System and method for marketing mix optimization for brand equity management

• US 8,452,639 B2
• Filed: 03/25/2008
• Issued: 05/28/2013
• Est. Priority Date: 01/20/2006
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Abe, Naoki, Pednault, Edwin P
• Primary Examiner(s): Sterrett, Jonathan G
• Application Number:
  US12/054,745
• Publication Number:
  US 20080177621A1
• Time in Patent Office: 1,890 Days
• Field of Search: 705/7, 705/10, 705/7.31, 705/14.52
• US Class Current: 705/7.31
• CPC Class Codes: G06Q 10/06315 Needs-based resource requirements planning or analysis
• CPC Class Codes: G06Q 10/0637 Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0244 Optimization

Abstract

A method of marketing optimization with respect to brand lifetime management formulates a problem of brand equity maximization utilizing Markov Decision Process (MDP) thereby casting brand equity management as a long term regard optimization problem in MDP, The marketing mix is optimized by formulating the mix as actions in MDP and, utilizing historical marketing and transaction data, aspects of the MDP are estimated.

Claims

1. A computer implemented method of marketing optimization with respect to brand lifetime management for a brand having a lifetime and that comprises one or more branded products, comprising the steps performed by a computer of:

storing data that relate to the one or more branded products of the brand but are not tied to individual customers, the data being characterized by impossibility to ascertain individual customers from the data;

wherein the storing step includes storing;

(a) transaction data that contain historical, dated records of transactions, with information that specifies what products were sold and corresponding sales amount with or without specifying profit amount, the data being characterized by impossibility to ascertain individual customers from the data;

(b) marketing data which consist of historical, dated records of marketing actions, the data being characterized by impossibility to ascertain individual customers from the data;

(c) product taxonomy data, the data being characterized by impossibility to ascertain individual customers from the data,

accessing the transaction data, marketing data and product taxonomy data and generating training data;

running a reinforcement learning procedure on the training data to generate a number of lifetime value models;

selecting a value model and, using scoring data on selected product types, scoring the selected product types using the selected model;

for a mix of marketing media, and using the stored data, performing a long term reward optimization in MDP;

optimizing the marketing mix; and

maximizing a net present value of profits and losses over a life cycle of the brand that comprises introducing the brand, developing the brand, maturing of the brand, fading of the brand, the brand driving other brands, the brand being profitable and the brand failing;

outputting an action vector that is a set of marketing actions which is a marketing mix optimization, optimized over the life cycle of the brand.

2. The method recited in claim 1, further comprising the step of displaying the generated models.

3. The method recited in claim 1, further comprising the step of evaluating the generated models and displaying results of the evaluation.

4. The method recited in claim 1, wherein scoring the selected product types is used to determine an optimal marketing action for any product item, given a value function model which is a function that determines an expected lifetime value given a state feature vector and an action feature vector.

5. The method of claim 1, including outputting a vector that specifies investment mix over a set of marketing media that comprises mass marketing, printing, TV, email and catalogues.

6. The method of claim 1, including steps of:

passing the stored transaction data, and marketing data and product taxonomy data, to a Data Preparation Module, wherein the passing is performed by a Data Storage Module;

followed by

generating training data and passing the generated data to a Reinforcement Learning Module, wherein the generating is performed by the Data Preparation Module;

running a batch reinforcement learning procedure on the input training data and generating a number of lifetime value models, wherein the running is performed by the Reinforcement Learning Module;

scoring a selected model, wherein the scoring is performed by a Scoring Module.

7. The method of claim 1, wherein the brand includes one product.

8. The method of claim 1, wherein the brand includes multiple products.