Tap in to the power of suggestion to bump sales

Creating a personalized online shopping experience is key to operating a successful Web shop and can be done in a variety of ways, but one route that's becoming more popular — and more profitable — is the "suggest a product" feature. There are now companies dedicated exclusively to this type of up-sell that claim the technology has improved from the days when you bought a gift for a toddler and then received recommendations for the latest Elmo doll every time you went back to the same site.

Kevin Stecko, president of 80sTees.com, says he isn't sure exactly how the technology works, and it can be complicated, but he does know that since implementing CleverSet last summer he's seen results.

"Set up involved putting some tags on the site as well as giving them access to our product catalogue. It helped us automate product recommendations without requiring us to build a system. Our order values as well as items per order have increased since using CleverSet," said Stecko, adding that the cost is based on the number of impressions, and "we have negotiated a price that I am not free to share."

Apparently, Stecko is not alone. A CleverSet survey found that online retailers increased revenue-per-visitor by an average of 13 percent, and in some cases up to 30 percent, this past holiday season by implementing relevant, personalized product recommendations on their sites. (The survey was conducted in late December among a group of CleverSet online retail customers.)

Easy Implementation, Increased Revenue

By offering customers personalized product recommendations based on analysis of their predicted future behavior, retailers such as aWineStore.com, 80sTees.com and Overstock Auctions were able to tempt shoppers to purchase items they would not have seen otherwise. According to the company, 63 percent of its surveyed clients were up and running with CleverSet in less than three days, with 25 percent able to implement CleverSet in just a few hours.CleverSet CEO Todd Humphrey says these results are typical of his company's "Smart Recommendations Technology."

"CleverSet doesn't make irrelevant product recommendations based on something a customer bought a year ago," said Humphrey. "Our technology captures and looks at the combination of customer behavior, product catalog information, Web analytics, demographic information and numerous other inputs to deliver timely, personalized and relevant recommendations to shoppers-presenting them with products they truly want to buy, every time they visit."

Developed by Dr. Bruce D'Ambrosio, an expert in statistical relational learning and professor emeritus of Oregon State, CleverSet uses the same technology deployed by the U.S. Department of Defense for intelligence operations to identify patterns and recommend products based on customer behavior during a particular shopping session.

Using statistical relational modeling, CleverSet identifies various "online clues" left by shoppers to provide instant product reviews based on predicted future behavior. The standard model looks at past behavior and uses rules-based technology, turning out mostly irrelevant recom­men­dations, according to the company.

"CleverSet mines myriad data sources and analyzes past, present and future customer behavior to present relevant recommendations tailored to each individual shopping session," said D'Ambrosio, CleverSet CTO and founder. "Our dynamic model analyzes not only customer behavior — what products they looked at, how they came to the site, how much time they spent looking at a certain product — but also mines Web structures and product information to find the right recommendations for the right customers at the right time."

Attribute-Based Model

Another company, ChoiceStream, also claims it has improved on the old techniques, namely item- or user-based collaborative filtering, used for suggesting products to online shoppers. ChoiceStream uses its own proprietary technology, called Attributized Bayesian Choice Modeling (ABCM).

"ChoiceStream's ABCM is based on the principle that in order to provide truly accurate, useful recom­men­dations, a personalization system must understand not just what users like, but why they like it," according to the company Web site. "By using techniques to classify content and products in terms of attributes people care about — attempting to represent content using the same characteristics that consumers consider when evaluating it — ChoiceStream's ABCM-based solution matches each individual's needs and interests with the content they are most likely to enjoy."

At its most basic level, ChoiceStream's ABCM-based perso­nali­zation:
  • Automatically classifies content according to the explicit and implicit attributes deemed the most powerful predictors of user preferences.
  • Creates an accurate profile of each user's preferences for those attributes.
  • Matches users with content based on their preferences.

When Profiling is a Good Thing
Another company specializing in this arena is Touch Clarity. This company operates on the premise that shoppers often visit a site as many as five times before making a purchase. From the first visit, Touch Clarity Targeting starts building a profile of each visitor, and recognizes him or her using an anonymous first-party cookie. Fully automated, the software learns in real-time what content will best achieve a given goal for each visitor and triggers the serving of that content.

Profiles are then used by the software to learn which content is the most relevant to present to each individual. Each time a visitor returns, any new data is combined with retrieved historical information to ensure that the most updated and most complete view of the visitor is used as the basis for content targeting decisions, according to the company.

Get Down With Upselling

Whether one of these products will work for you depends, of course, on the depth and range of products you sell. However, given the developments in this part of the industry, now might be the time to brush up on the potenital of up-sell technology.

Michelle Megna is managing editor of ECommerce-Guide.com.

By Michelle Megna

Posted in Blog on