Business Intelligence is one of the hot topics at the moment, promising customers ways to make better use of all that important information stored on databases
The enterprise Business Intelligence landscape
It is now commonplace knowledge that modern corporations must use information effectively to compete successfully. It is less widely recognised how dramatically information use is changing on an enterprise basis. Restructuring, deregulation and industry consolidation have increased the number of critical decisions to be made and reduced the time available to make them. Flatter organisations produce higher percentages of decision-makers throughout the enterprise. The amount of knowledge and data available on which to base decisions has grown exponentially.
The data warehouse paradigm has emerged as a flexible framework for enabling more people to make better decisions based on more data. At its best, a data warehouse sources data from multiple internal and external information systems, and makes it available via a range of options, from push-model, update systems that regularly, automatically and invisibly refresh reports on users' PCs, to fully interactive pull-model technologies that enable users to independently ask and answer their own questions, working at the speed of thought.
In the pre-data warehouse era, few individuals outside the IT organisation had the tools and the ability to easily access data to become active users of information for decision-making. Client/server systems, decision support databases and data warehouses have completely changed the level and distribution of active usage of information in many organisations.
In this New World, there are still a small number of IT and other professionals who are expert data users, and there remains a group of people who do not need anything more than simple reports. However, a broad "middle class" has emerged whose data use is critical to business success. The question for these people is not whether they will use a data warehouse, but how they will interact with it.
The IT challenge today is addressing support for decision-making in this new context. If the broad majority of the corporation's decision-makers are being supported by data warehouse solutions, how effectively is the solution dealing with them? How well are user differences recognised? How well are those different needs met? Most importantly, how well are the differences balanced with enterprise needs for IT standards, functional consistency and return on investment?
The enterprise Business Intelligence "supply chain"
Examination of numerous data warehouse solutions shows that information systems, and those interacting with them, collectively form a "supply chain" complete with raw materials (data), inventory and distribution systems (the data warehouse and corporate network), manufacturers or producers (IT) and consumers (end users).
In the supply chain model, information producers build and manage the data inventory and distribution system. Within this group, information technologists are the group who creates automated systems for delivering information and who also manufactures enterprise for information consumers in the form of applications, reports and analyses. Information technologists need full-function tools for all these activities so that they can continually predict and address the changing needs of everyone else in the information supply chain.
Another group of information producers is power users. They rely on information technologists to make structured data available, but thereafter they become information producers for their own purposes. These sophisticated users interact directly with warehouse data to create reports and analyses for themselves and selected information consumers. Power users have an insatiable hunger for data and a concomitant ability to work with it. They will master any tool or technology that might assist them in finding the answers they need. Like information technologists, power users require full-function tools that impel rather than impede their work.
The largest group of information consumers in most organisations is report viewers. Note that this group generally includes senior managers. Report viewers do not interact with warehouse data directly; instead they view information in a variety of ways in their day-to-day activities. Typically, they receive information through pre-built reports that have been developed by information producers. The primary concern of the report viewer is that their reports are timely, accurate and readily available.
The fastest-growing group of information consumers today is active analysts. These professionals want to answer "What about...?" questions that go beyond the bounds of pre-defined analyses and standard reports. As such, their warehouse access falls between that of a report viewer and a power user. They want to interact with the information on which reports are based, but don't necessarily need to access the data warehouse directly (nor would IT prefer to grant such access in most cases). Active analysts generally spend about 10-15 per cent of their time analysing information on an ad hoc basis; they need access tools that enable them to get in-depth answers whenever business questions arise, regardless of time or place.
Beyond client/server solutions
Some organisations have attempted to serve all information producers and consumers using a highest-common-denominator approach. Everyone gets a client/server connection to the data warehouse, and users determine their own level of interaction after that. Experience shows that this approach entails high and continuing investments in technology and maintenance relative to the impact on enterprise .
It is therefore with relief and excitement that many IT organisations have greeted the serendipitous arrival of the World Wide Web: a cost-effective way to provide anywhere, anytime access to information for the broad middle class of data users. Client/server access to the data warehouse can be limited to those who actually need it, and whose impact on enterprise actually warrants it. The Web and a corporate intranet can take care of the rest.
While web technology lowers the cost barrier to enterprise data access, it does not automatically meet the varied functional needs of the broad set of users described above.
The Web is clearly a part of the solution. The overall challenge remains, however: deploying effective access and analysis products for enterprise by embracing both the Web and intranets, as well as the client/server network, to ensure that all users are served effectively. The critical success factors for such access and analysis products are detailed in the following paragraphs.
Access and analysis products for end users must meet three criteria for integration into the enterprise. First, they must leverage existing hardware and software systems, especially the range of desktop operating systems found in the typical organisation. Access tools that do not connect with all desktop operating systems are like telephones that work only with certain telephone exchanges. Second, access and analysis products must take full advantage of other products important to, specifically including the data warehouse, its server platform and metadata, and middleware. Third, they must present a consistent interface, file structure and documentation across versions for all types of decision-makers. When a report viewer evolves into an active analyst, for example, she should find that nothing has changed except the level of her access to information and the tools available to her.
Similarly, access and analysis tools must meet three criteria for overall solution scalability. First, they must scale functionality along the entire information supply chain, so that individuals can change their level of data access as their jobs change or their level of sophistication evolves. Products that restrict user evolution tend to restrict the growth of overall enterprise. Second, price points should scale along the supply chain, to keep costs in line with overall impact. Organisations should be able to give everyone report viewer status at a low cost per seat, while investing in other levels of user access at the level their impact warrants. Third, access and analysis products must be able to scale to support large databases and large numbers of users. Growth in user sophistication leads to growth in the warehouse solution and the number of people who wish to benefit from it.
Low IT impact
Finally, access and analysis products must minimise their impact on the existing IT organisation and infrastructure. First, they must provide short "time to market" for IT and for users. Both installation and training should be simple and quick. Second, IT resources required to support the access and analysis products should grow more slowly than the number of users and the number of applications.
As the solution scales up, IT resources required should be proportionately reduced. Third, data warehouse access and analysis products must support a high degree of manageability, auditing capability and security. These qualities are absolute requirements for hardware and software investments elsewhere in IT; they should not be passed over here.
Recent studies also indicate that enterprise solutions can produce dramatic results. International Data Corp. (IDC) recently found that the three-year return on investment for successful data warehouse implementations averaged 401 per cent in the companies studied. With the right levels of enterprise integration, solution scalability and IT impact, enterprise solutions can do more than reduce operating costs. By improving enterprise decision-making, they can drive the profitability of the organisation as a whole.
( Brio Technology, Inc. 1997
Compiled by Paul Phillips