New Approach to Shelf Space and Product Assortment Optimization in Retail
The increasing product variety in the retail sector makes decisions such as how to most effectively allocate limited shelf space and which products should be included in the assortment more complex than ever. This dynamic environment creates a clear need for advanced, data-driven solutions to both meet customer demand and enhance operational efficiency.
Within the scope of the TEYDEB 1501 project “Sales Forecasting and Product Portfolio Optimization System Based on Store Characteristics and Substitution Effects,” the Obase R&D team has developed a decomposition-based solution tailored to these needs.
An Analytical Approach to Retail’s Most Critical Questions
By addressing shelf space allocation and product assortment decisions within a unified structure, the study provides scientific methodology for answering the key questions retailers frequently encounter:
• How much shelf space should be allocated to each category?
• Which SKUs should be included in each category?
• How many facings should each product have on the shelf?
• How do substitution behaviors influence demand?
These questions are central to both customer experience and store operations. The model developed in the study optimizes these decisions in line with the retailer’s objectives using historical data.
Core Methodology: Decomposition-Based Optimization
The system consists of three main components:
1. Category-Based Product Grouping
SKUs with demand cross-elasticity (substitution effects) are grouped to form substitution categories.
2. Portfolio Optimization Model (POM)
For each category, SKU selection and facing counts are determined under different shelf capacity scenarios. The model incorporates substitution rates, product visibility, replenishment costs, and space elasticity.
3. Shelf Capacity Allocation Model (SCAM)
The store’s total shelf space is distributed across categories to optimize the objective function—profit, revenue, or sales volume.
This decomposition approach enables shelf space planning and assortment selection problems to be solved efficiently within realistic timeframes.
Scalability Validated Through Realistic Test Scenarios
The models were tested under varying SKU counts, shelf capacity options, and category distributions, demonstrating strong scalability.
The results show that when product variety and shelf space management are handled effectively, retailers can achieve notable improvements in:
• profitability,
• replenishment costs,
• category performance.
From Scientific Model to Business Application
The proposed methodology was not only developed as a theoretical framework but also transformed into a commercial Product Assortment Optimization Software by the Obase R&D team. The software was designed to support the decision-making processes of category managers and to align with real-world retail conditions.
Our R&D article prepared within the scope of the “Sales Forecasting and Product Portfolio Optimization System Based on Store Characteristics and Substitution Effects” project has been published in the Journal of the Faculty of Engineering and Architecture of Gazi University (SCI–Expanded).
