Don't Shout, Listen

For Hudson and the rest of his team, the Whitestrips brand lent itself particularly well to market testing via the Web. Although Whitestrips fall under the Crest umbrella, the team wanted to put some distance between the new product and the established brand. At $44 per package, the new product costs far more than any other Crest product. And Whitestrips -- pieces of thin plastic material coated with hydrogen peroxide -- require users to absorb more information than most Crest products do. The flexible, interactive nature of the Web promised to make it easier for the team to experiment with those variables.

In early 2000, Hudson persuaded Michael Kehoe, his boss's boss, to invest $2 million in creating Whitestrips.com, a site where consumers could learn about the bleaching product and then buy it. That was a huge investment -- all the more so because Hudson wanted to test the product in traditional, real-world venues as well. "It was like being an oral-care venture capitalist," says Kehoe, 44, vice president of global oral care. "I remember thinking, I'm like the guy in the IBM commercial: 'Are we really ready for this?' "

Kehoe, Hudson, and their colleagues were ready enough. In July 2000, they started selling Whitestrips through dentists' offices and through brick-and-mortar retailers in two classic test cities: Grand Junction, Colorado and Pittsfield, Massachusetts. Then, two months later, Whitestrips.com opened for business, offering the product at the same price that it retailed for in test-city stores. Over the next nine months, the site logged 1 million users and achieved a conversion rate of 12% -- quadruple the rate at which most consumer Web sites are able to turn visitors into buyers.

But as with any test market, the main goal here was not near-term sales but long-term insight. For Hudson and his team, creating Whitestrips.com was a way to get close to potential customers -- to acquaint them with the Whitestrips brand, to learn more about who they are, to listen in on their responses to the product. "We collected demographic information at the site and with follow-up research to purchasers who opted in," says Hudson. "When it comes to pinpointing our high-leverage target market, this information is incredible."

Testing via the Web enabled the Whitestrips team to build a customer base that was more diverse, and hence more reliable, than most test populations are. "We knew that we'd have an accelerated curve of adoption once we went retail," says Hudson. Traditionally, in its test-market programs, P&G has focused on putting its products in cities where the "average" U.S. household (white, two parents, two kids) is disproportionately common. But there's no longer any such thing as an average U.S. household. Using the Web, P&G is broadening the demographic reach of its research to include people of color, people from nontraditional families, and people in major urban areas.

The virtual test market also gave Hudson and his colleagues quick, useful feedback on how traditional marketing efforts were faring. During the test period, Hudson placed ads for Whitestrips in various magazines, including People -- not a publication in which P&G would ordinarily advertise that kind of product. Each ad featured a special code, and readers were offered a discount on the product in exchange for visiting Whitestrips.com and entering that code. By tracking such codes, Hudson learned that lots of hits were coming from People readers. "People was on fire," Hudson says. "We had to change the way we were buying media. The Internet was like sonar technology. It was picking up on something that we never would have heard without it."

Members of the Whitestrips team took away another, more fundamental lesson from their experiment with virtual testing: They learned that to build a new brand in a new product category, they didn't need to live off the reputation of an old brand. "We borrowed the equity of the Crest name, because it gave the product a kind of health credibility," says Kehoe. "But we knew that the Whitestrips name could live on its own -- after we started selling it on the Net."

Hide the Brand?

Could traditional branding be a kind of noise -- a form of communication that actually hinders a company's ability to listen to customers? These days, P&G is willing to entertain even that heretical notion. In its boldest experiment yet with using the Web to reach consumers, the company has invested in a site that all but conceals any association with P&G or P&G-style branding. That site, Reflect.com, lets women "brand" their own versions of makeup, perfume, and other beauty-care products. It also lets P&G begin to map what may be the next frontier of consumer-product marketing: mass customization.