Multi-Stage Clustering for Retail Substitute Item Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Retail stores face challenges in managing shelf space due to limited availability, often removing low-sales items which can impact sales of interrelated items like traditional and variety substitutes, potentially losing customer base and reducing basket-building behavior.
Innovation Solution
A system performs multi-stage clustering analysis to identify non-substitute, traditional substitute, and variety substitute item-pairs, assigning propensity scores to determine the degree of interrelationship, thereby informing inventory management decisions to prevent removal of items with strong substitutes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If low-sales items are removed from inventory to increase shelf space, then shelf space availability is improved, but sales of interrelated substitute items may be negatively impacted
Solution Approach 1:
The system segments items into different substitute relationship categories (traditional substitutes, variety substitutes, complementary items) using multi-stage clustering analysis. This segmentation allows selective retention of items based on their specific relationship types, enabling precise inventory decisions that protect substitute item sales while optimizing shelf space.
Solution Approach 2:
The system performs preliminary clustering analysis and propensity score calculation before making inventory decisions. By pre-identifying substitute relationships and calculating interrelationship scores, the system can proactively determine which items to retain or remove, preventing negative impacts on substitute sales before they occur.
2Measurement precision
If multi-stage clustering analysis is performed to identify substitute relationships, then accuracy in identifying items to retain is improved, but computational complexity increases
Solution Approach 1:
The clustering process is divided into three distinct stages: initial clustering to identify potential substitutes, filtration to remove false positives, and final classification to categorize substitute types. This segmentation reduces computational complexity at each stage while maintaining high overall accuracy in identifying substitute relationships.
Solution Approach 2:
The system calculates propensity scores for all item-pairs in the initial clustering stage, then applies filtration to focus computational resources only on promising candidates. This partial action approach avoids unnecessary calculations on clearly non-substitute pairs while maintaining comprehensive analysis for potential substitutes.
Data Source
AI summary
Examples provide a multi-stage cluster component that performs a multi-stage clustering analysis on a plurality of items in a category associated with a selected item using a set of interrelationship factors. The multi-stage cluster component generates a cluster of non-substitute item-pairs, a cluster of traditional substitute item-pairs, and a cluster of variety item-pairs. The set of interrelationship factors includes at least one of measure of association, brand similarity, pack-size similarity, demographic similarity, item description similarity, lift, and/or percentage same-basket variable. A propensity score is generated for each item-pair. The propensity score is utilized to identify traditional substitute items and variety substitute items. Each substitute item is ranked based on the generated propensity score. The ranking is used to identify potential low-performance items for removal from inventory.