From December 2007

By Janet Groeber

Bricks-and-mortar stores have certain inherent advantages over their online brethren – salespeople to direct customers and answer questions, mounds of merchandise to touch and lively environments in which to do both.

Still, it’s tough to beat browsing online at midnight in your skivvies. But e-commerce consultants like Forrester and Jupiter think “Internet fatigue” will cause a predicted slowdown in online sales. Cupertino, Calif.-based Baynote, an online product and content guidance provider, hopes to rescue online retailers with its new technology that allows like-minded customers to “hang out” online.

Ninety-five percent of online customers “will bail if they don’t find what they’re looking for within just three clicks,” says Baynote founder and CEO Jack Jia. To help convert online browsers to buyers, Baynote developed the Community-Guided e-Commerce service, which tracks 20 characteristics in order to form groups of like-minded customers. Baynote’s “content guidance” matches customers to products that they are actually interested in at a specific point in time.

Rather than taking a behavioral approach – “You bought this last time, you should consider this” – Baynote uses contextual targeting powered by the ‘Wisdom of Crowds” – observing a site’s invisible crowds or silent groups of like-minded peers whose intent is determined by their implicit behavior instead of their explicit feedback (forums, posts or focus groups) or past shopping experiences.

“We really don’t care who you are or about your past [purchasing] history,” Jia explains. “We are against profiling and personalization, and against tracking your historical behavior.” Why? “Because all those things are going to match with other people at some random point,” he says. “We care about the holistic view of that visitor community.”

Baynote’s platform “is based on social science,” says Jia, “the idea that humans are not that unique after all. Our physical brain structure is quite similar even though we may be different culturally or from a gender perspective. We’ve found that the like-minded peer will behave 95 percent similarly.”

What’s in a click?
Still, Baynote also recognizes that an individual customer can be a father, an artist and an avid cyclist, so tracking all of these profiles makes it very difficult to predict what that customer might want on a given day. Baynote’s social science research found that the “click” doesn’t reveal anything: high click-through rates are most often the function of the area of the site. If you put a link on the front page, for example, you’ll get more clicks, Jia says.

Technology primarily built around attaching meta tags to content produces too many results, Jia says. What Baynote has done is attempt “to mimic the human brain. We’re tracking memory patterns among visitors and then finding like-minded peers who have similar memory patterns.” Even without their active participation “we can then link these customers in relevant ways.”

Baynote’s technology strives to intuitively lead visitors to the products or services that best satisfy their needs with its Affinity Engine. Visitors can then click boxes labeled, for example, Most Popular, Shoppers’ Picks or People also Considered.

Customers start out “generic,” Jia says. If they’re shopping for a refrigerator, they might prefer a side-by-side design. When they click on such a model, “they will immediately see additional recommendations … four or five other units that the customer can choose to click [on] for more information.” These are the models that other customers clicked on and are now their recommendations.

From there, customers will begin to diverge. “Some want a side-by-side with a freezer in the bottom, so that leads them to … the recommendations of those customers looking for similar models.” New groups form as they continue to search for additional features.

The scenarios Jia cites are real ones from Baynote’s beta testing with US Appliance, which went live with Baynote’s technology in March and already has experienced a double-digit increase in conversions, according to site manager Joe Nashif.

Revenue spike
US Appliance is a division of Pontiac, Mich.-based ABC Warehouse, which operates about 40 stores in Michigan, Ohio and Indiana selling everything from ice makers and wine chillers to built-in warming drawers and other high-end luxury home appliances. During the testing phase, half of US Appliance visitors saw an area called Shoppers’ Picks containing four or five product thumbnails.

“We immediately saw a huge revenue spike because of the recommendations,” Nashif says.

Recommendations “tend to lift awareness, but they also tend to make customers aware of the better, more expensive products and higher-margin items,” Jia says. Baynote’s customers (which include Glam.com, eBay and Motorola) have reported revenue boosts of anywhere from 18 to 107 percent, Jia says, and “if you get a 30 percent revenue lift, you’ll likely get a 60 percent profit lift.”

Improved navigation also helps explain why Baynote’s solution is working well for US Appliance. “We didn’t have any merchandising activity under way because we had never undertaken the effort it would require,” says Nashif, who points to the “heavy review” of its sales and web statistics and the resulting maintenance. “It was something I always knew we were lacking and could benefit from, but I also recognized it would be a big step to fully support it.”

Now, he says, “it’s all done automatically and we didn’t have to get involved or hire another person to handle it.”