Case study: Product range optimization - Efficient assortment

A discount store with more than 200 branches commissioned COSMO CONSULT to develop a solution that optimizes the product range of the individual branches on the basis of shopping behaviors and other influencing factors. Today, the company is benefiting from a higher profitability and reduced storage costs. In addition, it was possible to optimally adapt the product range to the condition, environment and competitive situation of each respective branch.

• Optimizing individual product range for individual branches dynamically
• Targeted evaluation and consideration of branches and branch groups
• Considering different underlying conditions and restrictions
• Placing products optimally for maximum sales
• Efficient action recommendations

The product range of a discount store with more than 200 branches throughout Germany grew so much over the years that it became economically inefficient to offer the full range of products in every branch. The company therefore had to implement a procedure that would reduce the product range in the branches while fulfilling as many customer requests as possible. Around 80 percent of the product range should be available in all branches, but the remaining 20 percent should be set individually depending on the environment and location of each branch.

The retailer commissioned COSMO CONSULT to optimize the product range of each individual branch in such a way that the market potential would be fully exploited while preserving the individuality of each location. Therefore, the local shopping behaviors of customers and other factors were analyzed.

The project team used stock list, sales and branch data to conduct the optimization. For example, aggregate analysis was used to determine the frequency at which certain combination of products are sold together from the sales data. The result was a network structure that clearly displays relevant information such as grouping, sales, revenue or quantities. The advantage: The structure illustrates the overall size as well as the influence of each product on the overall range of products. Depending on the respective strategy, different network points were used to rank the optimal product range. Then, an allocation algorithm was used to determine the best shelf placement of the product range in order to use composite effects and increase the sales quantity per receipt. In addition, the ranking can be used to initiate targeted actions for individual products that lead to high composite effects.

Today, the discount store uses the optimization process in several ways. First, an update script runs weekly in R and CPLEX, which performs the analysis and saves the results in a QlikView file. The results are then visualized using a QlikView dashboard to facilitate the decision-making process.

As part of the optimization process, it was not only possible to increase the sales quota, but also to make the action planning more targeted and efficient. From the history of actions, "flops" can be identified, which barely affects products outside of the action quantities. The optimal placement could increase profits and reduce storage costs. In addition, it was possible to optimally adapt the product range to the condition, environment and competitive situation of each respective branch.