A Fistful of Dollars
Many company executives would agree that the most important factor in the sales process is the close. But getting customers to actually make a purchase can be a difficult journey. This is especially true for companies that have similar product lines. Ask Ken Wirt, senior vice president of marketing and product management for Palm Inc. After discovering that many Web-based customers prematurely ended the buying process because they were confused when choosing between Palm's different handheld products, Palm began looking for ways to close the loop.
"We heard from many customers that they would come to our Web site, wanted to buy a Palm handheld, but didn't know which one to get," Wirt says. "Because of this, many customers would not buy a device at all."
To rectify this problem Palm turned to San Mateo, CA--based Active Decisions, makers of what the company calls guided selling lifecycle solutions. Basically, these are automated sales applications that virtually walk customers through the buying process. The Web-based applications process a customer's most important buying criteria, and suggest which inventoried products most closely fit their needs.
In March Palm launched Active Sales Assistant, created by Active Decisions, to help customers compare and decide between Palm handhelds. Customers start by choosing which individual feature or combination of features (price, screen type, memory, etc.) is most important. Once features are selected, the application will further drill down and allow customers to select more detailed options. For example, if a customer selects "screen type," the Active Decisions application shows the choices (e.g., color or monochrome) that are available. After selecting the preferred choice, customers are able to rate the choice using one of the following attributes: somewhat important, very important, or extremely important. Once all relevant choices and attributes are selected, the Active Sales Assistant returns a list of suggestions that closest match the customer's query.
During the initial implementation of Active Sales Assistant, Palm performed what the company calls a one-on-one usability study. Upon visiting the Palm Web site, half of the site's customers were presented with links to Active Decision application and the other half to the traditional commerce process. Data was gathered to compare the number of sales and the average sales price from both groups. This information was used to determine if the Active Sales Assistant supported an increase in sold products and sales amounts. After performing the usability test for more than a month, Palm learned that customers using the Active Decision application had fewer drop-offs and the ones who did buy a device usually spent more than they initially planned, Wirt says.
The Active Decision application is based on several personalization and adaptive recommendation technologies including multiaxial preference matching and fuzzy recommendation. Multiaxial preference matching evaluates a user's total preference system, weighted by their intended and derived product preferences. Using this technology the application mathematically models buying needs, which are used to score products and show customers the best available choices. Fuzzy recommendation uses algorithms that take into account the minimum and maximum price a user is willing to pay and sets a priority and fuzzy nature on the price (lower the price, the better for the user). For example, if the customer's maximum price is $350, a $380 product may still be shown if the product is otherwise a good fit.
"After implementing Active Decisions, we saw a 20 percent increase in revenue," Wirt says. "This was from the combination of fewer people dropping out of the sales process and a higher average of purchase-per-customer."
Reduce customer confusion to build online sales
A 20 percent increase in revenue generated from it Web site
Positive customer feedback regarding the online buying experience
Fewer "drop-offs" from potential customers
Increases in final sale dollar amount