Personalized Product Display via Purchase Probability Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online shopping platforms fail to provide personalized product recommendations, as they primarily showcase items popular among the general customer base, neglecting individual purchasing habits and preferences.
Innovation Solution
A system comprising a control unit that communicates with product, category, and customer purchase history databases to categorize products and calculate the probability of a customer being an underbuyer or overbuyer of a product category, using a Beta Negative Binomial distribution and customer preferences to tailor product displays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If products are displayed based on general population purchasing habits, then the catalogue reflects popular items, but individual customer preferences are not personalized
Solution Approach 1:
The system pre-calculates and stores purchase probability metrics for each customer-category combination before the shopping session. This preliminary analysis of historical purchase data enables rapid personalized product recommendation during the actual shopping experience, resolving the contradiction between personalization and real-time processing complexity
Solution Approach 2:
The patent introduces an intermediary layer (the server's analysis system) that processes historical purchase data and generates personalized product recommendations. This intermediary translates raw purchase history into actionable insights, enabling personalization without requiring complex real-time processing at the customer interface
2Loss of information
If the catalogue shows most-purchased products prominently, then general customer interests are addressed, but customers with different preferences see irrelevant products
Solution Approach 1:
The system performs preliminary filtering and ranking of products based on each customer's predicted purchase probability before the customer even views the catalogue. By pre-processing product relevance based on historical data, the system eliminates the need for time-consuming real-time filtering during the shopping experience
Solution Approach 2:
The system automatically analyzes customer purchase history and generates personalized product rankings without requiring customer input or manual curation. This self-service approach to personalization reduces time loss by eliminating manual intervention while maintaining high relevance of product information
3Ease of operation
If a virtual shopping basket model is used, then the ordering process is simplified, but the system cannot proactively suggest relevant products
Solution Approach 1:
The system continuously monitors customer purchase behavior and uses this feedback to dynamically update purchase probability metrics. This feedback loop enables the system to learn from customer actions and improve proactive recommendations while maintaining the simplicity of the virtual shopping basket interface
Solution Approach 2:
The system proactively identifies and presents relevant products to customers before they actively search for them, based on preliminary analysis of purchase patterns. This allows the system to enhance sales efficiency without complicating the ordering process, as recommendations are integrated into the existing virtual basket workflow
Data Source
AI summary
There is provided an apparatus and method for a webshop such that the products shown to a customer are related to the purchasing habits of the customer. A control unit is arranged to communicate with a product information database, a product category database and a customer purchase history database. The control unit includes a product categorising unit arranged to generate at least one product category based on product information in the product information database and to store the at least one generated product category in the product category database. A calculating unit is arranged to calculate a probability of a customer being an underbuyer/overbuyer of a type of product based on the customer's purchase history stored in the customer purchase history database and the at least one product category from the product category database.