PREDICTING CUSTOMER PURCHASE BEHAVIOR FOR EDUCATIONAL TECHNOLOGY PRODUCTS
The disclosed embodiments provide a system for processing data. During operation, the system obtains a set of features for a customer of an educational technology product. Next, the system uses the set of features to calculate an overall score representing a predicted purchase behavior of the customer with the educational technology product. The system then uses multiple subsets of the features to calculate a set of sub-scores that characterize different components of the overall score. Finally, the system outputs the overall score and the sub-scores for use in managing sales activity with the customer.
1 . A method, comprising:
obtaining a set of features for a customer of an educational technology product;
using the set of features to calculate, by one or more computer systems, an overall score representing a predicted purchase behavior of the customer with the educational technology product;
using multiple subsets of the features to calculate, by the one or more computer systems, a set of sub-scores that characterize different components of the overall score; and
outputting the overall score and the sub-scores for use in managing sales activity with the customer.
2 . The method of claim 1 , wherein using the set of features to calculate the overall score comprises:
applying a joint model to the features to produce multiple values of the overall score; and
combining the multiple values into a final value of the overall score.
3 . The method of claim 2 , wherein the joint model comprises:
a random forest; and
a gradient-boosted tree.
4 . The method of claim 1 , wherein using multiple subsets of the features to calculate the set of sub-scores for characterizing different components of the overall score comprises:
for each sub-score in the sub-scores, using a different statistical model to calculate the sub-score from a different subset of the features.
5 . The method of claim 4 , wherein using multiple subsets of the features to calculate the set of sub-scores for characterizing different components of the overall score further comprises:
iteratively adjusting one or more of the sub-scores until a sum of the sub-scores equals the overall score.
6 . The method of claim 1 , wherein the sub-scores comprise a similarity score representing a demographic similarity of the customer to existing customers of the educational technology product.
7 . The method of claim 6 , wherein a subset of the features for calculating the similarity score comprises:
a company characteristic;
a potential spending; and
a company statistic.
8 . The method of claim 1 , wherein the sub-scores comprise an engagement score representing a similarity in engagement with an online professional network between the customer and existing customers of the educational technology product.
9 . The method of claim 8 , wherein a subset of the features for calculating the engagement score comprises:
a number of visits to the online professional network;
a number of members of the online professional network;
a number of connections within the online professional network; and
a previous purchase behavior of the customer with one or more other products associated with the online professional network.
10 . The method of claim 1 , wherein the sub-scores comprise a learning culture score representing a similarity in learning culture between the customer and existing customers of the educational technology product.
11 . The method of claim 10 , wherein a subset of the features for calculating the learning culture score comprises:
a connectedness to educational technology entities in an online professional network;
a number of members with skills listed on the online professional network;
a number of learning decision makers; and
a number of e-learning certificates.
12 . An apparatus, comprising:
one or more processors; and
memory storing instructions that, when executed by the one or more processors, cause the apparatus to:
obtain a set of features for a customer of an educational technology product;
use the set of features to calculate an overall score representing a predicted purchase behavior of the customer with the educational technology product;
use multiple subsets of the features to calculate a set of sub-scores that characterize different components of the overall score; and
output the overall score and the sub-scores for use in managing sales activity with the customer.
13 . The apparatus of claim 12 , wherein using the set of features to calculate the overall score comprises:
applying a joint model to the features to produce multiple values of the overall score; and
combining the multiple values into a final value of the overall score.
14 . The system of claim 13 , wherein the joint model comprises:
a random forest; and
a gradient-boosted tree.
15 . The system of claim 12 , wherein using multiple subsets of the features to calculate the set of sub-scores for characterizing different components of the overall score comprises at least one of:
for each sub-score in the sub-scores, using a different statistical model to calculate the sub-score from a different subset of the features; and
iteratively adjusting one or more of the sub-scores until a sum of the sub-scores equals the overall score.
16 . The system of claim 12 , wherein the sub-scores comprise:
a similarity score representing a demographic similarity of the customer to existing customers of the educational technology product;
an engagement score representing a similarity in engagement with an online professional network between the customer and the existing customers; and
a learning culture score representing a similarity in learning culture between the customer and the existing customers.
17 . The apparatus of claim 12 , wherein the memory further stores instructions that, when executed by the one or more processors, cause the apparatus to:
use the set of features to calculate a potential spending of the customer with the educational technology product; and
output the potential spending with the sub-scores and the overall score.
18 . A system, comprising:
an analysis module comprising a non-transitory computer-readable medium storing instructions that, when executed, cause the system to:
obtain a set of features for a customer of an educational technology product;
use the set of features to calculate an overall score representing a predicted purchase behavior of the customer with the educational technology product;
use multiple subsets of the features to calculate a set of sub-scores that characterize different components of the overall score; and
a management module comprising a non-transitory computer-readable medium storing instructions that, when executed, cause the system to output the overall score and the sub-scores for use in managing sales activity with the customer.
19 . The system of claim 18 , wherein the sub-scores comprise:
a similarity score representing a demographic similarity of the customer to existing customers of the educational technology product;
an engagement score representing a similarity in engagement with an online professional network between the customer and the existing customers; and
a learning culture score representing a similarity in learning culture between the customer and the existing customers
20 . The system of claim 18 , wherein using the set of features to calculate the overall score comprises:
applying a joint model to the features to produce multiple values of the overall score; and
combining the multiple values into a final value of the overall score.