Paid clicks on ads across Google-owned sites and its advertising network jumped 33% during the quarter.
(Page 2 of 3)
When people log on, the site asks such things as what types of music they listen to, who their favorite performers are and how often they want to receive e-mail communications. For example, if a visitor says she likes the blues and the Indigo Girls, she’ll receive a monthly e-mail notifying her of new blues releases and reviews, discounts on Indigo Girls CDs and other information. On her next visit, the site creates Web pages designed around her interests.
The second type of data is gathered by consumers’ activity on Web sites. The CDnow/N2K site tracks which CDs a customer has purchased, which sound samples they have listened to and which articles they’ve read. That information is then added to the customer’s profile to generate more suggestions
and other information links. Personalization may soon be as common a function on Web store sites as electronic catalogs and shopping carts.
But building customer profiles and tracking consumers’ movements through a Web site raises privacy concerns, and retailers have to protect personal data collected on the Web the same way they protect consumer data collected at the point of sale and in other kinds of direct marketing.
“The customer only wants to give information about himself or herself if they feel confident you will use it to service them better, not to take advantage of them or make money off of it,” Pakman says. “So we do not sell or distribute customer information. That is our prize information, and we keep that confidential.”
Yet even if online retailers do develop and maintain strict guidelines about the kinds of information they collect from Web shoppers, that effort still isn’t enough for many privacy rights groups, which are taking take a wait-and-see attitude toward personalization software. Privacy and the security of credit card transactions over the Internet are still the two biggest reasons keeping more consumers from shopping online.
“Consumers are concerned about personalization because a lot of it seems to be done covertly, which means people don’t know how information is being acquired or used,” says Marc Rotenberg, executive director of the Electronic Privacy Information Center in Washington, D.C. “That’s typically the basis for privacy problems.”
Still, many people are deciding that the convenience of shopping on the Web and the advantages of more customized attention when they do outweigh the disadvantages of divulging personal information-knowingly or not. When retailers assemble customer profiles about people who visit their sites, they use them in several different approaches, Brightware’s Williams explains.
These programs include:
Frequent-customer programs. This approach uses customer profiles to tailor Web pages to suit individual visitors. “A Web developer can put some logic in an HTML template file so that when a user comes to the site and has profile indicators set in their profile, the logic chooses which content to display,” says Cliff Allen, president of GuestTrack Inc., a Los Angeles-based software development company that specializes in Web site personalization. “The logic says, if the customer is a golfer, then show this additional paragraph of material. You can do a lot with that.”
Collaborative filtering. This technology considers what someone is looking at, searches the database for profiles of other customers who have bought that item, identifies what else those people purchased, and then uses that information to make suggestions to the original customer.
For example, if a visitor to toysmart.com is purchasing a Brio train set, the Web site’s Toy Detective may suggest that the person also look at the wooden bridge that many people who bought the train set have also purchased.
Automated sales advice and customer assistance. This application follows rules-based logic to help consumers find what they are looking for on a Web site and educate them about the products. “We are the ‘Hi, can I help you folks?’ that you would find at a traditional store,” says Stephen Tomlin, chief executive officer, PersonaLogic Inc., a San Diego-based personalization company acquired last November by America Online Inc. PersonaLogic creates decision trees that bring people through the process of selecting which product such as a bicycle or cruise, for example, is right for them. And in doing so, it tells the consumer why the questions are relevant to the product, so they understand why certain items are recommended.
Brightware also implements this type of system on Web sites, offering its Advice Agent server and software to create an electronic sales advisory dialog. “The idea is to have the same kind of dialog you might have with an excellent salesperson who knows the products,” Williams says.
With so many different types of personalization technologies available, businesses must decide exactly what they want to accomplish and create a strong business plan before they buy. “This is not an off-the-shelf, slap-it-in kind of thing,” says Geoffrey Bock, senior consultant, Patricia Seybold Group, a Boston-based e-commerce consulting firm. “A company needs to think through what they are going to personalize and how to go about it.”
Going all the way
Personalization technology remains in its infancy, with work still to be done in the areas of data integration and the scalability of applications. And its full value will only be seen when retailers personalize the entire online experience, according to Forrester Research.
That’s a goal of many retailers. With N2K and CDnow merging their online platforms into one combined store under the CDnow banner, personalization will be key to as the stores expand together.
“We’d like to have the entire store be personalized,” explains Evan Schwartz, CDnow’s director of product management. “If you go to Pearl Jam’s discography page, we want to show you the Pearl Jam records in the order we think you’ll want to see them. So every single page in the store will be personalized just for you.”