Technology helps catalogers deliver precise online marketing

Marketing online is cataloging in reverse. Instead of mailing to your housefile, you use it to lure prospects to your Web site.

Using high-speed automated databases, Web sites can make judgment calls about what products to offer to which consumer and can treat valued customers differently than prospects.

That’s because you have a mix of customers who know you and customers who don’t, so changing your Web site to suit each customer is just as important as versioning your catalog or knowing which products will appeal most to a certain consumer.

Tailoring the online offer to the shopper increases the chances of purchase. Nothing is more welcoming to online shoppers than having a Web site greet them by name and then show them the best and newest product in their favorite colors. Customers are often willing to supply retailers with information if it will result in a better shopping experience. But if you waste or misuse the information, customers grow reluctant to register.

By mining and modeling their databases, catalogers find they can attract online customers using targeted offers. Off-line, catalogers model to build profiles of similar customers. Online, they use the same data to assemble merchandise profiles.

Offering the right products, combined with the right incentives and discounts, can convert lookers into buyers or entice a former shopper to return.

Reelin’ Em In

Williams-Sonoma has been working with e-commerce marketing company Net Perceptions to create targeted, personalized e-mail offers for its online store.

In October, Williams-Sonoma will re-launch its site using e-commerce and Personalization Network products from Net Perceptions. For the multi-channel, specialty supplier of cooking utensils, reaching customers in an appropriate way with the right offer was a challenging goal to meet.

Shelley Nandkeolyar, Williams-Sonoma’s vice president of e-commerce, says the company felt it was important to know what customers were looking for and offer it in a precise and real-time manner. When considering its options for personalization, Nandkeolyar says Williams-Sonoma had to factor in its own retail channels and those of its other retail companies: Chambers, Gardener’s Eden, Hold Everything and Pottery Barn, which launched its first site in August.

Many of Williams-Sonoma’s customers shop through one or more of its retail stores and often via multiple sales channels, requiring customer behavior to be integrated from all channels for use online and in e-mail campaigns.

“As we think about long-term opportunities to build our customer-management capability, we look at things that give deep customer behavior and understanding so we send the right messages,” says Nandkeolyar about the company’s decision to use online personalization. “I think it reinforces our basic position as an authority in the cooking space. The basic [customer] expectation is that [we] know what the customer needs and that we provide them with good counsel.”

The Personalization Network allows Williams-Sonoma to combine past online and off-line purchasing behavior, real-time shopping patterns and product acceptance rates to generate an automated, instantaneous merchandise recommendation to the online shopper.

Tracking and using real-time information to affect online product sales has been challenging for catalogers. While they are familiar with analyzing how merchandise performs and making the necessary alterations in between drops, the pace at which the Internet throws out data has been dizzying. In addition, catalogers are competing with companies that offer lower price points, not necessarily a strength of catalogers that are used to working on tight profit margins.

Automated personalization programs can eliminate some of the work. For instance, the Personalization Network can identify a cataloger’s best customers based on recency and purchase amounts. It then takes the customer’s history and generates a list of common products purchased by best clients. The data can then be used to offer the right incentives and discounts to only quality customers, so that the cataloger is not “giving away” product to bargain hunters who won’t become multiple buyers.

Nandkeolyar sees personalization as a component of customer service needed for effective marketing. He says that while marketing to consumers through personalization does require using existing data, he doesn’t consider it database marketing because unlike the current uses of database marketing it is real time and one to one.

“Data is a very perishable commodity, and if you don’t use it strategically at every interaction to enrich the customer experience or influence the purchase then it really isn’t being used well,” says Nandkeolyar. “We really are taking a data point and leveraging it into a marketing interaction. It makes it so much more meaningful.”

How it Works

Upon entering the Williams-Sonoma site, the customer will be asked to register, cuing the customer’s file. If the customer is new, he is asked to complete a customer profile supplying information about cooking abilities, interests and needs. That information is compiled with existing data, if the customer is a catalog shopper.

Using either a rules-based system in which products are offered based on certain responses, or collaborative filtering, by which merchandise suggestions are determined from real-time browsing and other shoppers’ purchases and profiles, Williams-Sonoma is able to supply a fairly accurate suggestion as to what the consumer wants. Nandkeolyar says this offers the convenience of always being offered appropriate products in the right colors at the right price. But it also provides the consumer with a suggested offering that might not be obvious but is favored by like-minded shoppers. It is similar to cross-selling but on a larger scale. Essentially the customer gets a lot more of what she wants—fast.

Enticing Offers

Using Net Perceptions’ Intelligence Channel, Garden.com created a targeted e-mail campaign to recapture quality customers who hadn’t yet purchased in 2000.

In identifying the past purchasers, Garden.com discovered that most had purchased roses. An overall popular item for Garden.com is the Princess of Wales Rose. By further analyzing the data, the Internet retailer selected a variety of complementary products to the rose and created an incentive package. Offering 50 percent off the rose, but cross-selling it on the Web with the affinity products, Garden.com still made a profit and reactivated some quality customers.

Garden.com used three simple queries to unearth high-spending, multiple-purchase buyers from its file:

•Identify profitable customers.
•What do they buy?
•What sells with roses?
The e-mail offer resulted in a 33 percent click-through rate, with 41 percent of those customers making their first purchase in 2000.

On target: Williams-Sonoma’s HTML-enhanced e-mails are personally addressed, based on past purchases and are a function of affinity profiling. Customers choose the frequency and type of content they want in their e-mails. The message above reads:

“Easter is a time to enjoy a grand celebration with family and friends. At Williams-Sonoma, we help make the most of this springtime holiday with a festive assortment of cookware, bakeware, specialty foods and gifts that will delight adults and children alike. We invite you to visit us online at www.williams-sonoma.com.”