A renowned US-based retailer with headquarters in the Pacific Northwest.
The client wanted to establish a chain of “Lite” stores closer to the city, as customers based in the cities found it inconvenient to travel out of the city to purchase food items. However, replicating the product assortment from the incumbent large stores would have been ineffective. Hence, the retailer wanted to optimize the assortment for its food category in the new “Lite” stores.
Leveraging our deep expertise in the consumer sector, we employed the following approach to resolving this situation:
We then determined the factors for clustering similar stores, and deployed K-means clustering on these factors to identify the stores that could be clustered. Using sales aggregation and ranking, the SG Analytics' team then identified the high-performing stores in each cluster, and finally determined the top-selling SKUs in these stores.
We tracked various historical transactions to understand consumer preferences, and based on weighted sales attribution, transferred demand of a particular SKU to other SKUs based on similarity. Using association rule mining, SG Analytics' team also determined the affinity sales SKUs (SKUs mostly bought together) using the Confidence-Support methodology by calculating affinity scores for the SKUs.
By aggregating transactional data across SKUs, our team estimated the demand for each SKU. Based on constraints such as minimum and maximum quantity, shelf space, etc., we implemented a SAS based optimization algorithm using PROC OPTMODEL. After applying constraints related to needs of the core customers, we determined the list of SKUs with estimated demand for sales value maximization for the target store.
Implemented a holistic solution which was accepted by all the category managers across the small stores.
Implemented the solution for some of the non-food categories as well.