Automated promotion response modeling in a customer relationship management system
First Claim
1. A computer-implemented method of creating a customer promotion response model for use in customer relationship marketing, comprising:
- (a) generating, in a computer, an input data set for the response model, wherein the input data set is generated using an Analytic Data Set Template containing one or more Analytic Variables that include both primitives that are base variables and conditions that are predicates, aggregates or other functions, wherein the primitives and conditions determine how the Analytic Variables are derived from operational data to produce the input data set, and wherein the Analytic Variables are subdivided into independent variables and their related dependent variables;
(b) splitting, in the computer, the input data set into a test sample and a validation sample;
(c) identifying, in the computer, the independent variables and their related dependent variables using the test sample;
(d) identifying, in the computer, a Transformation Type for each of the identified independent variables and their related dependent variables;
(e) estimating, in the computer, a Coefficient for each of the identified independent variables and their related dependent variables;
(f) generating, in the computer, a Model Equation for each of the identified independent variables and their related dependent variables using the identified Transformation Type and estimated Coefficient;
(g) validating, in the computer, the generated Model Equation by applying it to the validation sample; and
(h) scoring, in the computer, customers retrieved from a database using the validated Model Equation as a customer promotion response model for use in customer relationship marketing.
2 Assignments
0 Petitions
Accused Products
Abstract
A Customer Relationship Management (CRM) system includes a CRM Client, CRM Engine, and Relational Database Management System (RDBMS). The CRM Client, CRM Engine, and RDBMS integrate a suite of services that allow users to plan, manage, and execute promotional or marketing campaigns, build customer segments, score customers, and analyze customer behavior, product purchases, and response to promotional campaigns. The services include an automated Analytic Data Set Creation service, which simplifies and automates the process of creating analytic data sets useful for modeling and analysis out of operational data stored in the relational database, and a Response Modeling service, which automatically creates promotion response models to score individual customers based on that model in order to predict which customers are most likely to respond to a future promotional campaign.
14 Citations
21 Claims
-
1. A computer-implemented method of creating a customer promotion response model for use in customer relationship marketing, comprising:
-
(a) generating, in a computer, an input data set for the response model, wherein the input data set is generated using an Analytic Data Set Template containing one or more Analytic Variables that include both primitives that are base variables and conditions that are predicates, aggregates or other functions, wherein the primitives and conditions determine how the Analytic Variables are derived from operational data to produce the input data set, and wherein the Analytic Variables are subdivided into independent variables and their related dependent variables; (b) splitting, in the computer, the input data set into a test sample and a validation sample; (c) identifying, in the computer, the independent variables and their related dependent variables using the test sample; (d) identifying, in the computer, a Transformation Type for each of the identified independent variables and their related dependent variables; (e) estimating, in the computer, a Coefficient for each of the identified independent variables and their related dependent variables; (f) generating, in the computer, a Model Equation for each of the identified independent variables and their related dependent variables using the identified Transformation Type and estimated Coefficient; (g) validating, in the computer, the generated Model Equation by applying it to the validation sample; and (h) scoring, in the computer, customers retrieved from a database using the validated Model Equation as a customer promotion response model for use in customer relationship marketing. - View Dependent Claims (2, 3, 4, 5, 6, 7)
-
-
8. A computer-implemented system for creating a customer promotion response model for use in customer relationship marketing, comprising:
-
(a) a computer; (b) a customer relationship marketing system, performed by the computer, for; (1) generating an input data set for the response model, wherein the input data set is generated using an Analytic Data Set Template containing one or more Analytic Variables that include both primitives that are base variables and conditions that are predicates, aggregates or other functions, wherein the primitives and conditions determine how the Analytic Variables are derived from operational data to produce the input data set, and wherein the Analytic Variables are subdivided into independent variables and their related dependent variables; (2) splitting the input data set into a test sample and a validation sample; (3) identifying the independent variables and their related dependent variables using the test sample; (4) identifying a Transformation Type for each of the identified independent variables and their related dependent variables; (5) estimating a Coefficient for each of the identified independent variables and their related dependent variables; (6) generating a Model Equation for each of the identified independent variables and their related dependent variables using the identified Transformation Type and estimated Coefficient; (7) validating the generated Model Equation by applying it to the validation sample; and (8) scoring customers retrieved from a database using the validated Model Equation as a customer promotion response model for use in customer relationship marketing. - View Dependent Claims (9, 10, 11, 12, 13, 14)
-
-
15. An article of manufacture comprising a storage device embodying instructions that, when read and executed by a computer, result in the computer performing a method for creating a customer promotion response model for use in customer relationship marketing, comprising:
-
(a) generating, in a computer, an input data set for the response model, wherein the input data set is generated using an Analytic Data Set Template containing one or more Analytic Variables that include both primitives that are base variables and conditions that are predicates, aggregates or other functions, wherein the primitives and conditions determine how the Analytic Variables are derived from operational data to produce the input data set, and wherein the Analytic Variables are subdivided into independent variables and their related dependent variables; (b) splitting, in the computer, the input data set into a test sample and a validation sample; (c) identifying, in the computer, the independent variables and their related dependent variables using the test sample; (d) identifying, in the computer, a Transformation Type for each of the identified independent variables and their related dependent variables; (e) estimating, in the computer, a Coefficient for each of the identified independent variables and their related dependent variables; (f) generating, in the computer, a Model Equation for each of the identified independent variables and their related dependent variables using the identified Transformation Type and estimated Coefficient; (g) validating, in the computer, the generated Model Equation by applying it to the validation sample; and (h) scoring, in the computer, customers retrieved from a database using the validated Model Equation as a customer promotion response model for use in customer relationship marketing. - View Dependent Claims (16, 17, 18, 19, 20, 21)
-
Specification