US 10,769,651 B2 · Estimating prospect lifetime values

Abstract

The disclosure relates generally to estimating lifetime values, and more particularly to estimating Prospect Lifetime Values (PLTVs) for prospective customers for an organization. In one example, for estimating PLTV for a prospective customer, a distance of the prospective customer from each of existing customers of the organization is computed using existing customer data and prospective customer data. The existing customer data includes at least one of existing customer demographic variables and existing customer profile variables. Further, the prospective customer data includes at least one of prospective customer demographic variables and prospective customer profile variables. Subsequently, for the prospective customer, a pre-defined number of existing customers are determined from amongst the plurality of existing customers based on the distance. Thereafter, PLTV for the prospective customer is estimated using an average of Customer Lifetime Values (CLTVs) for the determined existing customers.

A system (100), implemented on a server, for accurately estimating and assigning Prospect Lifetime Values (PLTVs) for a plurality of prospective customers using a large existing customer base of an organization, the lifetime value estimation system (100) comprising:

a processor (102);
and
a memory 106 coupled to the processor 102, wherein the processor 102 executes a plurality of modules 108 stored in the memory 106, and wherein the plurality of modules 108 comprises:
an interaction module (112) to obtain, from one or more data sources, information related to the plurality of prospective customers and a plurality of existing customers of the organization, wherein the data sources include one or more of social media websites, external marketing datasets, and internal database of the organization;
a computation module (114) to,
identify existing customer data and prospective customer data that are common to both prospective customers and existing customers from the obtained information, wherein the existing customer data and the prospective customer data represent a string of values associated with different variables, and wherein the variables received in a non-numeric format are converted into a numeric value or a computer-processable format;
compute a distance of each of the plurality of prospective customers from each of the plurality of existing customers of the organization using the existing customer data and the prospective customer data, wherein a large computed distance represent that the prospective customer is not similar to the existing customer;
generate distance vectors for the plurality of prospective customers, wherein a distance vector for the prospective customer includes a set of distances of the prospective customer from the plurality of existing customers;
sort the generated distance vectors such that an existing customer who is closest to the prospective customer in terms of similarity is placed on top in the distance vector;
a determination module (116) to,
reduce processing time and processing load used for accurately estimating PLTVs from the large existing customer base of the organization, wherein the processing time and processing load are reduced by, determining for each of the plurality of prospective customers, one or more existing customers from amongst the plurality of existing customers based on a selection rule and the distance, such that the determined one or more existing customers are similar to respective prospective customer, wherein the selection rule includes one of:
comparing the distances between the plurality of prospective customers and the plurality of existing customers with a pre-determined threshold distance value;
and
selecting a pre-defined number of existing customers from amongst the plurality of existing customers, wherein a value of the pre-defined number of the existing customers is determined using k-Nearest Neighbors (KNN) algorithm by iterating on an optimal number of nearest existing customers in order to minimizes the Mean Absolute Percentage Error (MAPE) on the prospective customer data and the existing customer data;
and
an estimation module (118) to,
estimate and assign a PLTV for each of the plurality of prospective customers using an average value of Customer Lifetime Values (CLTVs) for respective one or more existing customers that were determined based on the selection rule, wherein the PLTV for a prospective customer is indicative of revenue to be generated by association of the prospective customer with the organization, and wherein the CLTV for an existing customer is indicative of revenue generated by association of the existing customer with the organization;
a database to store the estimated PLTV values, the existing customer data, the prospective customer data and data generated as a result of execution of the plurality of modules, and wherein the database is periodically updated.

15 more elements

Estimating prospect lifetime values

Life of this patent

Estimating prospect lifetime values

Life of this patent

Abstract

Claims9