Identifying target customers for campaigns to increase average revenue per user

• US 8,762,193 B2
• Filed: 12/01/2005
• Issued: 06/24/2014
• Est. Priority Date: 11/11/2005
• Status: Active Grant

• Current Assignee: Accenture Global Services Limited (Accenture PLC)
• Original Assignee: Accenture Global Services Limited (Accenture PLC)
• Inventors: Maga, Matteo, Canale, Paolo, Bohe, Astrid
• Primary Examiner(s): Boswell, Beth V
• Assistant Examiner(s): FEACHER, LORENA R
• Application Number:
  US11/291,438
• Publication Number:
  US 20070112614A1
• Time in Patent Office: 3,127 Days
• Field of Search: 705/7.29, 705/7.31, 705/7.33
• US Class Current: 705/7.31
• CPC Class Codes: G06Q 10/06375 Prediction of business process outcome or impact based on a proposed change
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0204 Market segmentation

Abstract

A method and system that provide analytical tools for increasing average revenue per user (ARPU) allows users to design and execute marketing campaigns that target customers with a statistically significant likelihood of accepting a marketing campaign offer and generating the greatest increase in revenue. The method and system creates statistical models to determine an individual customer'"'"'s propensity to respond positively to a campaign and propensity to generate increased revenue. The method and system scores the customers according to the customers'"'"' propensities, and uses the scoring results to select an optimal mix of customers to contact during the marking campaign.

Claims

1. A system comprising:

a processor;

a user interface, controlled by the processor, configured to;

display data on a display based on first and second scoring criteria for scoring a customer; and

receive the first and the second scoring criteria for scoring the customer, where one of the first and second scoring criteria comprises criterion selected from a group comprising of;

billed usage increase propensity criteria that identify characteristics of customers subject to a threshold increase in billed usage over a period of time, and

customer loyalty criteria that identify a customer'"'"'s propensity to stop using a product;

a data mart, controlled by the processor, with customer data about individual customers in a customer population;

an application program comprising processing instructions that when executed by the processor load the data mart with the customer data;

a data mining tool, controlled by the processor, configured to;

analyze customer data;

prepare first and second statistical models for scoring customers based on customer attributes quantified in said customer data;

generate, using said first and second statistical models, first and second scores for the customers in the customer population by quantifying the customer attributes in said customer data using the first and second scoring criteria; and

identify a subset of customers from the customer population based on the first score satisfying the first scoring criterion and the second score satisfying the second scoring criterion, by

calculating for each of the customers;

a revenue trend line;

revenue slope classes based on the revenue trend line; and

a size of customer population within the revenue slope class;

classifying each of the customers into the revenue slope classes based on the revenue trend line;

defining a first plurality of distribution classes for grouping customers according to the customer scores generated by the first statistical model and define a second plurality of distribution classes for grouping customers according to the customer scores generated by the second statistical model,

wherein the first statistical model is configured to calculate a customer'"'"'s propensity to generate increased revenue and score the customer based on said propensity to generate increased revenue, and

wherein the second statistical model is configured to calculate a customer'"'"'s propensity to respond to a marketing campaign and score the customer based on said propensity to respond;

assigning customers to distribution classes within said first plurality of distribution classes based on their first score and assign customers to distribution classes within said second plurality of distribution classes based on their second score;

filtering the customers assigned to the distribution classes of said first plurality of distribution classes by the distribution classes to which the customers are assigned in said second plurality of distribution classes;

sizing, on a bubble chart, a bubble for each of the revenue slope classes to correspond to the size of the customer population within the revenue slope class as the percentage of overall customer population; and

identifying the subset of customers from the customer population by the size of the bubble for each revenue slope class on the bubble chart;

a data manipulation module comprising processing instructions that when executed by the processor;

prepare the customer data stored in the data mart for data mining;

transport said prepared data to the data mining tool; and

store the first and second scores for the customers in the data mart; and

a reporting tool, comprising processing instructions that when executed by the processor;

access the customer data stored in the data mart, including the first and second scores for the customers; and

report, by the user interface displaying, customer distributions based on said first and second customer scores,

where the user interface is further configured to;

receive commands from the user to manipulate; and

contrast the customer data based on the received first and second scoring criteria by displaying the customer distributions in the bubble chart for a customer segment distributed between the revenue slope classes and an average revenue of each of the revenue slope classes.

2. The system of claim 1, wherein the reporting tool, further comprising processing instructions that when executed by the processor:

define a first plurality of distribution classes for grouping customers according to the scores generated by the first statistical model; and

define a second plurality of distribution classes for grouping customers within one of the distribution classes of said first plurality of distribution classes according to the scores generated by the second statistical model, and

wherein said prepared customer data comprises customer analytical records that identify customer attributes of customers who have already responded positively to the first and second scoring criteria for scoring a customer.

3. The system of claim 2, wherein the first statistical model is configured to calculate a customer'"'"'s propensity to respond to a marketing campaign and score the customer based on said propensity to respond.

4. The system of claim 3, wherein the second statistical model is configured to calculate a customer'"'"'s propensity to generate increased revenue and score the customer based on said propensity to generate increased revenue.

5. The system of claim 2, wherein the first statistical model is configured to calculate a customer'"'"'s propensity to generate increased revenue and score the customer based on said propensity to generate increased revenue.

6. The system of claim 5, wherein the second statistical model is configured to calculate a customer'"'"'s propensity to respond to a marketing campaign and score the customer based on said propensity to respond.

7. The system of claim 1, wherein the first statistical model is configured to calculate a customer'"'"'s propensity to respond to a marketing campaign and score the customer based on said propensity to respond, and wherein the second statistical model is configured to calculate a customer'"'"'s propensity to generate increased revenue and score the customer based on said propensity to generate increased revenue.

8. A method comprising:

loading, by a processor executing an application program, customer data into a data mart;

receiving, through a user interface controlled by the processor, first scoring and second scoring criteria, where one of the first and second scoring criteria comprises criterion selected from a group comprising of;

billed usage increase propensity criteria that identify characteristics of customers subject to a threshold increase in billed usage over a period of time, and

customer loyalty criteria that identify a customer'"'"'s propensity to stop using a product;

developing, by the processor executing the application program, a first statistical model with the first scoring criterion for scoring customers according to the first scoring criterion;

developing, by the processor executing the application program, a second statistical model with the second scoring criterion for scoring customers according to the second scoring criterion;

applying, by the processor executing the application program, the first and second statistical models to the customer data to generate first and second scores for individual customers in the customer population;

calculating for each of the customers a revenue trend line;

grouping, by the processor executing the application program, customers according to said first and second scores of the customers, and

calculating revenue slope classes based on the revenue trend line;

calculating a size of customer population within each of the revenue slope classes;

classifying each of the customers into revenue slope classes based on the revenue trend line;

defining a first plurality of distribution classes for grouping customers according to the customer scores generated by the first statistical model and define a second plurality of distribution classes for grouping customers according to the customer scores generated by the second statistical model,

wherein the first statistical model is configured to calculate a customer'"'"'s propensity to generate increased revenue and score the customer based on said propensity to generate increased revenue, and

wherein the second statistical model is configured to calculate a customer'"'"'s propensity to respond to a marketing campaign and score the customer based on said propensity to respond;

assigning customers to distribution classes within said first plurality of distribution classes based on their first score and assign customers to distribution classes within said second plurality of distribution classes based on their second score;

filtering the customers assigned to the distribution classes of said first plurality of distribution classes by the distribution classes to which the customers are assigned in said second plurality of distribution classes;

sizing, on a bubble chart, a bubble for each of the revenue slope classes to correspond to the size of the customer population within the revenue slope class as the percentage of overall customer population; and

analyzing, by the processor executing the application program, the customer groupings to identify a subset of customers from the customer population by the size of the bubble for each revenue slope class on the bubble chart, where the first score and the second score for each of the subset of customers satisfy the first scoring criterion and the second scoring criterion, respectively, indicating a statistically significant likelihood of accepting a campaign offer and generating increased revenue; and

displaying, through the user interface controlled by the processor, the subset of customers from the customer population, including displaying customer distributions, based on the first score and the second score, in the bubble chart for a customer segment distributed between the revenue slope classes and an average revenue of each of the revenue slope classes.

9. The method of claim 8, wherein the customer groupings comprise:

a first grouping of the subset of customers identify first scores that satisfy the first scoring criterion; and

interface

a second grouping of the subset of customers identify second scores that satisfy the second scoring criterion; and

wherein analyzing customer groupings comprises analyzing distribution of the second scores for the subset of customers from the second grouping according to the first score for each of the subset of customers from the second grouping.

10. The method of claim 9 wherein the one of said first and second scoring criteria is from the group of the criterion where the group further comprises of:

a customer'"'"'s propensity to accept a marketing campaign offer, and the other of said first and second scoring criteria is a customer'"'"'s propensity to generate increased revenue.

11. The method of claim 8 wherein the first statistical model uses the first scoring criterion to identify a customer'"'"'s propensity to accept a campaign offer and the second statistical model uses the second scoring criterion to identify a customer'"'"'s propensity to generated additional revenue after accepting the campaign offer.

12. The method of claim 11, wherein the identified subset of customers identifies those customers from the customer population with a high propensity to accept the campaign offer and highest propensity to generate additional revenue as a result.

13. The method of claim 12 further comprising training, performed by the processor executing the application program, the first and second models on historical customer data relating to past implementation of a substantially similar marketing campaign.

14. A method of creating a marketing campaign list of customers to target, the method comprising:

loading, by a processor executing an application program, customer data into a data mart using an application program;

receiving, through a user interface controlled by the processor, first scoring criterion for scoring a customer'"'"'s propensity to generate increased revenue and second scoring criterion for scoring a customer'"'"'s propensity to respond positively to a marketing campaign, where the first scoring criterion is selected from a group comprising of;

billed usage increase propensity criteria that identify characteristics of customers subject to a threshold increase in billed usage over a period of time, and

customer loyalty criteria that identify a customer'"'"'s propensity to stop using a product;

creating, by the processor executing the application program using the first scoring criterion, a first statistical model for determining a customer'"'"'s propensity to generate increased revenue;

creating, by the processor executing the application program using the second scoring criterion, a second statistical model for determining a customer'"'"'s propensity to respond to a marketing campaign;

scoring a plurality of customers, by the processor executing the application program using the customer data, according to each of the customer'"'"'s propensity to generate increased revenue and propensity to respond positively to a marketing campaign;

calculating for each of the customers a revenue trend line, and

calculating revenue slope classes based on the revenue trend line;

classifying each of the customers into revenue slope classes based on the revenue trend line;

defining a first plurality of distribution classes for grouping customers according to the customer scores generated by the first statistical model and define a second plurality of distribution classes for grouping customers according to the customer scores generated by the second statistical model,

wherein the first statistical model is configured to calculate a customer'"'"'s propensity to generate increased revenue and score the customer based on said propensity to generate increased revenue, and

wherein the second statistical model is configured to calculate a customer'"'"'s propensity to respond to a marketing campaign and score the customer based on said propensity to respond;

assigning customers to distribution classes within said first plurality of distribution classes based on their first score and assign customers to distribution classes within said second plurality of distribution classes based on their second score;

filtering the customers assigned to the distribution classes of said first plurality of distribution classes by the distribution classes to which the customers are assigned in said second plurality of distribution classes;

calculating a size of customer population within each of the revenue slope classes;

evaluating, by the processor executing the application program, the results of the scoring to identify a group of customers from the plurality of customers identified to have a high propensity for generating increased revenue and a high propensity to respond positively to a marketing campaign, using a bubble chart that includes a customer segment distributed between revenue slope classes and average revenue of each of the revenue slope classes, the bubble chart comprising a bubble for each revenue slope class, where each of the bubbles is sized to correspond to the size of customer population within the revenue slope class as a percentage of overall customer population; and

adding, by the processor executing the application program, the identified group of customers to the marketing campaign list of customers to target.

15. The method of claim 14, wherein evaluating the results of the scoring comprises:

grouping customers according to the customers'"'"' propensity to respond to a marketing campaign scores and propensity to generate increased revenue scores;

selecting a group of customers from the grouping of customers based on one of the customers'"'"' propensity to respond to a marketing campaign scores or propensity to generate increased revenue scores, wherein the selected group of customers comprises a distribution of the customers; and

analyzing the distribution of customers within the selected group according to the other of the customers'"'"' propensity to respond to a marketing campaign scores or their propensity to generate increased revenue scores.