Optimizing Fresh Product Planograms at Scale for a Leading Grocery Retailer

By integrating expiration-aware planning, Days of Supply targets, and advanced merchandising rules, DecisionBrain helped transform the way fresh products are allocated across store shelves.

A leading supermarket retailer in the United States operates hundreds of stores across multiple states in the Southeast region. To support store operations, the company relied on planogram optimization to determine how products should be allocated across shelves, tables, and display areas.

While the company already had an optimization solution for dry grocery categories such as cleaning products and bottled water, extending these capabilities to fresh products presented a significantly more complex challenge.

Unlike dry goods, fresh products introduce constraints that directly affect inventory management and store operations. Consider a bakery shelf stocked with 40 croissants where daily customer demand is only 15 units. Without expiration-aware modeling, the excess product spoils before it can be sold, generating direct waste and eroding margins. Multiply that across hundreds of SKUs and dozens of stores, and the operational impact becomes substantial. Beyond spoilage, food safety regulations require physical separation between raw and cooked products, further constraining how available shelf space can be used.

The company wanted to extend its existing optimization platform to support fresh categories while introducing new merchandising capabilities, such as shared facings, and improving the satisfaction of target inventory requirements across stores.

They needed a solution capable of:

• Optimizing shelf allocation for fresh and perishable products.
• Meeting target inventory levels while minimizing product waste.
• Enforcing food safety requirements, including separation between raw and cooked products.
• Supporting advanced merchandising strategies such as shared product displays.
• Providing store planners and shelf stockers with a practical and scalable planning tool.

Building a New Generation of Planogram Optimization

DecisionBrain partnered with IBM to extend the company’s existing optimization engine and create a more flexible framework capable of handling the operational realities of fresh products. The project was delivered through three successive phases.

Phase 1: Refactoring the Optimization Foundation

The first phase focused on restructuring and refactoring the existing optimization code. This work established a scalable architecture capable of supporting additional business rules and more sophisticated optimization constraints without compromising performance.

The objective was not only to improve maintainability, but also to prepare the model for the introduction of fresh-product requirements that were not considered in the original implementation.

Phase 2: Extending Optimization to Fresh Products

The second phase introduced constraints specific to fresh and perishable categories such as bakery and meat products.

A key enhancement was the integration of expiration-aware inventory planning. The optimization model was extended to consider the relationship between demand, shelf capacity, and product shelf life, limiting the number of units that could be allocated to a display in order to reduce the risk of spoilage and waste. The model was also enhanced to support Days of Supply (DOS) targets, balancing inventory objectives with shelf capacity and assortment priorities while maintaining feasible planograms.

Additional food safety constraints were also incorporated. The model automatically reserved separator spaces between raw and cooked products, ensuring compliance with hygiene requirements while maintaining feasible shelf layouts.

Phase 3: Introducing Shared Facings

The final phase focused on improving merchandising flexibility through the implementation of shared facings.

Previously, a single facing could only accommodate one product. The enhanced optimization model introduced rules allowing multiple products to share the same facing under specific conditions. For example, a bakery display table could present croissants in the front row and cakes behind them, enabling more efficient use of space while preserving visual merchandising standards.

Implementing this capability required the introduction of new optimization rules governing minimum quantities, partially filled facings, and display consistency, including a constraint ensuring that when an item occupies more than one facing, only one of those facings may be left partially filled.

Business Impact

Several years after its initial deployment, the optimization platform remained a critical component of the retailer’s pricing operations, supporting large-scale decision-making across multiple markets.

To ensure continued performance, scalability, and maintainability, a comprehensive optimization health check was conducted, reviewing:

The enhanced model improved the satisfaction of target inventory requirements while preserving assortment priorities and merchandising objectives. By integrating expiration constraints, inventory objectives, merchandising rules, and food safety requirements into a single optimization framework, the company gained a more accurate and operationally realistic approach to shelf planning.

Key outcomes include:

• Improved inventory management through expiration-aware shelf allocation.
• Reduced waste through better alignment of shelf quantities with expected demand and product shelf life.
• Enhanced merchandising flexibility through support for shared facings and modern product displays.
• Automated compliance with food safety requirements through built-in separation rules.
• Increased planning efficiency through fewer manual adjustments and reduced planogram infeasibilities.
• Unified planning capabilities supporting both strategic planning and day-to-day store execution.

Today, the solution is used by planners and store shelf stockers across the company’s network, helping ensure that planograms are not only optimized mathematically but also practical to execute in day-to-day store operations.