T 1954/08 (Marketing simulation/SAP) 06-03-2013

Summary of facts and submissions

I. This appeal is against the decision of the examining division to refuse European patent application No. 04727275.2, entitled "Predicting marketing campaigns using customer-specific response probabilities and response values". The application was filed as international application PCT/EP2004/003930 and published as

A2: WO-A2-2004/090765 (21 October 2004).

II. The examining division refused the application for lack of inventive step (Article 56 EPC 1973) over a notorious data processing system comprising a plurality of terminals and computers connected via a communications network. They considered the marketing prediction method as a business scheme, an abstract administrative plan, an intellectual scheme and a mental act for performing mathematical steps. The division did not identify any technical detail beyond the use of commonplace communication and data processing means.

III. The appellant requests that the decision under appeal be set aside and that the case be remitted to the department of first instance with an order to grant a patent on the basis of one of four amended claim sets (main request, 1st to 3rd auxiliary requests) filed with the statement setting out the grounds of appeal (24 September 2008). Oral proceedings have been requested on an auxiliary basis.

(a) Claim 1 according to the main request reads:

"1. A computer-implemented method of predicting outcomes of marketing campaigns, the method comprising:

determining a response probability for each of a plurality of customers, the customers being intended targets of a marketing campaign;

determining a response value for each of the customers that indicates a predicted value of a response to the marketing campaign by the customer; and

predicting an outcome of the marketing campaign using the response probability and the response value,

wherein at least one campaign step (206) in the marketing campaign comprises a plurality of alternative campaign elements (207, 208, 209),

further comprising assigning the customers to the campaign elements (207, 208, 209) using an optimizing algorithm, and

wherein the optimizing algorithm assigns and reassigns the customers to the campaign elements (207, 208, 209) while evaluating the predicted outcome of the marketing campaign, but does not reassign a customer to a campaign element to which the customer has previously been assigned, and where each assignment of a customer to a campaign element is recorded in a binary map, such that the optimizing algorithm provides a best goal value for the marketing campaign."

(b) The first auxiliary request appends the following paragraph to the end of claim 1 of the main request:

"and where the optimizing algorithm is terminated after a user-defined number of customer reassignments does not improve the most recent best goal value."

(c) The second auxiliary request appends the following three paragraphs to the end of claim 1 of the main request:

"the customer reassignments are recorded in a list after finding the most recent best goal value, where

the optimizing algorithm is terminated after a user-defined number of customer reassignments does not improve the most recent best goal value, and where

upon said termination, the assignment of customers to campaign elements corresponding to the best goal value is determined by reversing all the assignments made since finding the most recent best goal value."

(d) As compared to the second auxiliary request, the third auxiliary request inserts four paragraphs in the middle of claim 1 so that the complete claim reads (italics added to highlight the inserted paragraphs):

"1. A computer-implemented method of predicting outcomes of marketing campaigns, the method comprising:

determining a response probability for each of a plurality of customers, the customers being intended targets of a marketing campaign;

determining a response value for each of the customers that indicates a predicted value of a response to the marketing campaign by the customer; and

predicting an outcome of the marketing campaign using the response probability and the response value,

wherein the marketing campaign comprises at least first and second campaign steps, and wherein predicting the outcome of the marketing campaign further comprises:

using the response probabilities for the plurality of customers to predict a number of responses to be received if the first campaign step were performed toward the plurality of customers;

selecting a target group (168) of customers from the plurality of customers using the response probabilities, where a customer is included in the target group if a randomly generated number is less than the customer’s response probability, the target group (168) being substantially equal to the predicted number of responses; and

predicting an outcome of performing the second campaign step toward the target group (168); and

wherein at least one campaign step (206) in the marketing campaign comprises a plurality of alternative campaign elements (207, 208, 209),

further comprising assigning the customers to the campaign elements (207, 25 208, 209) using an optimizing algorithm, and

wherein the optimizing algorithm assigns and reassigns the customers to the campaign elements (207, 208, 209) while evaluating the predicted outcome of the marketing campaign, but does not reassign a customer to a campaign element to which the customer has previously been assigned, and where each assignment of a customer to a campaign element is recorded in a binary map, such that the optimizing algorithm provides a best goal value for the marketing campaign, where

the customer reassignments are recorded in a list after finding the most recent best goal value, where

the optimizing algorithm is terminated after a user-defined number of customer reassignments does not improve the most recent best goal value, and where

upon said termination, the assignment of customers to campaign elements corresponding to the best goal value is determined by reversing all the assignments made since finding the most recent best goal value."

(e) Technical problem presented in the statement setting out the grounds of appeal

The skilled person was a computer scientist or a computer programmer who received the task of computing the outcome of a marketing campaign comprising a plurality of campaign elements. The business manager who wanted to use the prediction method might require the optimising algorithm to provide a best goal value for the marketing campaign. However, in order to accomplish that task, the computer scientist had to design and implement an algorithm which computed the optimum result. Thus, the choice of algorithm was not dictated by business considerations but was connected to the particular manner of implementation. The algorithm taught a technical professional how to find an acceptable best goal value in a reasonable amount of time. Thus, the skilled person was confronted with the objective technical problem of computing the best goal value for a marketing campaign efficiently and reliably.

(f) Inventiveness argumentation presented in the statement setting out the grounds of appeal

A general-purpose computer was provided with additional functionality without simply choosing an existing method. Claim 1 taught the skilled person how to design and implement an optimising algorithm which finds the best goal value efficiently and reliably. The innovative algorithm was specifically tailored to avoiding a sub-optimal assignment of customers to elements of a marketing campaign. Neither existing computational algorithms (e.g. iterative improvement or simulated annealing), nor mathematical formulae (e.g. a formula for calculating an absolute maximum) suggested the claimed combination of steps.

IV. The Board summoned the appellant to oral proceedings, scheduled for 6 March 2013. In an annex to the summons, the Board voiced its preliminary opinion that claim 1 (all requests) did not involve an inventive step over a general computerised method for processing data according to any existing mathematical algorithm. Finding a maximum revenue or profit value (A2, page 2, paragraph 2) of a simulated marketing campaign appeared to be a commercial rather than a technical purpose. Therefore, the iterative mathematical algorithm according to claim 1 remained a mere mathematical contribution which did not enter into the examination for an inventive step.

V. In response to the summons, the appellant defined an objective technical problem as how to implement a method of predicting outcomes of marketing campaigns on a computer. The manner of implementing such a method required technical considerations. While the skilled person had various ways to implement such a method on a computer, the solution of claim 1 was influenced by technical limitations of a computer (reference to T 258/03-Auction method/HITACHI, OJ EPO 2004, 575, reasons 5.8).

In particular, a human seeking to build a target group of customers would not consider using random numbers. Surprisingly, that method was more efficient than a human approach, particularly for a large data set (i.e. a large number of customers).

The use of a binary map in the context of an optimisation algorithm made it more likely that an optimal set would be determined, and enabled the determination to be performed efficiently. The binary map was used in consideration of technical limitations of a computer system, since the intuitive approach of a human facing a similar problem was difficult to implement in a computer system.

A list of reassignments to return to a previously found goal value (when subsequent iterations failed to find a higher value) was more efficient than storing each set of customer assignments to campaign elements and showed consideration of the memory limitations of a computer system. A set of customer assignments to campaign elements could require storage of one million values (A2, page 16, lines 16 to 18), whereas the method might terminate if 10,000 reassignments did not improve the most recent best goal value (A2, page 17, lines 4/5).

VI. By a letter received 12 February 2013, the representative informed the Board that neither the appellant nor the representative would attend the oral proceedings. A decision according to the state of the file was requested.

VII. The Board held oral proceedings in the appellant's absence on the appointed date (6 March 2013) based on the above-mentioned four requests (filed on 24 September 2008).