Machine Learning Demand Transfer Estimation for Retail Assortment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current demand transfer estimation methods in retail fail to consider simultaneous product interactions and major causative factors, leading to inaccurate demand shifts due to limited product space and numerous sales drivers in diverse formats.
Innovation Solution
A processor-implemented method and system that collects and aggregates sales drivers from various sources, generating a multivariate multi-structure data matrix to analyze product behavior using a machine learning model, enabling simultaneous consideration of all products and sales drivers for accurate demand transfer estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional demand transfer estimation methods are used, then the estimation process is simple, but the accuracy of demand transfer estimation is poor because simultaneous product interactions and major causative factors are not considered
Solution Approach 1:
The patent segments the demand transfer estimation problem into multiple independent components: product interaction effects, promotion effects, pricing effects, and other sales driver effects. Each component is modeled separately using machine learning algorithms, allowing the system to capture complex relationships while maintaining computational tractability and model interpretability
Solution Approach 2:
The patent introduces machine learning models as intermediary components that process raw sales data, product attributes, and sales driver information. These models act as mediators between the input data and the final demand transfer estimation, enabling the system to handle complex non-linear relationships and interactions without requiring explicit mathematical formulations
2Ease of manufacture
If product attributes are bucketed to reduce distinct values, then the modeling requirement is satisfied, but bias is introduced based on user decision making
Solution Approach 1:
The patent transforms categorical product attributes into numerical representations suitable for machine learning processing. Instead of bucketing categorical variables, the system uses one-hot encoding or embedding techniques to convert high-cardinality categorical attributes into a format that preserves information while being compatible with statistical models, thereby avoiding the bias introduced by arbitrary bucketing decisions
3Measurement precision
If all sales drivers are processed at granular level, then the explanation of sales variation is maximized, but the data processing complexity increases
Solution Approach 1:
The patent segments sales drivers into distinct categories (product interactions, promotions, pricing, inventory, and external factors) and processes each category separately. This segmentation allows the system to handle granular information for each driver type while managing overall data processing complexity through modular architecture and specialized algorithms for each driver category
Solution Approach 2:
The patent creates structured representations (copies) of sales driver data in standardized formats that are optimized for machine learning processing. By transforming diverse granular sales driver information into consistent data structures with defined schemas and relationships, the system preserves detailed information while simplifying subsequent analysis and model training processes
Data Source
AI summary
This disclosure relates to a system and method to estimate demand transfer of a product while considering performance of all the products of a category simultaneously. It would be appreciated that the demand of a removed product transfers to other products of same category in a store. In addition the demand transfer is influenced by sales drivers such as product level promotion and competitor prices, store location, weather and seasonality. By considering these factors the proposed approach provides a method to estimate demand transfer of a product. It is addressed by creating multivariate multi structure machine learning models and estimating demand transfer values by using suitable scenario generator for product availability. It enables to estimate more holistic demand transfer values by simultaneous consideration of individual product behaviours with respect to other products availability and other sales drivers.