by Shaku Atre
Business Intelligence: From Theory to Reality (Part I)
Business intelligence (BI) decision-support initiatives are costly and difficult. A staggering 60% of all BI projects ultimately fail by curtailing the initial goals. But it needn’t be that way.
BI projects are costly because disparate business data must be extracted and merged from a variety of data sources, including online transaction processing (OLTP) systems, batch systems, and external syndicates. Also, many BI decision-support initiatives require new technologies, new tasks, and new roles and responsibilities. Making the projects even more costly, analysis and decision-improving applications must be delivered promptly and with high quality.
BI projects are difficult to implement, and as a result they fail in such great number due to a long list of mistakes and missteps. These include inadequate planning, missed tasks, missed deadlines, poor project management, undelivered business requirements, and low-quality deliverables.
There is an alternative. With this alternative, the high costs of BI implementations can be reduced. The failure rates can be lowered. But to achieve these improvements, technology managers must use the right tools and an appropriate method for implementing the BI decision-support systems. They must also open their eyes to the four new realities of BI.
First, technology managers must realize that BI is neither a product nor a system. Instead, it is an architecture and a collection of integrated operational systems. BI can also include decision-support applications and databases that ease access to business data.
Examples of BI decision-support applications include:
- Multi-dimensional analysis (OLAP)
- Click-stream analysis
- Data mining
- Business analytics
- Balanced scorecard
Examples of BI decision-support databases include:
- Enterprise-wide data warehouses
- Data marts (functional and departmental)
- Exploration warehouses (statistical)
- Data mining databases
- Web warehouses (click-stream)
Second, managers must learn that BI implementations, like nearly every kind of engineering project, pass through six stages between inception and implementation. These six stages are:
- Justification: An assessment is made of a business problem or a business opportunity, which gives rise to the engineering project.
- Planning: Strategic and tactical plans are developed, which lay out how the engineering project will be accomplished.
- Business Analysis: Detailed analysis of the business problem or business opportunity is performed to gain a solid understanding of the business requirements for a potential solution (product).
- Design: A product is conceived, which solves the business problem or enables the business opportunity.
- Construction: The conceived product is built, and it is expected to provide a return on the development investment within a predefined timeframe.
- Deployment: The finished product is implemented (or sold) and its effectiveness is measured to determine whether the solution meets, exceeds, or fails to meet the expected return on investment.
Third, managers must realize that BI systems, like many other engineering processes, are iterative. This means that after a BI system is deployed, it’s continually improved and enhanced, based on feedback from its business users. Each iteration, in turn, produces a new product release, or version. In this way, the product evolves and matures. Also, BI’s iterative process produces an important side effect: It renders the system-development practices of the past both inadequate and inappropriate.
In the past, every system had a clear beginning and an end. Each system was also developed for one set of knowledge workers from one business line. As a result, cross-organizational activities were unnecessary. In fact, early designers viewed cross-organizational activity as a barrier that only slowed project progress.
Today, however, these development practices are no longer suitable. Modern, integrated BI initiatives must take into account the cross-organizational activities necessary to sustain an enterprise-wide decision-support environment. That, in turn, requires new development skills, practices, and techniques.
Fourth, managers must learn that the dynamic, integrated BI decision-support environment is iterative in nature. In other words, it cannot be built in one big bang. Instead, data and functionality must be rolled out in releases. Each deployment will then likely trigger new requirements for the next release.