What to Expect
Conventional wisdom says that knowledge is power. An undifferentiated mass of information is not knowledge. Intelligence results from a company's using its intellectual resources and capabilities to bring focus, clarity, and meaning to the large volumes of information and data made available by today's technology. However, attempting to harvest knowledge without having the necessary ana- lytical tools can render a company powerless. As companies adopt responsive, event- driven e- business models, they have to invest in new knowledge frameworks to help them respond to chang- ing market conditions and customer needs.
The challenge is how to transform the incredible amount of valuable data locked away in a compa- ny's applications, storage platforms, and databases into new revenue opportunities. Converting data into knowledge is the job of applications known as business intelligence (BI). BI is an emerging group of applications designed to organize and to structure a business's transaction data so that it can be analyzed in ways beneficial to company decision support and operations.
In this chapter, we examine BI's capability to tailor information content, format, and interaction to the needs of individual users. The BI trend provides customized and personalized information to customers and other end users through a variety of channels: e-mail, pagers, faxes, and Web pages.
We review the basic elements on which BI is built—personalization, analysis and segmentation, reporting, what-if analysis—that form the foundation of BI applications. We also examine several areas in which BI is being used, including employee benefits management, customer management, and information retrieval. The chapter concludes with an easy-to-follow manager's guide for estab- lishing a BI framework in your firm.
The main objective in war, as in life, is to deduce what you do not know from what you do know.
—The Duke of Wellington
In business, knowledge is neither a product nor a capability. Rather, knowledge is a critical frame- work of a fully evolved information economy. Let's consider two examples.
A multibillion-dollar retailer of electronics with more than 5,000 stores nationwide delivers an online weekly sales report to managers, who use this report to identify "hot spots"—locations where prod- ucts are selling at a faster rate than in the rest of the country. By identifying these hot spots, the retailer can inform its manufacturing partners what products are in demand in which regions, ena- bling them to manage inventory levels in response to real-time sales events.
An insurance claim software provider offers more than 200 auto insurance companies the ability to access and to analyze insurance claim data via the Web. Its consumer database alone includes profiles of more than 1 million consumers. Auto insurance companies access and analyze nation- wide insurance claim data, including repair-cycle times and the amounts paid for vehicle parts, and compare their claim-resolution performance against industry averages and historical trends.
245
These companies and many more like them are running "about the business" applications. For example, Wal-Mart, which operates more than 2,400 stores, supercenters, and 450 SAM's Clubs in the United States and more than 725 stores internationally, has developed a 100-terabyte data warehouse that monitors and captures each transaction in each store. The company's goal is better inventory management and improved collaboration with suppliers, which in turn enables Wal-Mart to merchandise each store on an individual basis and to provide superior local shopper satisfaction.
Wal-Mart's software, Retail Link, gives suppliers access to sales, inventory, and pricing information at Wal-Mart stores and SAM's Clubs. More than 7,000 suppliers access the data warehouse via Wal-Mart's Retail Link, enabling suppliers to know exactly what is selling where, plan their production accordingly, and keep inventories under control. With sales and in-stock information transmitted between Wal-Mart and their suppliers over the Internet, buyers and suppliers have a single source of information, thereby saving a significant amount of time over more traditional systems.
The Wal-Mart data warehouse, powered by NCR's Teradata servers, runs more than 30 business applications, supports more than 18,000 users, and handles some 120,000 complex queries a week.
[1] As Wal-Mart captures shoppers' transactions, the data warehouse receives 8.4 million updates every minute during peak times—detailed data on each item purchased. Wal-Mart's sophisticated knowledge management strategy is based on a simple premise: Success comes from anticipating customer needs—even before they do.
Whereas the first generation of e-business applications focused on buying and selling goods via the Web, second-generation e-business applications focus on organizations' gaining insight from the data collected with each transaction. These applications analyze data more effectively to develop customer loyalty and to enhance profitability, analyzing a business's customer interactions and helping optimize its customer relationships. Second-generation e-business applications aid in both interpreting what has happened in past transactions and using this knowledge to support decisions about which direction the company should be headed.
Evolution of Knowledge Management (KM) Applications
You've got operational data, transactional data, and a boatload of e-commerce data. Today's com- panies are looking for application solutions to help make sense of the information gathered. Com- panies want to know
• How to make effective use of raw data
• How to convert raw data into revenue
For example, FedEx reengineered its database marketing process from marketing and campaign planning to customer segmentation, evaluation, and refinement. A knowledge application has been instrumental in FedEx's efforts to automate its database marketing process. FedEx reports a time reduction in direct-marketing campaign cycles and a major improvement in "prospecting" cam- paigns.[2] Companies are demanding more than access to data. They want processed and refined information that will help reach effective tactical decisions.
At Best Buy, it used to take months to profile customers and to reasonably predict their purchasing behavior. Now, the company's new Web-based applications take less than half that time. The Web browser has become the de facto interface for the "information at your fingertips" paradigm. The browser's speed enables the retailer to direct services to its most valued customers much more aggressively in a highly competitive marketplace.
The foundation of any KM framework is information sorting, extraction, packaging, and dissemina- tion. Retailers, manufacturers, and financial institutions have spent millions of dollars to build data warehouses containing masses of information about their customers and their transactions. At headquarters, managers use query engines and reporting tools to extract useful information from data warehouses. The reactive, data centric world of today is gradually changing into the proactive, query- driven knowledge world of tomorrow. Figure 11.1 shows how KM has evolved in the past decade. This evolution has occurred in five waves, the last two of which are ongoing.
Figure 11.1. Evolution of Knowledge Management Applications
Wave 1: Group Memory Systems
Most companies will tell you that their two greatest assets are employees and the knowledge they possess. In its broadest sense, group memory is the sharing of information throughout the company.
Group memory systems included discussion boards or bulletin systems, such as Lotus Notes and corporate intranets. This technology provided a company's employees with instant access to data and reporting information that had previously taken days or weeks to obtain.
For example, Intraspect, a knowledge management start-up company, used its applications to in- tegrate collaboration, organization, searches, and subscription in a single location called a group memory (GM). Contribution and subscription to the group memory is easy because of GM's tight integration with the company's e-mail and desktop applications. Search tools help locate and reuse information in the group memory, the company's Web servers, and legacy databases.
Lotus Notes and Intraspect applications form the core of group memory systems. The main goal of GM is to enable companies to use their own data to determine best practices, to retain the tacit knowledge and experience of individuals, and to classify employee expertise. GM also makes it easier for corporations to react more quickly and decisively to problems, as well as to competitors.
The methods and combinations of products needed to implement a GM application, however, vary widely.
Although GM was the buzzword of the 1990s, it failed to live up to its promise, being long on cost and short on results. Consequently, companies have been reluctant to adopt the technology, for the following reasons.
Business Intelligence: The Next Generation of Knowledge Management 247
• Few can define it. —Vendors of document management systems, data warehousing applica- tions, and push technology all claim to provide GM tools, but some consulting firms help clients indiscriminately develop GM strategies for virtually any process without consideration for the enterprise-wide scope that is needed. Firms are faced with a deluge of contradictory and con- fusing efforts.
• Software vendors are distancing themselves from GM as a product. —Riding the corporate fascination with the GM wave is no more. GM was to the late 1990s what reengineering was to the early part of the decade, a fad spawned by consultants and vendors to generate demand for their products and services.
• Costly group memory efforts aren't delivering expected returns on investment. —The efforts of one company have produced a number of knowledge databases using Lotus Notes. The most widely used is the "gossip and rumors" database. In fact, GM is referred to as the "knowledge scam," owing to the miniscule ROI that firms receive from their GM efforts. At most companies, GM-like efforts have evolved into corporate intranets.
Unfortunately, GM's hype and failure have resulted in businesses' throwing the baby out with the bath water. However, new ideas and approaches, once suppressed, have a way of reappearing in new forms. Does your company have a group memory effort? What have results been so far?
Wave 2: Corporate Intranets and Decision Support Portals
Using corporate intranets and decision support portals, companies seek to create more complete and uniform linkage of the data resources scattered throughout the organization. The technology enabling this development is the corporate intranet. Moving from departmental solutions, in which data and reports are developed for small, specialized communities of users, to a corporate intranet opens up a company's data resources to a broader base of users by using the browser as a standard interface.
However, dumping reams of information on employees' desktops isn't effective. Key information should be disbursed just in time when the user needs it. For example, a large software company improved its sales-close ratio when it tracked the sales force's progress in the sales cycle and then distributed competitor intelligence or industry-specific information only at relevant points in the sales cycle.
Clearly, corporate intranets alone do not create knowledge, much less manage it. For knowledge to be created, data aggregation must be complemented by data analysis. One trend is to perform data analysis by using decision support portals. With databases growing larger every day and the time available for in-depth business analysis shrinking, automating the predictable components of a decision maker's routine, where possible, makes a lot of sense. Managers also use decision support portals to conduct data mining, the process of analyzing large quantities of data to discover relationships and patterns to support better decision making.
Decision support portals built on corporate intranets are a prerequisite for creating a responsive business model. For example, Home Depot, the do-it-yourself building supplies retailing giant, is pushing knowledge to the staff level. Home Depot carries a diverse inventory in its more than 500 stores. Most of its stores have installed Home Depot's Mobile Ordering platform, a radio-frequency- transmitted data warehouse link. The system uses cart devices to let floor clerks and department managers access data analyses of the store's past and present inventory in order to make mer- chandise ordering decisions while standing in front of the merchandise itself.[3]
As this example illustrates, decision support applications streamline the process of turning decision into action. The objective is to help users address any critical issues facing their businesses by answering highly focused, industry- specific management questions.
• Retail: —What products or groups of products should our company sell? Where? At what price?
How much shelf space should be allocated for specific products? How much promotion should each product receive? Which products sell well together? How much inventory should be car- ried? What was the in-stock position and stock-to-sales ratio of the ten most profitable and ten least profitable items?
• Banking and finance: —What are the 100 most profitable customers by branch, and how are they contributing to income? What portion of the contribution comes from fees? Interest income?
Overdraft charges? Whom should I target for direct-marketing efforts? What is the proper pricing strategy for a given set of financial products? Which customer groups are credit risks? How much fraudulent activity is occurring?
• Telecommunications: —Of the customers who have switched carriers in the past month, what are their average call volumes and dollars spent since they signed up with my company? What are the same metrics for the 3 months before they quit?
• Healthcare: —What is the range of outcomes for a given treatment? How frequently is this treat- ment prescribed? Which drugs, hospitals, doctors, and health plans are most effective? Which patient groups are most at risk? How efficient and effective is a given technique for treating a specific illness?
The promise of intranet-based decision support applications is to offer decision makers the oppor- tunity to ask and to answer mission-critical questions about their businesses, using transactional data assets that have been captured, but not exploited, to their fullest extent.
Wave 3: Extranets and Interenterprise Portals
As companies begin to implement supply chain strategies, they will move select parts of the internal corporate information infrastructure outside the firewall so that suppliers and trading partners can access them. These extranets are driven by fast information access, customized data, and respon- siveness.
Extranet-based interenterprise linkages create new requirements: the ability to manage huge data volumes, data breadth coverage, cross-platform support, response-time speed, and a broad range of interface choices. For example, every day, DaimlerChrysler's massive Mopar parts operation ships about 220,000 items from 3,000 suppliers to its 15,000 dealers. A logistical nightmare, the operation is a monitoring and forecasting balancing act to ensure that the parts pipeline stays just full enough but not too full. Dozens of planners and forecasters manage atop this network, making decisions about how many engines to ship to Phoenix or correcting an undersupply of brakeshoes in Boston. If the pipeline is constricted, customers go without the parts they need, meaning that costs rise as parts are rushed to them. If the pipeline is full, DaimlerChrysler ends up paying for needless inventory. Viewing these transactions in real time using a Web browser provides invaluable data to DaimlerChrysler's management team to make pipeline adjustments as the parts are distrib- uted from the group's 15 distribution centers.[4]
Extranet-based applications encourage trading partners to improve profits by managing inventories in the supply chain. In order to obtain information visibility and reliability, a company's partners may be willing to offer more favorable terms, invest more in comarketing, make available increased levels of supplies, provide more shelf space, or pay higher prices. The extranet-based KM strategy goal is to give preferential treatment to one another in exchange for detailed ordering and inventory information that provides greater process reliability and visibility up and down the supply chains.
For example, Lexmark, which manufactures and distributes laser and inkjet printers for the office and home markets, is using extranet-based solutions from Microstrategy to help customers manage their inventories. Through Lexmark's data warehouse and an inventory management application called the Retail Management System (RMS), Lexmark aims to help dozens of large retailers man- Business Intelligence: The Next Generation of Knowledge Management 249
age their inventories of printers. The company is using an approach called vendor-managed inven- tory (VMI), which is used by the retail packaged-goods industry to track inventory. VMI helps re- plenish inventory before it's depleted. In contrast, stores without VMI often don't order inventory until after it runs out, because they're too busy to act in advance.[5]
Potential users of Lexmark's RMS application include a firm's vendors, distributors, partners, out- sourcers, resellers, and financing sources. Initially, the RMS project was used to provide customer inventory information to about 35 field sales people but later was expanded to provide management reports to about 75 top executives and line managers at headquarters. The number of potential supply chain DSS users can range from hundreds to tens of thousands.
In addition to boosting sales, RMS gives Lexmark a better idea of where its customers are and the best locations for its products. The data warehouse replaces a system in which inventory figures were compiled by paper. Using the data warehouse, sales and inventory information that once took 4 or 5 days to turn around is now compiled in half a day.
Lexmark's RMS application is an example of an interenterprise knowledge portal. As such, it pro- vides access to valuable retail sales information that can be used to design new products, refine marketing campaigns, develop optimal pricing schemes, rationally allocate inventory, and proac- tively schedule factory production.
Wave 4: e-Commerce and Click-Stream Analysis
Successful online interactions with customers don't happen by chance. The rise of e-commerce has helped new forms of KM management to emerge, such as user click-stream analysis, e-mail man- agement, and multichannel knowledge portals.
User Click-Stream Analysis
Click-stream analysis provides the electronic footprints that show where people go on the Web, what they do or buy, and when they return. What marketers need is the ability to always be aware of every customer activity and purchase, as well as to be able to analyze and to understand their buying preferences and to anticipate their changing expectations.
Collecting and analyzing all the mouse clicks of a company's prospects and customers is a night- mare for companies ill-prepared to handle this amount of information. How can businesses prepare?
User click-stream data is accumulating so quickly at some Web sites that it's testing the limits of conventional approaches to database management. As a result, a new generation of much larger, faster-growing databases in the 5-, 10-, or even 15-terabyte range is being developed, forcing man- agers to invent new roadmaps on database design, storage, backup, and archiving.
e-Mail Management
Another rapidly growing KM area is intelligent e-mail management. Companies providing these applications include eGain, Kana Communications, and Siebel. Until recently, a company's rela- tionships with its customers and partners were based on in-person, telephone, or written interac- tions. In order to communicate effectively, companies invested substantial resources in call centers.
Typical call centers were customer service oriented, often using costly technology, and were not scalable. Also, outbound direct-marketing calls are expensive and only minimally effective in their conversion and response rates. With the advent of new channels and the proliferation of e-mail, business communications with customers and partners has fundamentally changed.