Store Clustering

Store Clustering is a popular analytic technique in the retail domain. It is employed to build relevant store segments which are homogenous in certain behavioral aspects and can be targeted using the same marketing strategy. Store Clustering can be done on an existing segmentation, for example, Customer lifestyle segmentation or Trip segmentation.

A store segmentation can help a retailer group together stores which have a similar customer-base (For example, a store in the heart of the city may cater to customer-base similar to a store in another area of town which may be different from the store in the suburbs). The retailer can then have differentiated marketing strategies for each of the store segments appealing to different sets of customers.

Alternately, stores can be clustered together based on the trip mission distribution mix. Trip-missions tell us the customer motivations behind a shopping trip. Stores which have a similar trip mission distribution are catering to similar customer needs or motivations. For example, if majority of the trips in a store are ‘Discount-oriented’, the customers are coming to the store for bargains. On the other hand, if a majority of the trips are pantry stock-up trips, the store proposition is likely to be value-for-money and wider assortment.

Store clustering is not unlike other kinds of segmentation and relies on statistical and non-statistical methods to group together observations (stores) that are similar across certain pre-selected variables. One salient feature of this segmentation is the size of data. While a retailer can have millions of customers, the number of stores is usually more limited (100 to 10000). This means that resource-intensive techniques like neural networks can also be used to perform this exercise. K-means clustering and self organizing maps are popularly used techniques for store clustering.