Filter out archaic, one-size-fits-all strategies

March 27, 2000

BY MICHAEL KRAUSS

I used to be a short but proud Keebler cookie Elf. At least, that’s how I fancied myself years ago when I was a new product development manager working at 1 Hollow Tree Lane in Elmhurst, Ill.

We did a lot of collaborating as we worked to create Keebler’s flagship chocolate-chip cookie brand, Chips Deluxe. We collaborated with R&D on the product specifications, with manufacturing and distribution and with the ad agency. We even collaborated pretty well with the finance and the sales departments.

It took such collaboration to bring our child—our product—to market, and we delivered a good-tasting cookie with big, soft chocolate chips (bigger chips than in Chips Ahoy, if memory serves me right). Consistency was the key, tonnage was the name of the game and customizing the product offer to individual buyers wasn’t anywhere on our horizon.

Of course, there were voices in the wilderness who predicted we were missing something in our product development efforts. If we’d listened to Alvin Toffler and read Future Shock [ISBN# 0-553-27737-5] back in 1970, we might have realized that customization and personalization of the product offer would become significant.

Professor and author Stan Davis laid the foundation and coined the term "mass customization," which we thought to be an oxymoron when his 1987 book, Future Perfect [ISBN# 0-201-11513-1], was published.

Consultant Joe Pine explained this phenomenon in his 1993 work, Mass Customization [ISBN# 0-87584-372-7] and, of course Don Peppers and Martha Rogers filled our marketing heads with the idea of The One to One Future [ISBN# 0-385-48566-2] in 1993.

Back in 1997, Clay Christensen’s, The Innovator’s Dilemma [ISBN# 0-87584-585-1] hit the charts. He plainly spelled out the impact of today’s disruptive new technologies and why we marketers thought we were doing everything right, but in fact were missing the biggest opportunities in product development.

Then Pine returned last year with his Experience Economy [ISBN# 0-87584-819-2], suggesting that rich environments and consumer experiences are the pathway to high margins and marketing success.

Thanks to the rise of the Internet and today’s new technologies, it’s possible to conceptualize, develop and deliver customized product offers and experiences for individual buyers, just as this train of authors exhorted us to do all along.

Actually, the technology making this feat possible was quietly invented in 1968; that was the year the global publishing industry created the ISBN (International Standard Book Number) system. By establishing a unique labeling system for all books and published monographs, the foundation for Jeff Bezos’ future fortune was laid. It isn’t complex for a software program to examine a pool of customers’ purchases by ISBN number. From relatively small amounts of data, the system can make some pretty good recommendations, helping online marketers enhance the buyer’s experience while customizing their offerings to the individual.

Pattie Maes, director of the Software Agents Group at Cambridge, Mass.-based Massachusetts Institute of Technology Media Lab, was thinking along those lines when she founded Firefly Network, Inc. (www.firefly.net) in Cambridge—one of the first companies to commercialize software agent technology. (Firefly was acquired by Microsoft in April 1998.)

Says Maes in a recent Association for Computing Machinery Journal article, "Firefly recommends products through an automated ‘word-of-mouth’ recommendation mechanism called ‘collaborative filtering.’ Firefly uses the opinions of like-minded people to offer recommendations of such commodity products as music and books, as well as more difficult-to-characterize products, such as Web pages and restaurants."

A former physics department programmer at Philadelphia’s University of Pennsylvania got the concept. Steven Snyder founded Minneapolis-based Net Perceptions Inc. in July 1996. His clients (including CDnow, Egghead.com and Ticketmaster) use collaborative filtering to learn from each customer interaction to adjust marketing messages and product offers in real time.

Delivering customized product offers to individual customers requires that marketers think in new ways, discover and adopt new skills and learn about a whole new set of marketing services suppliers. In addition to buying advertising from Young & Rubicam, Giant Step or Leagas Delaney, you may have to buy a whole new technology platform from Firefly, Net Perceptions, Broadvision, Blue Martini, Engage, LikeMinds or Vignette. Firms such as these have developed sophisticated software for mining customer data and delivering customized offers.

Traditionally, direct mail and database marketers—the forerunners of today’s online marketers—coded into their applications a set of decision rules, and these so-called rules- based systems assessed you as a shopper. Based on a set of predetermined guidelines—defined by the presumably prescient marketing manager—these sites provided you with a set of customized product offers. The possibilities for customizing offers using a rules-based approach is limited only by the creativity of the marketing manager. And the problem with rules-based systems is they are only as good as the individual creativity of the marketing or product development manager or team; these systems don’t take full advantage of the community’s buying power.

Using collaborative filtering, however, the marketing manager can better leverage the power of the community of shoppers that come to a Web site. Essentially, the marketing manager can engage the community in determining future product offers.

A step ahead, but while collaborative filtering applications have gained broad acceptance, "There are limitations," says Bennet Harvey, vice president of product management for Esurance.com, an online personal lines insurance provider based in San Francisco.

"At Esurance.com, we want to provide a totally customized product offer for each and every shopper. Techniques like collaborative filtering have their place, but solutions such as Net Perceptions’ tend to be developed for particular vertical industries. They don’t always travel well."

Another concern with these techniques, Harvey adds, is that "they serve up offers based on assumptions of buyer similarity. At Esurance.com … we don’t want to make offers based on other customers serving as surrogates for the buyer."

While there are advocates on both sides of this argument, some things are clear. First, the days of selling a one-size-fits-all product offer—even if it’s for chocolate chip cookies—are long gone. And, second, the new product development manager needs to start collaborating with state-of-the-art technologists and learn about these new techniques—especially those who live and work, like I did, in a Hollow Tree.

Michael Krauss is a partner with Diamond Technology Partners in Chicago.
He can be reached at news@ama.org.