Hanssens, Dominique M., Leonard J. Parsons, and Randall L. Schultz. Boston: Kluwer Academic Publishers (2001)

In order to provide our clients with the most accurate sales models possible, Marketing Analytics Inc. does a lot of model testing and evaluation. This helps us, it helps our clients, and additionally, we’re proud that it has made us valued contributors in the academic as well as the business sales modeling community.

The recently released 2nd edition of the respected sales modeling textbook: Market Response Models, Econometric and Time Series Analysis, references a number of the contributions Marketing Analytics Inc. has made in the sales modeling field. Our findings are cited on proper aggregation in marketing mix models, model testing and evaluation methods, modeling seasonality, pantry loading, and shelf space.

Finally, we are especially proud that a Coefficient Generator™ application was used as a case study of a successfully implemented sales modeling system. This price and promotion system Marketing Analytics Inc. implemented for Kraft Foods, a recognized leader in fact-based marketing, was our first major automated modeling project. This is why we really do modeling methodology research: to help our clients make better business decisions.

Market Response Models, Econometric and Time Series Analysis 2nd Edition by Dominique Hanssens of UCLA, Len Parsons of Georgia Tech, and Randy Schultz of the University of Iowa, ISBN #0792378261, published 2001, is available from Kluwer Academic Publishers, 781-871-6600, http://www.wkap.nl /p>

At Kraft Foods, technological advancements have made it possible to easily and rapidly estimate price-promotion models across a wide portfolio of products and categories. The result of these models can be used by each product team for price and promotion planning. This "mass application" of standard price and promotion models enables Kraft to have ready information at the point of decision making about the implications of marketing actions. There have been two keys to the positive impact of market response modeling on promotion planning at Kraft.

Planners not Analysts. The key organizational decision made at Kraft has been to hire and develop skilled analysts, and then use them in the role of planning rather than as analysts only. While their analytical skills are critical, just as critical is their ability to sit at the small table where decisions are really made. There they are advocates for what the scanner data and other information sources are saying would happen as a result of marketing actions by Kraft or its competitors. Category Business Teams bring small group, fast, cross-functional management to a wide range of categories and products by including representatives from Operations, Advertising, Sales, Marketing, Finance, Product Development and Marketing Information. The Category Business Teams depend on the Marketing Information group to provide knowledge and insight based on market data. The models and forecasting tools Marketing Information uses enable them to have a data-driven-voice in the decision making.

Standardized and Automated Modeling. Kraft has invested in software to provide automated and standard modeling results and on the information infrastructure to rapidly feed data into this software. This investment has made possible the "mass production" of standard econometric models of price and sales promotion across hundreds of product groupings in dozens of product categories. With a strong partnership with ACNielsen to provide the data and modeling software from Marketing Analytics Inc., Kraft has been able to implement consistent models across all categories. While a standard model does not meet all the needs of each brand (each one has different issues to study), it does provide a common basis and methodology from which the company can address the special concerns of each business. This also enables Kraft to provide a common benchmark with which to compare products across categories and divisions of the company. One additional benefit of this process of automated modeling is a cross-sectional database of model results that can be further analyzed to produce company-wide insight into the effect of key measures, such as price elasticity and trade merchandising effectiveness, on brand performance. As Kraft continues to build its information base of model results, these "meta-analyses" of lessons learned will be more and more important in establishing norms and expectations for the businesses of the company.