Return to Newsletter Contents...

Understanding the Full Range of Business Intelligence

by: Tom Meers, President

For those of us who have enjoyed decades of exposure to the IT industry, and I use the word enjoy somewhat loosely here, the concept of “buzz words” is nothing new. Among the many notable hot new trends that culminated in a single phrase only to be abused by every vendor in existence are infamous terms like “client-server” and “portal”. Of late the term Business Intelligence (BI) has gained favor with everyone from the IT department of both large and small companies to leading industry analysts who claim that BI is the fastest growing segment of the software industry.

In spite of all this hoopla, or maybe because of it, most IT professionals, senior managers and end users don’t have any idea what BI really is or what it means. If you ask five different people for a definition of BI you will likely get five different answers, all of which include some reference to reporting or analysis of data. In the past BI was the end result of a data warehouse which housed and organized data from a variety of operational systems. The rule was that BI represented the toolset that you used to provide analysis and reporting from the data warehouse because data in the underlying systems was not adequately organized to answer complex questions that included things like trend analysis and dimensions.

Additionally, there were (and are) concerns about the impact such question asking might have on the organization’s production environment. However, now many vendors are promoting the concept of “operational BI” which suggested that analysis can be done on real-time or near-real-time data. The concept of data federation where source data from a transactional system can be related and transformed in real-time is growing in popularity. These ideas may hold some water, but they also represent danger to the stability and availability of operational systems.

With the never ending redefinition of buzz words it inevitably becomes necessary to think more about what it is you want and need and to stop thinking about what the industry is saying. For example when “client-server” technology was all the rage every vendor in the industry seemed to jump on the band wagon. Not only were database systems considered client-server solutions, but vendors like Novell who offered network operating systems claimed their solutions were “client-server” based. In the end most customers recognized that client-server was a concept that solved a set of real world problems and today most of us see client-server as the foundation for relational database technology which addresses data stability, performance and availability issues that used to pervade PC database and mainframes. Business Intelligence will eventually reach a similar level of maturity. At the risk of contributing to the problem I will endeavor to provide a reasonably complete definition of BI from a practical perspective. The primary solutions that BI systems are designed to provide can be divided into six basic areas:

1. Enterprise Reporting: The creation and distribution of reports in both static (off-line) and parameter driven (on-line) form. Such report distribution can be on-line via a Web browser where a user runs the report based on values he provides to determine what data is returned or in the form of conventional static formats like PDF or Excel sent via email or to a Web site for downloading.
2. Office Application Integration: The ability for end users to access valid and secure data from the BI platform in office applications like word processors, spreadsheets and presentation software. Such access, in this case, will be dynamic. This means that each time the office application is opened the data is refreshed automatically. This shouldn’t be confused with an export of data to an office application.
3. Ad-Hoc Query & Analysis: The ability for end users to create their own reports and analytics by leveraging a semantic layer such as a data cube or meta-data layer that insulates them from the complexities of the underlying data structures. This is typically done on-line and allows users to control all aspects of the content including things like grouping, columns, ranking, summaries, and charting. It also allows users to analyze data in different ways without re-querying the database.
4. Dashboards: Interfaces that summarize data at a high level and provide visual elements such as gauges, charts and trend lines to represent key performance indicators (KPI). Dashboards are typically used to get a high level understanding of the status of something. If properly engineered they will also offer the ability to interact with the data to provide insight into tends and potential outcomes. They will also provide a drill-through feature that allows users to see more detailed levels of the data in the form of linked reports and analytics.
5. Scorecards: An enhancement to dashboards that allows key performance indicators (KPI) to be correlated to predefined goals. This provides a foundation for what is called “Performance Management”. In such cases goals and trends should carry enough sophistication to support dependencies on pre-defined rules. For example a monthly sales goal should be something the BI system can calculate so that it varies from month to month.
6. Data Mining: This element of BI is designed to provide users with the ability to identify patterns in the data that are unknown. Where Ad-Hoc Queries allows users to choose the way in which information is organized and summarized, data mining looks at the data and “find” patterns that are inherent to the data. An example is predictive analysis where data mining calculations help predict future behavior based on historical trends and patterns.

In the final analysis BI is simply a collection of software technologies that makes it possible for end users to access and analyze data. This can, and should, lead to better decisions and faster response to business critical situations. However, it is not a panacea. In spite of what many vendors claim BI by itself will not revolutionize your business. To realize the most from BI, your company must begin with a commitment from management to standardize and plan the business structure and to carefully review who can benefit from improved information distribution and analysis.

While enterprise reporting by itself can represent a significant improvement to a business, misunderstanding the other elements of BI and misapplying them can actually cost your company money by reducing productivity. In the scores of data warehouse and BI system implementations in which our team has taken part several common issues have emerged to suggest that how you approach such an endeavor will, in most cases, dictate the outcome. Beware of half hearted, cheap planning concepts. Without a well vetted and carefully planned blueprint your BI solution will at best offer a questionable return on investment and at worst result in an irrelevant and unusable mess. However, if you take the time upfront to fully understand the different elements of BI and who they will help the most in your business you stand to gain a great deal from the solution. While this approach will cost you time and money up front it will pay you back with success and value in the end.

Go to Top  |  Return to Newsletter Contents