Check where the online order originated and whether the store merely served as a convenient pickup point. Conversely, an apparent run on walking sandals may have been a ship-from-store routing decision made hundreds of miles from the shopper. A conversation with the manager can reveal that a display moved beside the entrance midway through the season. Each observation changes the question to ask of the numbers. None, by itself, warrants a new store group.

Each group also has to support its own buy. Four stores might share a taste for premium dress sandals but lack the combined sales to cover the vendor’s minimum in two colors and a full size run, in which case the proposed range exists only in a deck. The merchant can narrow the colors, use a common core with one local option, or test those styles in a subset of stores where they will get enough depth. A cluster has to survive both customer preference and the economics of making a buy.

Give The Old Range A Fair Chance to Beat The New One

A practical trial begins with comparable stores where the evidence favors a change. Within those stores, create matched pairs on past category performance, fixture space and weeks of availability. Where feasible, randomly give one store in each pair the local range and let the other keep its usual selection, with the same number of style slots and a comparable inventory investment. Choose the decision metric and selling window before receipts arrive. Otherwise a team keen on localization can declare victory on whichever number looks best after the fact.

For the sandal trial, full-price gross margin per fixture-week would be a sensible primary measure. Imagine the test stores generate $9,600 of that margin across 24 fixture-weeks, or $400 per fixture-week, while the comparison stores deliver $9,120 across the same exposure, or $380. The observed difference is $20 per fixture-week. Before committing the chain, ask whether the test stores generally ran ahead by roughly that amount anyway, whether one outlier explains the gap and whether unsold sizes have accumulated there. A preliminary gain in full-price margin can disappear when seasonal stock is marked down.

Look beyond the two trial slots. A new walking sandal might sell well by diverting customers from a more profitable core style already on the same fixture, and counting only the new pair would make the change look more valuable than it was. Judge the whole fixture, examine the displaced styles and record the extra buying and handling effort. Persuading the buyer to manage a twelfth version of a range should require a stronger return than extending a version the team already knows how to execute.

Leave Room for A Store to Change Its Mind

An unusual event can temporarily alter the customer base. A long-running hotel renovation might drain one location of occasion shoppers for a season; a nearby transit closure might divert lunchtime foot traffic. Put an end date on any assignment that depends on an event like that. For a new store with scant history, choose a provisional group based on comparable space, offer and trading pattern, then identify which early observations could change the assignment. A store manager is far more likely to report contrary evidence if somebody can explain what it would take to move groups.

At the next line review, present the decision in the same physical terms the buying team used at the start: which six styles stay common, which two slots change, in which stores the change earned its space after markdowns, and where the original range should return. Framed that way, assortment planning lands back where customers experience it. The two $240,000 stores can remain peers on a finance report, while their sandal walls can finally reflect what each store’s shoppers came in to buy.