Customer State Segmentation for LTV Growth Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for identifying incremental growth of a customer's lifetime value (LTV) result in non-continuous functions, making it difficult to estimate derivatives and predict effective transitions for individual customers.
Innovation Solution
A system and method that segment customers into distinct states based on their purchase history and behavior, using machine learning models to predict LTV and coordinate targeted online advertisements to transition customers from lower to higher LTV states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional systems are used to identify incremental growth of LTV for individual customers, then overall LTV growth paths can be identified, but the functions become non-continuous making derivative estimation very difficult
Solution Approach 1:
The patent segments customers into distinct states (e.g., new customer, active customer, lapsed customer, lost customer) based on their behavior and purchase history. This segmentation transforms the continuous but complex LTV growth problem into discrete state transitions that are easier to model and analyze, resolving the contradiction between measurement precision and system complexity.
2Ease of operation
If customers are segmented into distinct states with machine learning models, then interpretable transitions and targeted marketing actions become possible, but the system complexity increases
Solution Approach 1:
By dividing the customer base into distinct, interpretable states and defining clear transition paths between them, the system makes marketing actions more understandable and actionable. The segmentation creates a simplified view of complex customer behavior that is easier to operate with, despite the underlying machine learning models.
Solution Approach 2:
The patent introduces customer states as intermediary concepts that mediate between raw customer data and marketing actions. These states serve as a simplified representation layer that makes the relationship between customer behavior and marketing interventions more interpretable, reducing the operational complexity.
3Loss of information
If non-continuous functions are used for LTV identification, then overall growth paths can be captured, but derivative estimation becomes very difficult
Solution Approach 1:
The patent segments the continuous LTV growth trajectory into discrete customer states and transition probabilities. This segmentation allows the system to capture overall growth information while avoiding the mathematical difficulties of estimating derivatives from non-continuous functions, as the discrete transitions provide clear, interpretable growth paths.
Data Source
AI summary
Systems and methods including one or more processing modules and one or more non-transitory storage modules storing computing instructions configured to run on the one or more processing modules and perform acts of determining a lifetime value (LTV) for customers of a retailer, segmenting the customers into customer states based upon one or more purchases made by each customer at the retailer within a predetermined period of time, determining a first average LTV for customers in a first customer state and a second average LTV for customers in a second customer state lower than the first average LTV, coordinating a first display of a first online advertisement for customers in the first LTV to transition the customers from the first customer state to the second customer state, and coordinating a second display of a second online advertisement for customers in the second state.