Tools for centralizing scattered data and using it to define and predict individual customer behavior are available in many forms, from standalone extract, transform and load (ETL) applications to proprietary embedded database plug-ins like Oracle Personalization. Among the newest of these tools, context servers offer the promise of objectifying widely disparate bits of customer data and restructuring it as a graphical guidance system for targeting product and service offers to that customer.
Context servers reach out to the furthest outposts of an enterprise regardless of hardware platform, operating system (including legacy Cobol systems) or software application to gather and analyze every bit and byte of information known about a customer and his or her previous interactions with the company. Think of a chess-playing application that collects information from every square on the virtual board, tries thousands of moves, and presents you with a palette of five or six optimum ones from which to choose.
"You're not talking about just pulling in the name and address from over here and the order history from over there and the call center history from a third place," says Forrester Research Analyst Eric Schmitt. "What you are doing is centralizing some of the data, of course, but more importantly you're also defining at a real high level what the customers look like and what their behaviors are."
Schmitt notes that the type of enterprises that would benefit most from this convergence of raw data, analysis and object-modeling technology are those which have multiple sales channels and deal with customers on more than one level--firms with online, phone, brick-and-mortar and direct-mail wholesale and end-user sales with a large range of products and services geared to different clientele, for example. Other target companies would be those using specialized technologies for each of their business units.
"It's not just strictly the number of channels, it's the touchpoints within those channels," he says. "Companies using one call center for sales and marketing, but some other call center package for service and support may also benefit from context server implementation."
Context Server Contenders
Several companies pioneering such open-platform context server solutions are YellowBrick Solutions, Miosoft and eLoyalty.
YellowBrick's Visitant is described by the company as a Customer Experience Management application that "assimilates call center, marketing automation, sales, Internet and wireless systems with back office applications, databases, and analytic tools" to present an object-modeled representation of a customer profile.
According to the company, Visitant incorporates "memory" of all previous customer interactions in each profile and automatically modifies profiles based on what it "learns" from each subsequent customer interaction. This "living profile" is immediately broadcast to every employee and executive involved with that customer or any policy or decision that may affect that customer.
Miosoft's Customer Context Server is an object-oriented application that tracks and integrates a wide range of customer financial, personal relationship and buying habit information across a variety of software and hardware platforms. Then it applies a set of analytical and company-specific rules to that information and presents the results as an automatically updated graphic model of that customer.
Global customer-retention consultant eLoyalty has also entered the context server arena with its Loyalty Suite 5.0. The Loyalty Suite contains six engines that, the company claims, do everything from transforming raw data into "critical customer information," to helping executives optimize business policies and rules impacting customer loyalty, to providing a Decision Engine that applies those rules across multiple channels and applications.
Although context servers may become an invaluable tool for accurately predicting customers' future behavior based on computer analysis of their past, Schmitt cautions that it is "still early days." None of the applications currently available have been fully field tested and proven, he says, and none have been developed to the point where "you are going to be able to plug them in and assume that they have connectors and work with your front and back office interfaces."
Like other data-mining applications, context servers are hostage to the quality of raw data existing throughout the enterprise. If the call center scripts do not pose the right questions and the company's Web survey isn't collecting relevant data, context server profiles will be about as accurate as the pundits in last November's election.
There are also potential invasion of privacy issues involving context servers. With Miosoft, for one, recommending the inclusion of such items as credit reports and the "dynamic relationship" of "people living together," it might be well to heed the Gartner Group's recent advice to gather such personal data now because "by 2001, consumers will start to boycott Web sites disrespectful of privacy."