FRAMINGHAM (04/18/2000) - Privacy watchdogs may be on red alert as Web sites collect bushels of data about their shoppers and visitors, but they won't have to go into attack mode anytime soon against many of the online retailers.
Most retailers are still struggling to analyze the data they have, sometimes from multiple channels in different databases, so they can figure out how to use that data to effectively personalize content, promotions, marketing campaigns and customer service for their customers, analysts said.
"You're seeing an awakening to the possibilities," said Seamus McAteer, an analyst at New York-based Jupiter Communications Inc. He noted that there is a greater appreciation for "the value that can be extracted from pulling together the various silos that [compose] the typical data architecture of the Web venture."
But companies that have used personalization effectively caution that there are still tough technical challenges.
Jamie O'Neill, chief operations officer at Austin, Texas-based Garden.com Inc., uses transaction information and data that users voluntarily provide to deliver to customers all sorts of useful content, including regional newsletters, tips on gardening in dry or wet climates and even Princess of Wales rose sales promotions.
O'Neill cautions about the difficulties in delivering Web content really fast in real time.
"You have to build really robust data warehouses. Your analytics engine has to be really good to make sure you're clustering intelligently," O'Neill said, adding that there are human challenges, too. "You have to have a lot of really smart people" who understand data analysis and marketing.
SmarterKids.com Inc.'s site wouldn't exist without personalization. The Needham, Massachusetts-based company makes product recommendations for parents based on surveys they fill out or on test results that provide information about their children's learning styles. The company then uses the rich profiles stored in its database to make product recommendations.
SmarterKids can continually expand product attribute tags so that undesirable items can be screened out. For instance, if a parent doesn't want competitive games, those products wouldn't be recommended.
But CEO David Blohm recognizes the limitations of the technology.
"A lot of it is still guesswork," he said. "Looking at all that data, you have to make some inferences. You have to ask the right questions. And you've got to be smart about what you're asking the data to tell you."
For many online retailers, the technology has nothing to do with their decision to take a gradual approach to personalization. They think the risk of delivering content that might puzzle or even alienate customers is still too high.
"I feel very strongly that personalization gone wrong will do more harm than not doing anything at all," said Sally McKenzie, director of merchandising and operations at Eddie Bauer Inc.'s interactive media division in Redmond, Washington.
Eddie Bauer allows visitors to customize their shopping experience by storing their addresses for faster checkout, setting up wish lists and offering a special event reminder service. The Spiegel Inc. unit has plans to expand, letting customers store addresses of gift recipients and indicate products they would like to see.
McKenzie noted that "apparel is a tricky category" and that her company is still in learning mode. With books and CDs, a retailer might be able to predict other artists a customer might like, based on past purchases. But just because a customer bought five sweaters doesn't mean he wants to buy another one, McKenzie said.
"We've been cautious because our products are unique and different," said Kathryn Grant, a senior manager of Internet strategy at Sharper Image Corp. in San Francisco. "We know that just because you bought a Turbo Groomer the last time you were here doesn't necessarily mean you wouldn't want to buy the talking [crystal] ball."