Business Intelligence Systems Defined
Business intelligence systems (BIS) are interactive computer-based structures and subsystems intended to help decision makers use communication technologies, data, documents, knowledge, and analytical models to identify and solve problems. The new generation of BIS offers the potential for significantly improving operational and strategic performance for organizations of various sizes and types.
During the 1990s, most large organizations engaged in enterprise data warehousing projects. The scope of these efforts ranged from combining multiple legacy systems to developing user interface tools for analysis and reporting. The data warehouse is the underlying structure that is used to generate a variety of reports and analyses. In the past, business intelligence amounted to a set of weekly or monthly reports that tended to be unconnected.
Two salient features of the new generation of BIS are integration and visualization. Typically, this information flow is presented to the manager via a graphics display called a Dashboard. A BIS Dashboard serves the same function as a car’s dashboard. Specifically, it reports key organizational performance data and options on a near real time and integrated basis. Some BIS industry pundits claim that Dashboards are simply “eye candy” for executive managers. This perspective suggests that these systems are merely a new fad being promoted by consultants and vendors. While these claims may have some merit, Dashboard based business intelligence systems do provide managers with access to powerful analytical systems and tools in a user friendly environment. Furthermore, they help support organization-wide analysis and integrated decision making.
Typically, BIS can be categorized into two major types: model-driven and data-driven. Model-driven systems tend to utilize analytical constructs such as forecasting, optimization algorithms, simulations, decision trees, and rules engines. Data-driven systems deal with data warehouses, databases, and online analytical processing (OLAP) technology. A data warehouse is a database that is constructed to support the decision making process across an organization. There may be several databases or data marts that make up the data warehouse. OLAP is increasingly utilized by managers to help process and evaluate large-scale data warehouses and data marts.
In five years, 100 million people will be using information-visualization tools on a near daily basis. And products that have visualization as one of their top three features will earn $1 billion per year.
~ Ramana Rao, founder and chief technology officer,
Inxight Software Inc.*
Today, there is an ongoing requirement for more precise decision making because of increased global competition. Generally speaking, decision making should be based on an evaluation of current trends, historical performance metrics, and forecast planning. New and improved BIS continue to emerge to help meet these ongoing requirements.
Within three years, users will begin demanding near-real-time analysis relating to their business — in the same fashion as they monitor stock quotes online today. Monthly and even daily reports won’t be good enough. Business intelligence will be more focused on vertical industries and feature more predictive modeling instead of ad hoc queries.
~ Thomas Chesbrough, executive vice president
BIS vendors are offering a variety of new systems that provide necessary links and end user interface for managers to access and receive selective information such as competitor behavior, industry trends and current decision options. To increase organizational acceptance and use, these new systems feature distributed decision making, which helps leverage organizational visibility. Specific attention is being given to the user interface as highlighted by the following list of standard end user features:
- Filter, sort and analyze data.
- Formulate ad hoc, predefined reports and templates.
- Provide drag and drop capabilities.
- Produce drillable charts and graphs.
- Support multi-languages.
- Generate alternative scenarios.
There are a number of approaches for linking decision making to organizational performance. For example, in the manufacturing industry, decisions may focus on resource allocation optimization and waste reduction, as supported by the Lean Manufacturing Methodology. From a decision maker’s perspective, the new BIS visualization tools such as Dashboards and Scorecards provide a useful way to view data and information. Outcomes displayed include single metrics, graphical trend analysis, capacity gauges, geographical maps, percentage share, stoplights, and variance comparisons. A “Dashboard” type user interface design allows presentation of complex relationships and performance metrics in a format that is easily understandable and digestible by time pressured managers. More specifically, such interface designs significantly shorten the learning curve and thus increase the likelihood of effective utilization. Figure 1 presents an example of a dashboard design.
Figure 1: Example of a Dashboard
A “scorecard” is a custom user interface that helps optimize an organization’s performance by linking inputs and outputs both internally and externally. (The Balanced Scorecard is the specific methodology associated with the Kaplan and Norton model). To be effective, the scorecard must link into the organization’s vision. Over the next few years the differences between dashboards and scorecards will become increasing blurred as these interface structures become fully integrated. Figure 2 illustrates the current adoption of BIS throughout the organization.
Figure 2: BIS Adoptions by Management Area
Figure 3 illustrates the basic structure of how the Dashboard fits into the decision making process. The Dashboard integrates the data warehouses and analytical models directly into the decision making process. This is a continuous process based on ongoing environmental scanning and feedback from current performance metrics, e.g., inventory turns. Behind the graphical interface lie the supportive analytical systems such as statistical analysis for data validation, combined forecasting algorithms, and expert systems for decision options analysis and recommendations.
Figure 3: The Dashboard Interface Structure
The Importance of Training
Training at all levels is a key ingredient in the successful application of BIS. In many applications, training occurs at the last minute and is simply geared towards how to use the system for specific assignments. Intensive training before, during, and after system implementation helps create the cultural change needed to maximize acceptance throughout the organization. Training simulators represent one approach for both improving system utilization and increasing organizational buy in.
Within two to three years, companies will ditch the traditional model of making business adjustments on a quarterly basis. Instead, they’ll use business intelligence and performance management tools to make real-time shifts in strategy to respond to changes in the marketplace.
~ Rob Ashe, president and chief operating officer
Some current technical challenges facing this evolving industry are presented below:
- Integrating optimization based models with enterprise resource planning systems.
- Developing an observation oriented approach to data modeling that includes manual and automated processing.
- Combining decision support, knowledge management, and artificial intelligence in a data warehousing framework.
- Designing intelligent agents that can be used to support decision making processes.
- Formulating adaptive and cooperating systems that use evaluation and feedback to improve the decision making process.
Additionally, speech recognition represents a significant development for improving the human/computer interface. Specifically, a speech interface system would allow the manager to increase the decision making flow volume as well as to explore a broader range of unstructured decision applications.
In the next two years, business intelligence capabilities will become more democratized, with a far greater number of end users across the enterprise using the tools to get better visibility into the performance of their segment of the business. Think of it as executive dashboards for worker bees.
~ Steve Molsberry, senior consultant
Stonebridge Technologies Inc.*
Highlighted below are some specific examples in which dashboards have been successfully applied to improve organizational performance. Following each abstract is a link that will take you to the actual study.
- Hospital Bed Management – The current crisis in the nation’s health care system has triggered an intensified focus on increasing productivity and reducing costs. Two primary goals of a hospital bed management dashboard system are to optimize bed resources and reduce emergency department wait times. The system consists of a number of modules, which include both bed placement and data mining models. Specific displays include real time bed availability forecasts and capacity alerts. In many respects this BI system is like an air traffic controller for hospital beds. For example, it both schedules patient bed assignments as well as facilitates the transfer of patients from other departments. (Bed Management)
- Conflict of Interest Assessment – Prior to taking on a new client, many law firms routinely check throughout the organization to determine the potential for a conflict of interest. Historically, this has required many man hours of effort with the possibility of errors that could significantly affect operating performance. This dashboard based system, which connects attorneys and staff, automatically checks organizational records and results in reduced operating expenses and improved worker productivity. Specifically, the system has reduced the time to conduct conflict checks by 75%. (Conflict)
- Product Development Management – Historically, measuring the performance of ongoing product/service development (PD) has been a hit or miss proposition. This inconsistency has often led to significant overruns and in some cases, total failure. Estimating product development cycle time is key to any effective assessment process. A typical PD dashboard system is designed to report results to date as well as to indicate the potential for continuing success/failure. Project compliance is one key dashboard PD metric. A gauge reports the fraction of new product launches that occurred on schedule and budget. Another standard dashboard metric is the fraction of products/services that has received a favorable trade journal review. Additionally, the dashboard should have the capability of identifying new product/service opportunities. (Product (hyperlink no longer accessible))
- Financial Management – Many financial and investment organizations have concluded that it is essential to have real time updates of key performance metrics such as revenues and profits in order to remain competitive in today’s marketplace. Traditionally, many organizations have relied on quarterly reports to support the decision making process, a practice which has often led to uneven performance. A financial dashboard provides an integrated and real time overview of performance that can be directly correlated to the business model. Specific metrics include balance sheets, income statements and competitor performance. Additionally, the dashboard can display alerts identifying negative trends that require immediate attention.
Each of these applications was developed based on a well designed business intelligence strategy.
Building the Business Intelligence Strategy
Developing an effective business intelligence strategy is predicated on three key drivers: perceived value, organizational utilization and a cost effective solution. The development of a BIS strategy should be tied to specific organizational performance goals and operational objectives. Examples of the latter include increasing customer retention and reducing turnover of key personnel. The proposed solution must be adaptable, scaleable and maintainable. Often a phased schedule in implementing the BIS is best since it tends to minimize risk as well as increase organizational acceptance. Such an approach allows elements of the system to be checked out prior to full system deployment.
Presented in the following list are the major steps involved in developing an effective BIS strategy:
- Establish BIS objectives. (Specifically, what do you want the system to do?)
- Evaluate the current in-house support capability, including the present system’s architecture.
- Perform a gap analysis on existing data systems, including response time.
- Identify alternative technical solutions.
- Formulate an implementation timeline.
- Conduct organizational “Town Hall” meetings to solicit ideas and to enhance the cultural climate for change.
- Determine the need for outsourcing support.
Outsourcing some or all of the implementation process can offer significant benefits to organizations with limited internal technical capabilities or an already strained IT department. Outsourcing also brings the latest in technological development. A first step when considering outsourcing is to assess the organization’s internal infrastructure. This assessment is essential since BIS applications can become very expensive whether developed internally or outsourced. The initial investment for developing a BIS ranges from $1 million to $20 million plus, depending on organizational goals, current IS capabilities, and the projected number of users. The annual system operating expenses can often equal a significant proportion of the initial investment.
Table 1 presents a list of selected BIS vendors. (This list does not imply an endorsement of any vendor. These are presented as examples only.) Generally it is a good idea to start the selection process with the development of a request for proposal (RFP). There are a number of standard RFP formats (no longer accessible) available on the Internet. Obtaining multiple bids will insure both a competitive process as well as serve as a forum to generate additional ideas and technical approaches. Keep in mind that only 50% of all IT oriented projects are completed on budget and on time. A careful check of the references cited by the vendor is essential.
Table 1: Selected BIS Vendors
- The use of BIS throughout most organizations is on the increase as a result of growing global competitive pressures. Improved user friendliness through the use of graphic interfaces is a primary characteristic of the new generation of BIS applications. Specifically, managers require interactive interface systems such as dashboards that are easy to understand and use. Organizational integration represents another important characteristic of BIS.
- Current industry challenges include improving system integration and developing cooperative and adaptive systems that incorporate feedback and evaluation automatically into the decision making process. More specifically, real time speech recognition represents a new technology for improving the human/computer interface that is essential for use by managers at all levels.
- Developing a BIS strategy involves three key issues: perceived value, organizational utilization, and a cost effective solution. The development of a BIS strategy should be tied to specific organizational performance goals. With a carefully crafted plan, organizations can realize significant increases in productivity and insights into the marketplace.
- Ongoing management training is essential for insuring the continued effective use of the BIS. Simulation is one training strategy that provides an effective and dynamic structure for introducing and supporting new BIS applications. Many organizations should consider outsourcing for implementing their BI strategy. The initial investment for a BIS can range from $1 million to $20 million plus, depending on the specific operational requirements.
- Some potential implementation barriers include failure to establish viable performance metrics, failure to fund adequate post-system training, and failure to obtain organizational “buy in.” *Quotations are from “The Future of Business Intelligence,” Computerworld.com.
 Kobana Abulkari and V. Job, “Business Intelligence in Action,: CMA Management, 77, Issue 1, (March, 2003): 15.
 Eric Bonabeau, “Don’t Trust Your Gut,” Harvard Business Review, 81, Issue 15, (May, 2003): 116.
 John Orefica, “Moving to the Next Level,” Health Management Technology, 22, Issue 17, (July, 2001): 46.
 Robert S. Kaplan and David P. Norton, “Using the Balanced Scorecard as a Strategic Management System,” Harvard Business Review, 74, Issue 1 (January/February, 1996): 75.
 Sarah F. Gale, “For ERP Success, Create a Culture Change,” Workforce, 81, Issue 9, (September, 2002): 88.
 Christer Carlsson and Efraim Turban, “DSS Directions for the Next Decade,” Decision Support Systems, 33, Issue 2, (June, 2002): 105.
 Carl M. Rebman, Milam W. Aiken, and Casey G. Cegielski, “Speech Recognition in the Human-Computer Interface,” Information & Management, 40, Issue 6 (July, 2003): 509.
 Sanjay K. Singh, Hugh Watson, and Richard T. Watson, “EIS Support for the Strategic Management Process,” Decision Support Systems, 33, Issue 1, (May 2002): 71.