Of all the "active" channels to customers, the Web is growing the fastest. Approximately 1.4 billion people, 21% of the world's population, are now online, according to www.internetworldstats.com, a site that tracks Internet usage and population trends. Web use ranges from 73% of people in North America to 5.3% in Africa; about half of Europe is connected. Transaction volumes are rising on every channel; however, the mix of use is changing, with the Web channel growing the fastest.

In retail banking, for example, the number of transactions across all channels will increase 9.9% through 2010, according to the Tower Group's 2007 Web seminar "Going from 'In Line' to Online: Transaction Migration in the U.S." By 2010, 40% of transactions will be on the Internet, with 23% to the call center and 19% each at a branch or ATM.

Internet use is not limited to PCs. "Online" increasingly means mobile phone connectivity via Web access protocols (WAPs), which accommodate phone screen sizes. The continued growth in mobile phone sales—now at 3 billion—along with WAP capabilities means continued unprecedented growth in the number of Web transactions. The latest technologies let users download custom coupons and boarding tickets, as well as transfer money—mobile cash—onto their phones.

Smarter Web systems
The Teradata depiction (see figure, below) of the decision-making maturity continuum can be adapted to highlight opportunities to use enterprise information and insights to make Web systems smarter.

Stages 1-3 show how the enterprise data warehouse (EDW) is used to develop strategic insights. Stage 1 might include reports on how many people are using the Web, when, and how much product they buy. Another report might show the number of customers who are "lookers" versus "bookers" and how many "lookers" are buying at a store instead of online. Other examples are scoreboarding reports of internal operations, such as which packages are at risk of missing guaranteed delivery deadlines or how many widgets are being produced per hour.

Stages 2 and 3 focus on using the data warehouse with business intelligence (BI) and Web analytics packages to do ad hoc and predictive analysis of customer activities, including Web browsing. By capturing click-stream sequences in the EDW, an organization can see the history of each customer's interactions. The tools can be used to spot where dwell times are highest—perhaps because individuals are reading about products of interest—or analyze where they might be getting confused and bailing out. These insights might be used to drive Web site redesigns.

In Stage 3, organizations use predictive tools like SAS and KXEN to build models of customer segments and their cross-channel activities, to provide answers to various questions: What is the next best product to offer based on up-to-the-moment store purchases when the customer returns to the Web site? Does this customer have a high propensity to churn, based on clues from Web-browsing behavior, like looking at new call plans? By conducting pricing experiments on the Web, an analyst might build price elasticity insights—at what price this customer segment will buy online.

Stages 4 and 5 concentrate on the use of these strategic insights to improve front-line, operational systems. Such insights can be applied to improve real-time, tailored variations of the Web for each customer:

Customized Web screens. Provide recommendations on custom advertising in one portion of a home page for the next best travel deal, an invitation to a free wealth management consultation on a financial Web site or a suggestion pop-up to switch to lower-priced generic drugs on a healthcare site.

Customized sequence of Web screens. Add "decision points" based on insights to enrich the customer experience with greater depth and relevance. For example, instead of a telco home page providing lists of all available products, the Web rendering engine might use insights to trim it to service pages for the products a customer already has, coupled with sales pages for only the products that person is most likely to buy.

enlarge

In the Web world, most companies have built standardized portals, often three—one each for customers, partners and internal employees. The focus is on widening simple information access and using the insights to drive context-aware sequences of Web pages that anticipate what users are trying to do based on roles or personas. The system uses current data about the customer (Where is the person, perhaps based on GPS feeds or ATM location?) along with historical context (Have we seen this situation before?) and system state/optimization rules (What can we do to optimize this customer's experience?).

An example from the airline industry is a passenger missing a connection. The airline knows the passenger is on the first leg of a trip and that a flight delay will prevent a planned connection. It knows whether the individual is en route on the plane or waiting to board. It also knows the rest of the system's status, so it can construct alternate connection plans. Additionally, the airline is aware of the value of this particular customer and the likelihood that the person will defect because of other recent incidents like lost bags or canceled flights. When the plane lands and the customer accesses the Web via a mobile phone, an "active" airline would make the first Web screen be the contextually most relevant one. In this case, "Your revised travel options" might appear first on the WAP page.

Making it happen
As Web use grows, it's more important than ever to drive projects that connect data-derived insights to your online channels. But what will that take?

An Active Enterprise Intelligence approach works only if you can forge the right triad of application architects, operational system owners and database administrators (DBAs), coupled with a change agent who sees the opportunities generated by your data warehouse investment. A DBA can start by educating the owners of the Web, then participating with corporate architects in building a company-wide, customer-centric vision of how to use and reuse information.

With business groups, you might need to foster a "customer dialogues" project, working with your customer relationship management team leaders and BI users, and possibly a governance committee. This ensures that customer insights are captured in one place, documented and systematically reused across various channels and departments for competitive advantage.

Finally, "making it happen" requires good project management skills, because cross-organizational projects are difficult to direct. But with focus and persistence, you can activate the Web channel with deeper insights, resulting in wider use of your database investment, as well as fostering better, more consistent customer experiences. T

ABN AMRO ABN AMRO, an international bank and insurance company with 19.8 billion euros in revenues in 2007, uses its data warehouse for strategic insights and drives those insights into Web operational systems. A typical application involves customized Web advertising. At any time, the marketing group has 50 advertisements ready to display on the home page when a customer or prospect comes to the site. The question is: Which offer should be made to each customer? Using a call-out to its Teradata system from the Web engine for guidance, ABN AMRO displays the best-suited ad within two seconds. With 175,000 Web sessions per day, the result is 63 million personalized offers per year. Does it pay off? A typical non-targeted bank ad achieves a 0.2%, or 1 in 500, click-through rate. But by applying better insights, ABN AMRO reports click-through rates of 1.1% to 5.5%, resulting in purchases of additional bank products. And a bonus from their disciplined approach to next best offers is that the same insights are reused on call center agent screens. —D.S.

Norfolk Southern A prime example of operational use of the Web comes from Norfolk Southern Corp. (NS), a $9.4 billion-per-year railroad company. More than three years ago, NS was looking into helping the power users help themselves faster. Wider sets of users—internal front-line groups as well as more technologically savvy business partners—needed access to up-to-date and historical information about shipments by themselves without waiting for NS to help. The approach taken by Blair Hanna, manager of e-commerce, and Mark Wittl, manager of customer applications, was to build a Report Wizard, with access to more than 125 fields of information so that users could modify existing reports or easily build their own. The philosophy was "Serve yourself," anytime, even at 3 o'clock in the morning! It worked. Expanding access to information led to wider use of the Teradata system, with more than 12,000 customers now using the accessNS Web portal. In addition to the 1,900 standard reports that NS Support provides to more than 30,000 users each week, users themselves created 9,500 variations of the reports and 4,400 new reports. They can set run schedules and delivery options for their reports. The system holds 8TB of data, with 4TB of user data. Various data elements are refreshed at different rates—by the minute, hour, day, week or month—depending on the business needs. "The users love it," Hanna says. "It takes most business users only five to 10 minutes to customize their reports, and sometimes new users never even need to contact us at all while creating their own reports." New Web reports, created by NS Support, are typically built in a half-day or day, depending on the complexity. These capabilities definitely improve the ease of doing business at NS. —D.S.

Dave Schrader is director of Strategy and Active Enterprise Intelligence Marketing for Teradata.

Teradata Magazine-September 2008