Marketing Workbench Laboratory

Professor John M. McCann
Fuqua School of Business
Duke University

I formed the Marketing Workbench Laboratory (MWL) at the Fuqua School of Business in 1986 following the publication of The Marketing Workbench, a book about the use of data and computers in consumer package goods marketing and sales. In that book, I made the following forecasts:

• Scanner databases will explode in size as they move from market-level data to account-level data, and eventually to store-level data.
• Marketing will become more regional in nature.
• Power will shift towards the retailers.
• Manufacturers will reorganize in ways that will blur the distinctions between marketing and sales.
• Brand groups and sales managers will be confronted with data overload.

These forecasts have come true, as discussed in The Changing Nature of Consumer Goods Marketing and Sales.

I went on to conclude that firms need to move beyond their focus on Management Information Systems and look to expert systems as a say to deal with these upcoming conditions. The lab was formed to learn expert system technologies, methods, and philosophies so we could invent new types of computer systems for use in the sales and marketing departments in consumer goods firms.

It focused on computer applications which would improve the effectiveness of managers and analysts in the analysis, modeling, and implementation of market-place information. The firms in the industry responded to the lab's formation by agreeing to become corporate sponsors and thus provide significant financial support to the lab's operations. The following is a list of the sponsoring firms:

A.C. Nielsen           Beechum                    Bristol-Myers       
Campbell Soup          Carnation                  Clorox              
Coca Cola              Colgate Palmolive          Frito-Lay           
General Foods          General Mills              Giant Foods         
IBM                    James River                Johnson & Johnson   
K-Mart                 Kraft                      Kroger              
Nabisco                Ocean Spray Cranberries    Ogilvy & Mather     
Pillsbury              Procter & Gamble           Quaker Oats         
R.J. Reynolds Tobacco  SAMI Marketing Research    Sara Lee            
Steinberg              Sterling Drug              Thrifty Drug        

Managers from these firms attended quarterly meetings where they reviewed the MWL's research conducted during the previous three months and made suggestions for future research. In addition, one or more of the authors of this hyperbook visited almost all of these firms for presentations and further discussions. Specific projects were undertaken and completed with several of the firms.

Additional seminars and presentations were held at Duke under the sponsorship of IBM in which the MWL's research results were presented and discussed. As a result of these activities, hundreds of marketing, sales, merchandising, marketing research, and information system managers have been exposed to the work reported in this volume. Because these applications evolved over a period of four years, these professionals had an opportunity to influence the work through their comments and suggestions.

Early work of the Laboratory is reported in Expert Systems for Scanner Data Environments, Kluwer Academic Publishers, 1990. The following is the Preface of that book:

"This book is about the role of expert systems in marketing,
particularly in the consumer goods industry. Section I describes
the changing nature of consumer marketing and presents the rationale
and need for expert systems. The remainder of the book combines
a tutorial on expert systems with a series of expert system prototypes.
The tutorial material is presented in three places. First, section
II is devoted to introducing expert systems in general. Chapter
3 provides a general introduction to the topic, which is contained
in Chapter 4 where a small expert system (the Promotion Advisor)
is used to illustrate the important features of a backward-chaining,
rule-based system. The promotion theme is extended in Chapter
5 where a larger system is presented. The material in all three
of these chapters was designed as an introduction and tutorial
on the most common technology for building applied expert systems:
the backward-chaining, rule-based inference engine.
Tutorial material is also contained in the body of the chapters
that describe the prototypes. This material is usually in the
form of sample rules and a description of the process for applying
the rules. 

The third location of the expert system material is
in chapters that follow discussions of the prototypes. Chapter
7 is a technical chapter on the coupling of expert systems to
traditional systems. Chapter 11 describes model-based reasoning
and its application to DEALMAKER, which is presented in Chapter
10. This material introduces frame-based systems, which go beyond
the backward-chaining, rule-based inference engine. Chapter 12
describes additional research in the application of expert systems
to marketing problems. Chapter 13 discusses the implementation
of marketing expert systems in a number of different firms."

This concepts and computer applications discussed in this hyperbook should be useful to manufacturers such as P&G. Another book discusses the use of databases and knowledge systems by retailers: John M. McCann and John P. Gallagher, Databases and Knowledge Systems in Merchandising, Van Nostrand Reinhold, 1991.