High-precision customer-based targeting by individual usage statistics

• US 20050010472A1
• Filed: 07/08/2003
• Published: 01/13/2005
• Est. Priority Date: 07/08/2003
• Status: Active Grant

• Current Assignee: Inmar-Youtech LLC
• Original Assignee: You Technology Incorporated
• Inventors: Karhinen, Anssi, Wasserman, Eric G., Quatse, Jesse T.
• Granted Patent: US 8,412,566 B2
• Time in Patent Office: Days
• US Class Current: 705/14.130
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0211 Determining the effectiveness of discounts or incentives
• CPC Class Codes: G06Q 30/0254 based on statistics

Abstract

A system for distributing limited numbers of promotional offers to individual customers, the promotional offers being targeted to customers based on the customers'"'"' individual probabilities of accepting the offers in such a way that each customer can receive a limited number of offers that are estimated to be most likely to be acceptable by the customer. Customer-Based targeting analyzes each customer'"'"'s past purchasing behavior relative to a master list of promotional offers made available to all customers. From that master list Customer-Based targeting selects a preset limit of promotional offers for each individual customer according to the likelihood that, given the opportunity to select any offers of the master list, each customer would prefer those few offers selected specifically for the customer. Various techniques are disclosed for providing an offer acceptance probability profile tailored for individual customers for use in the Customer-Based targeting technique. Product groupings and market segments are taken into account. Empirical Bayes techniques are applied to the estimation of the offer acceptance profile, and techniques suitable for handling sparse data are applied. Various marketing strategies are incorporated into the system. A graphical technique is provided for adjusting the offer acceptance profile that enables a user to override a system computation and manually set the relative offer acceptance probabilities for an individual user or class of users.

Claims

1. A method of distributing limited quantities of promotional offers from a plurality of promotional offers to a plurality of customers comprising:

providing, for each combination of customer and promotional offer from said pluralities, a measure of the probability that the customer will accept the promotional offer; and

for at least one customer from said plurality of customers, selecting a limited quantity of offers from said plurality of offers for distribution to said at least one customer,

wherein said limited quantity of offers are selected substantially in descending order of said measures of probabilities provided for all combinations of said at least one customer with said promotional offers.

2. The method of claim 1 wherein the quantity of offers for distribution to said at least one customer is limited by selecting no more than a pre-set fixed number of offers for said at least one customer.

3. The method of claim 1 wherein the quantity of offers for distribution to said at least one customer is limited by selecting only those offers having a measure of probability for said at least one customer greater than or approximately equal to a pre-set fixed cutoff value.

4. The method of claim 1 wherein a limited quantity of offers are selected for distribution to each and every customer of said plurality of customers, the offers for each individual customer being selected substantially in descending order of said measures of probabilities provided for that individual customer.

5. The method of claim 1 wherein a limited quantity of offers are selected for distribution to a selected subplurality of customers from said plurality of customers, the offers for each individual customer of said subplurality being selected substantially in descending order of said measures of probabilities provided for that individual customer.

6. The method of claim 5 further comprising:

partitioning said plurality of customers into a plurality of market segments; and

wherein said subplurality corresponds to one of said market segments.

7. The method of claim 1, further comprising:

grouping a plurality of distinct products into a plurality of product groupings;

providing a product grouping probability profile associating with each said product grouping a measure of the probability that a customer will purchase a product from said product grouping; and

deriving said measure of probability for each said combination of customer and promotional offer from the measure of probability associated with each product grouping containing a product subject to the promotional offer.

8. The method of claim 7, further comprising:

providing access to a transaction history database for at least a substantial portion of said plurality of customers, wherein the database associates with each database customer an identification of transactions engaged in by the database customer and an identification of products previously purchased by the database customer in each of the transactions;

providing a transaction summary data structure associating with each database customer the total number of transactions the database customer has engaged in and the numbers of transactions including each said product grouping;

averaging the product groupings per transaction from said transaction summary data structure for at least a portion of said database customers; and

deriving said measure of probability associated with each said product grouping from the averaged product groupings per transaction for the associated product grouping.

9. The method of claim 7, further comprising:

normalizing said product grouping probability profile for an individual customer to reflect a relative probability of said individual customer purchasing from a product grouping with respect to an average probability for a customer to purchase from said product grouping.

10. A method of distributing limited quantities of promotional offers from a plurality of promotional offers to a plurality of customers comprising:

grouping a plurality of distinct products into a plurality of product groupings;

providing a product grouping probability profile associating with each said product grouping a measure of the purchase probability that a customer will purchase a product from said product grouping;

applying preprogrammed targeting criteria embodying a marketing strategy to said product grouping probability profile to provide a profile of offer scores;

providing, for each combination of customer and promotional offer from said pluralities, a measure of the acceptance probability that the customer will accept the promotional offer,

wherein said measure of acceptance probability is derived from said profile of offer scores;

for at least one customer from said plurality of customers, selecting a limited quantity of offers from said plurality of offers for distribution to said at least one customer,

wherein said limited quantity of offers are selected substantially in descending order of said measures of acceptance probabilities provided for all combinations of said at least one customer and said promotional offers.

11. The method of claim 10, wherein

said marketing strategy includes at least one targeting product grouping and a promoted product grouping linked to said at least one targeting product grouping;

and

said promotional offers are distributed only to customers having a high probability of acceptance for said at least one targeting product grouping.

12. The method of claim 11, further comprising:

providing a taxonomy of said product groupings;

wherein said at least one targeting product grouping is defined in reference to said taxonomy.

13. The method of claim 11, wherein said marketing strategy includes a MoveStock strategy.

14. The method of claim 11, wherein said marketing strategy includes an UpSell strategy.

15. The method of claim 11, wherein said marketing strategy includes a CrossSell strategy.

16. The method of claim 11, wherein said marketing strategy includes a Reward strategy.

17. The method of claim 11, wherein said marketing strategy includes a BrandChange strategy.

18. A method of adjusting the distribution of limited quantities of promotional offers from a plurality of promotional offers to a plurality of customers comprising:

providing, for each combination of customer and promotional offer from said pluralities, a measure of the acceptance probability that the customer will accept the promotional offer;

presenting the measures of acceptance probabilities for an individual customer in a graphical display,

wherein said graphical display includes a plurality of graphic elements, one said graphic element being associated with each said measure of acceptance probability provided for said individual customer at least for the highest ranking of said measures;

enabling adjustment of said measures of acceptance probability by movement of the associated graphic elements; and

selecting a limited quantity of offers from said plurality of offers for distribution to said individual customer,

wherein said limited quantity of offers are selected substantially in descending order of said measures of acceptance probabilities as adjusted in said enabling step.

19. The method of claim 18, wherein said graphical display comprises a bar chart, said graphic elements comprise individual bars of said bar chart, and said movement comprises dragging said bars to lengthen and shorten them and thereby increase and decrease the associated measure of acceptance probability.

20. A method of distributing limited quantities of promotional offers from a plurality of promotional offers to a plurality of customers utilizing a transaction history database for at least a substantial portion of said plurality of customers, wherein the database associates with each database customer an identification of transactions engaged in by the database customer and an identification of products previously purchased by the database customer in each of the transactions, said method comprising:

deriving a historical purchase probability profile from said transaction history database for at least a portion of the customers in said database and for a plurality of product groupings in said database, said historical purchase probability profile providing for each individual customer and for each individual product grouping a measure of the probability that said individual customer will purchase a product from said individual product grouping;

applying a statistical model to said purchase probability profile for a given individual customer to determine estimated probabilities that said given individual customer will purchase one or more products from said product groupings;

selecting a limited quantity of offers from said plurality of offers for distribution to said given individual customer,

wherein said limited quantity of offers is selected substantially in descending order of said estimated probabilities.

21. The method of claim 20 wherein said statistical model is an empirical Bayesian statistical model.

22. The method of claim 20 wherein one or more of said product groupings includes one and only one product.