Apparatus and method for enhanced message targeting

Patent metadata

Field	Value
Publication number	US-11308506-B2
Application number	US-202016984785-A
Country	US
Kind code	B2
Filing date	Aug 4, 2020
Priority date	Mar 30, 2015
Publication date	Apr 19, 2022
Grant date	Apr 19, 2022

Abstract

A method, apparatus, and computer program product are disclosed for improved promotion targeting. An example apparatus includes a processor configured to cause retrieval of historical information regarding transactions associated with a plurality of identifier entities. The example apparatus further includes modeling circuitry configured to train a statistical model using the retrieved historical information, and estimate, using the statistical model, values for expected identifier entity transaction requests associated with each of the plurality of identifier entities. The processor of the example apparatus may further be configured to select, based on the estimated values, a subset of identifier entities to receive impressions of the promotion, and the apparatus may further include communications circuitry configured to transmit an impression of the promotion to each identifier entity in the subset.

First claim

What is claimed is:

1. An apparatus for improved machine learning with a statistical model to facilitate improved targeting of a promotion, the apparatus comprising:
a processor of a server of a promotion and marketing service, wherein the processor is configured to cause retrieval, by communications circuitry of the server and from a database of the promotion and marketing service, of historical information regarding transactions associated with a plurality of identifier entities, wherein each identifier entity represents a specific consumer and is associated with one or more consumer devices used by the consumer to access the promotion and marketing service, and wherein the historical information includes measured characteristics including, for each identifier entity, (i) a number of purchases associated with the identifier entity within a predetermined period of time; and
modeling circuitry of the server of the promotion and marketing service, wherein the modeling circuitry is configured to train the statistical model using the retrieved historical information, wherein training includes predicting expected incremental booking values by offering the promotion to the plurality of identifier entities, wherein the training includes using a retrieved historical information data set describing a control group of consumers who are not provided with a promotion and a target group of consumers who are provided with the promotion to estimate expected impact of the measured characteristics on expected incremental booking value, and for a given promotion the model predictively analyzes the measured characteristics of a selected consumer and predicts the expected incremental booking value associated with the selected consumer, and estimate, using the statistical model, values for expected identifier entity transaction requests associated with each of the plurality of identifier entities, and use the values in order to train the model to predict the expected incremental booking value.

2. The apparatus of claim 1 , wherein the processor is configured to select, based on the estimated values, a subset of identifier entities to receive impressions of the promotion and to select the subset of identifier entities by:
ranking the plurality of identifier entities; and
selecting a predetermined percentage of highest ranked identifier entities,
wherein the subset of identifier entities to target comprises the predetermined percentage of highest ranked identifier entities.

3. The apparatus of claim 2 , wherein the processor is configured to select the subset of identifier entities that maximizes expected revenue generated by the promotion.

4. The apparatus of claim 2 , further comprising:
communications circuitry configured to transmit an impression of the promotion to each identifier entity in the subset.

5. The apparatus of claim 1 , wherein the processor is configured to cause the retrieval of the historical information from at least one of the plurality of identifier entities, one or more merchants, or a memory.

6. The apparatus of claim 5 , wherein the retrieved historical information further includes, for each identifier entity, at least one of: (ii) a number of purchases associated with the identifier entity through the promotion and marketing service, (iii) click indications regarding a number of impressions associated with the identifier entity, (iv) a total revenue associated with the identifier entity over all promotions, (v) registration information associated with the identifier entity, (vi) email receipt indications regarding a number of email impressions associated with the identifier entity, and (vii) impression receipt indications associated with the identifier entity; (viii) a number of promotions purchased by the identifier entity within three zip codes of a given location within a predetermined period of time; (ix) an average cost of transactions associated with the identifier entity; (x) a number of active channel subscriptions associated with the identifier entity; (xi) a total number of bookings associated with the identifier entity; (xii) a highest price of a promotion purchased from the promotion and marketing service that is associated with the identifier entity; (xiii) impression receipt indications regarding a number of promotions in a predetermined price range associated with the identifier entity; and (xiv) a median price of bookings associated with the identifier entity.

7. The apparatus of claim 1 , wherein the statistical model comprises an ensemble learning model.

8. The apparatus of claim 7 , wherein the ensemble learning model comprises a gradient boosted regression model.

9. A method for improved machine learning with a statistical model to facilitate improved targeting of a promotion, the method comprising:
retrieving, by a processor of a server of a promotion and marketing service and from a database of the promotion and marketing service, historical information regarding transactions associated with a plurality of identifier entities, wherein each identifier entity represents a specific consumer and is associated with one or more consumer devices used by the consumer to access the promotion and marketing service, and wherein the historical information includes measured characteristics including, for each identifier entity, (i) a number of purchases associated with the identifier entity within a predetermined period of time;
training, by modeling circuitry of the server of the promotion and marketing service, the statistical model using the retrieved historical information; wherein the training includes predicting expected incremental booking values by offering the promotion to the plurality of identifier entities, wherein the training includes using a retrieved historical information data set describing a control group of consumers who are not provided with a promotion and a target group of consumers who are provided with the promotion to estimate expected impact of the measured characteristics on expected incremental booking value, and for a given promotion the model predictively analyzes the measured characteristics of a selected consumer and predicts the expected incremental booking value associated with the selected consumer,
estimating, by the modeling circuitry and using the statistical model, values for expected identifier entity transaction requests associated with each of the plurality of identifier entities; and use the values in order to train the model to predict the expected incremental booking value.

10. The method of claim 9 , comprising selecting, by the processor and based on the estimated values, a subset of identifier entities to receive impressions of the promotion, wherein selecting the subset of identifier entities includes: ranking the plurality of identifier entities; and
selecting a predetermined percentage of highest ranked identifier entities, wherein the subset of identifier entities to target comprises the predetermined percentage of highest ranked identifier entities.

11. The method of claim 10 , wherein selecting the subset of identifier entities maximizes expected revenue generated by the promotion.

12. The method of claim 10 , further comprising: transmitting, by communications circuitry, an impression of the promotion to each identifier entity in the subset.

13. The method of claim 9 , wherein retrieving historical information regarding the transactions associated with the plurality of identifier entities comprises:
retrieving the historical information from at least one of the plurality of identifier entities, one or more merchants, or a memory.

14. The method of claim 13 , wherein the r

Assignees

• Groupon Inc

Inventors

• Parekh Rajesh Girish
• Balakrishnan Raju
• Zhang Guoxian
• Gaurav Prashant

Classifications

• G06Q30/0201 Primary

  Market modelling; Market analysis; Collecting market data · CPC title

• G06Q30/0211

  Determining the effectiveness of discounts or incentives · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N20/20

  Ensemble learning · CPC title