Planogram Optimization for Food Categories
Executive Summary
We upgraded a genetic algorithm-based optimizer that recommends category arrangements by size and adjacency. By strengthening selection, mutation, and randomization operators, run time dropped from minutes to near real-time and uplift improved materially.
A provider of image recognition tech and market data services for tier-one manufacturers needed to dramatically reduce optimization run time from 6-10 minutes to near real-time.
Business Challenge
Slow Optimization Speed
Optimization took 6-10 minutes per scenario, preventing real-time testing during merchandising sessions.
Limited Sales Impact
Sales uplift plateaued at ~3%, below industry benchmarks and client expectations.
Poor Store-Layout Fit
Optimizer produced academically optimal layouts that didn't match real-world store characteristics.
Batch Processing Bottleneck
Category managers couldn't explore what-if scenarios interactively, limiting decision quality.
What We Built
Data and Signals
Product Data
- • SKU dimensions and packaging
- • Category adjacency rules
- • Price points and margins
- • Brand hierarchies
Store Context
- • Shelf dimensions and fixtures
- • Store format and size
- • Traffic flow patterns
- • Demographics and location
Sales History
- • Product velocity by store
- • Seasonal patterns
- • Promotion lift factors
- • Cross-category affinities
Constraints
- • Manufacturer agreements
- • Safety requirements
- • Inventory limits
- • Visual merchandising rules
Modeling Approach
Enhanced Genetic Algorithm
Redesigned selection, crossover, and mutation operators with adaptive rates that accelerate convergence while maintaining solution diversity.
Smart Fitness Function
Multi-objective optimization balancing sales potential, operational efficiency, and visual appeal using store-specific weights and constraints.
Parallel Evolution
Population-based parallelization with island models that explore different regions of solution space simultaneously, converging on optimal layouts faster.
Planning and Simulation Tool
Interactive web application where category managers can define constraints, run optimizations in real-time, compare scenarios side-by-side, and export planograms directly to store execution systems.
Real-time API
RESTful endpoints supporting sub-minute optimization requests with progress streaming
Result Caching
Redis-based caching for instant retrieval of similar scenarios and incremental refinements
Export Integration
Direct export to JDA Space Planning, Apollo, and other planogram software formats
Results and Impact
Operational Outcomes
- Category managers can test scenarios during planning sessions
- Better layouts leading to measurable sales increases
- Competitive advantage for the service provider
- Foundation for AI-driven merchandising insights
Financial View
- Doubled sales uplift from 3% to 7-8%
- 15x increase in daily optimization throughput
- New revenue streams from real-time optimization services
- Reduced time to market for new product introductions