Imagine the margin implications when you’re selling a million or more $45 to $55 CD packages every year, with few if any royalties. Titles like Mood Music That Will Live Forever. Now add to that no list costs. The CDs are being marketed entirely to the house file, customers who have already purchased at least one of your products. Clearly, the product line is a cash cow, and even under these circumstances you are able to squeeze out still more profit.

I’m speaking of Reader’s Digest, of course, undoubtedly one of the foremost practitioners of statistical modeling in the direct marketing industry. With uncanny accuracy, Digest managers are able to score the already well-segmented house file and know precisely where to make their cutoff to ensure that each “bucket” of names is profitable. To replicate what Reader’s Digest has established during the past 30 years is daunting task, but given the potential rewards, more and more businesses are making the attempt and now Dr. Ronald Drozdenko, professor of marketing and department Co-Chair at Western Connecticut State University, and Perry Drake, a long-time Digest veteran, have teamed up to write a new book on the subject, Optimal Database Marketing: Strategy, Development and Data Mining.

Until now, The New Direct Marketing: How to Implement a Profit-Driven Database Marketing Strategy by David Shepard Associates, has justifiably billed itself as “The only text to approach direct marketing statistics and predictive segmentation modeling from a uniquely non-mathematical perspective.” Now in its third edition (1999), The New Direct Marketing remains a classic, but with the publication of the Drozdenko and Drake book, the “only” claim is no longer valid.

Written for Business Drake and Drozdenko, who is also an advisor to the Direct Marketing Educational Foundation, have collaborated to produce a highly readable and useful book. Like the Shepard book, Optimal Database Marketing is written for business people, and provides a framework for developing marketing databases, and understanding the role they can play in helping an organization achieve its objectives. Unlike the Shepard book, however, it is not intended as an encyclopedic text. Instead Drozdenko and Drake follow the career of a fictitious marketer, Keri Lee, as she rises through the ranks from Account Executive to Corporate Vice President. This narrative technique sets the stage in each chapter for the resolution of Keri’s latest business challenges, which increase in complexity as she assumes greater and greater responsibility within the organization.

To begin with, we learn one of the fundamentals of successful database marketing: planning must precede the development of the database. “Marketing database development should start with [a] strategic plan. The strategic plan provides a guide for database development and defines how it will assist in achieving organizational objectives in the near and distant future.”

The first five chapters are devoted to a detailed discussion of the key considerations in database design and development:

• Performing a situational analysis
• Specifying objectives
• Developing strategies
• Implementing marketing programs
• Monitoring and control

Since database development can be costly in terms of both human and financial resources, the more careful the planning, the more likely the database will be worth the investment.

Once Your Database Is In Place
It’s once your marketing database is in place that all your efforts start to pay off. Here the authors begin with an explanation of sampling methodologies and then go on to explain how the samples thus derived can be analyzed using techniques such as univariate tabulation, cross-tabulation, and correlation analysis. This is the kind of information that enables marketers to characterize and distinguish between individuals, such as responders and non-responders to a marketing campaign, information that can then be used to refine future campaigns.

From sampling, the authors move on to database segmentation. The goal of segmentation is to identify different groups within a database based on behavior or demographics. “The underlying premise for segmentation is that not all customers residing on the database are alike and therefore should not be treated alike.” To the extent that a marketer can correctly identify unique market segments and develop appropriate campaigns to target them, the more effective they will be. Segmentation can be achieved using a variety of techniques such as the univariate and cross-tabulation methods mentioned above, as well as formal RFM (Recency, Frequency, Monetary Value), CHAID, factor, and cluster analysis, each of which is described at length.

All of this provides the foundation for understanding the most powerful tool in the analyst’s toolkit: multiple regression modeling. Multiple regression modeling enables the marketer to predict the probability that a customer will perform an action. As the authors put it: “Think of a multiple regression model as a set of specific criteria used to create a score for each customer that indicates their relative likelihood of ordering (or paying, canceling, renewing, etc.).” In particular, logistic regression models, which use binary variables such as ordered/did not order, are highly favored for direct marketing applications. Employing regression models allows a marketer to select only those names that score high enough to meet the objectives of the marketing program, reducing costs and increasing the likelihood of achieving a profitable campaign.

No Magic Box
Throughout the book, the authors describe and illustrate several statistical applications, including SAS, SPSS, Excel as well as some of the newer data mining tools. Their inclusion, however, elicits a warning: “These tools have one thing in common: they are statistically based and therefore should be operated by someone familiar with the applications such as a data analyst or trained statistician. Data mining tools are not a magic box into which raw customer data can be fed and out comes the solution.”

The strength of the Drozdenko and Drake book is that it is an orderly and focused approach to a complicated subject. Although direct marketers claim to know all about modeling, in truth, relatively few actually make use of modeling techniques. According to Direct magazine’s 2001 subscriber survey, only 31% (up from 27% the prior year) of consumer and business-to-business database marketers use such techniques. Reasons for this, as the book points out, include both the difficulty of the current statistical packages and the inadequacy of the database itself. Regardless of the techniques involved, any modeling based on a database in disarray will be unreliable. The authors encourage their readers to think through all of the issues involved in building a marketing database, beginning by asking the most fundamental question of all: Why are we building it?

Optimal Database Marketing is an excellent guide to aspiring database marketers, and is sure to take its place alongside the Shepard book as a classic of the industry. The principles of direct marketing haven’t actually changed much. What is changing are the tools used to manage direct marketing, and the more powerful they become, the more concepts like one-to-one marketing become attainable.

----------

Jane Weber is the founder of Ask Jane Direct, LLC, which works with companies interested in implementing marketing technology solutions. She also serves on the Board of Directors of the Direct Marketing Club of New York, and on the DMA International Echo Awards Executive Committee.

More Features