2007 Mar 01
Marketing Mix Modeling
David M. Raab
DM Review
March 2007

Consumer packaged goods marketers have generally been considered laggards at customer relationship management. The fundamental problem is they sell through retailers and therefore have little direct contact with their customers. As a result, viagra buy marketing technologies like customer databases, vialis 40mg targeting systems and real-time interaction management have been pioneered in industries that do have direct customer contact: financial services, order telecommunications, retail and travel.

Yet packaged goods marketers have invested heavily in technologies suited to their industry. These include marketing resource management, consumer surveys, and models to estimate the impact of marketing spending.

Today, such technologies are increasingly employed outside of the packaged goods industry. The reason is simple: as the number and variety of channels increases, even marketers selling directly to customers are finding it harder to contacts with each customer. In other words, their world is looking more like consumer packaged goods.

The furthest advanced example of this technology transfer is marketing resource management systems. Originally developed to help packaged goods marketers manage their huge, world-wide advertising programs and related marketing materials, their value is increasingly apparent to other marketers who find themselves running many more campaigns across multiple channels. Basic marketing resource management features—planning, budgeting, project management, content management, campaign execution and analysis—are now part of most major customer management software.

We are just starting to see a similar transfer in marketing results measurement. Here, the relevant consumer packaged goods technology is marketing mix modeling.

Marketing mix models use statistical techniques to estimate the impact on sales of different promotion elements. These elements include advertising by channel, pricing, product features, product assortment, retail distribution, and in-store promotions. To explain sales accurately, marketing mix models for consumer goods usually take into account competitors’ activities in addition to the company’s own. The competitive data is acquired from syndicated compilers like Nielsen and Information Resources, Inc. Models also include other external inputs such as weather conditions. The modeling system is loaded with historical data and tuned until it can simulate previous results satisfactorily. The model is then used to estimate the impact of proposed changes in different marketing components. Typically, users develop scenarios to test alternative resource allocations. Some systems can automatically devise an optimal plan within user-specified constraints.

Nearly all model inputs are related to geography—store locations, regional promotions, local media markets, weather, and so on. Models therefore treat each region separately. To the extent that different regions differ in demographic characteristics such as average income, education level, racial composition or population density, the models can infer gross relationships between customer characteristics and marketing results. But the input data does not identify individual consumers, so the precise segmentation and response analysis of traditional customer management systems are not possible.

Marketing mix modeling is still very much an art. Companies including Upper Quadrant (www.upperquadrant.com), M-Factor (www.m-factor.com), Marketing Management Analytics (www.mma.com), SAS (www.sas.com, through its acquisition of Veridiem), and DecisionPower (www.decisionpower.com) position themselves as software vendors, but their systems mostly allow users to do reporting and analysis after the models themselves are built by experts at the vendor. Other vendors present themselves more as consultants, including Hudson River Group (www.hudsonrivergroup.com), Analytic Partners (www.analyticpartners.com), Pointlogic (www.pointlogic.com), Marketing Analytics Inc. (www.marketinganalytics.com), Copernicus Marketing Consulting (www.copernicusmarketing.com), Strategic Oxygen (www.strategicoxygen.com), iknowtion (www.iknowtion.com), and Management Science Associates, Inc. (www.msa.com). Yet most of these firms also provide software to help clients test alternatives and review results.

At most vendors, marketing mix modeling is part of a larger suite of services aimed at optimizing return on marketing investment. These can include reporting systems that track actual results, compare them with expectations, and prepare updated projections. Other components of a marketing optimization suite may help manage market research, analyze customer preferences, test advertising messages, forecast new product sales, simulate business results across multiple product lines, and estimate brand equity. A few vendors also provide traditional customer-level segmentation and response modeling.

A customer-level marketing database isn’t needed for marketing mix modeling, although such a database can be a convenient source for sales information and promotion history. In fact, customer-level data may be the only practical way to determine the geographic distribution of direct customer contacts such as mail or Web advertisements.

Apart from a customer database, the main technical challenge in marketing mix modeling is integrating the external data feeds. This is usually handled by the vendor providing the modeling service, who assembles the data within the modeling system itself. Clients are typically limited to accessing results and setting scenario parameters.

As marketing mix modeling becomes more common, we can expect clients to demand a larger role in the process. Experts will debate whether highly automated, client-driven systems are really feasible or desirable. But such systems will emerge regardless as some clients accept less accurate results in return for quicker turnaround, lower cost and greater control. In the longer run, marketing mix models, and marketing optimization in general, will become part of the larger marketing resource management suites. This hasn’t happened yet because too few clients are interested. But that will change as all types of marketers realize that marketing mix modeling is the only way they can optimize their investments in a multi-channel world.

* * *

David M. Raab is a Principal at Raab Associates Inc., a consultancy specializing in marketing technology and analytics. He can be reached at draab@raabassociates.com.

Leave a Reply

You must be logged in to post a comment.