In the beginning, organizations simply delivered reports to users. The reports were predefined, static and paper based and shed light on past activity. Over the years, reporting moved online and became more interactive. Some parameterized reporting screens with drop-down pick lists make users think they are performing ad hoc queries. Nevertheless, reporting was the first significant dimension of BI.
But in the 1990s, reporting became passé. Many users clamored for personal analysis tools that would allow them to explore and analyze data in a relatively unfettered fashion. Consequently, vendors delivered ad hoc query tools and
However, two things happened at this point to cause a major crisis of faith among BI tools vendors. First, it turned out that only a small minority of users really had much interest in using ad hoc query and sophisticated OLAP tools. Most users found these tools much too hard to use or navigate without getting lost. "Slice and dice" turned into "lost and frustrated" and created a host of BI shelfware.
Second, most BI vendors abandoned reporting and almost too late rediscovered that most users prefer viewing reports to exploring data in an unfettered manner. In the past several years, leading BI vendors have scrambled to cover their reporting flanks by developing or acquiring reporting tools. With reporting and analysis tools in hand, users would have a complete set of BI tools, or so vendors thought.
Figure 1. The five dimensions of BI
Figure 1. BI tools evolved from reporting to analysis tools, which many users simply used to dump large data sets into Excel to do planning. Today, companies are adding a monitoring layer and new advanced analytics and visualization to the stack.
Unfortunately, BI vendors gradually discovered that most users were using analysis and reporting tools as glorified extract programs to dump data warehouse or source system data into Excel. Although they were glad for the sales, users often complained of poor query performance. It turns out that business analysts and managers who needed to create business plans and budgets, model scenarios, or recalculate forecasts would use the BI tools to grab large data sets and move them into Excel, Access or a statistical modeling program. These runaway queries would bog down query performance for everyone else.
Today, many BI vendors now sell business performance management software. What they're really doing is acknowledging the third dimension of BI tool evolution: the planning layer. Today, most BI vendors offer much tighter integration with Excel because it is by far and away the best all-purpose business planning and modeling tool available, and probably always will be. But some BI vendors now offer budget and planning tools that are fairly well integrated with reporting and analysis tools, creating a business performance management (BPM) suite.
But a combination of reporting, analysis and planning still doesn't do the trick for most users. The BI tools make it too hard to find the right data and too easy to get lost in it. The planning tools are not well integrated with their analysis tools. What most users really want is a layer on top of analysis, reporting and planning tools that pulls them all together in a highly simplified and intuitive interface called a dashboard or scorecard. These tools enable business users to track or monitor the few metrics they care about most, compare actual performance to predefined targets and trigger alerts when performance strays too much from goals.
These dashboards and scorecards represent the fourth dimension of BI. Today, most organizations use different, non-integrated BI tools to deliver reporting, analysis, planning and monitoring capabilities. Ideally, all work harmoniously on a common BI and data infrastructure.
Figure 2. BI platforms
Most leading BI vendors are now working feverishly to deliver such integrated solutions. Many now offer BI platforms that provide a common BI and data infrastructure to support a variety of BI activities. (See Figure 2.)
Figure 2. BI platforms provide integrated BI tools or modules that are aligned with business strategy and run on a common BI and data infrastructure.
New dimensions -- Advanced analytics
But while BI vendors rush to offer a BI platform that supports reporting, analysis, planning and monitoring, user requirements keep moving ahead. There is now a groundswell of interest in exploiting new dimensions of BI, namely data mining, text mining and advanced visualization. Many companies are now asking, "How can we deliver further value from our data warehousing/BI investments?" and they are turning to these solutions.
Some companies have already created custom solutions or used early adopter technologies to create sophisticated statistical models to detect fraud, automate cross-sell recommendations, define loan prices and terms, predict customer behavior and machinery breakdowns, among many other things. Similarly, some companies are analyzing call center interactions, email responses, Web pages, video and image files to understand key trends and patterns in text and other unstructured. Because there is so much data involved in this type of deep analysis, many are leveraging advanced visualization techniques to quickly detect key patterns and rules or display data in usable forms, such as maps.
While I have defined five dimensions of BI, there are probably many more. BI is a rich field, one built on the continuous exploration of knowledge and pursuit of truth in our businesses. Since knowledge and truth is fleeting and evanescent, we will always need new tools to help us capture insights required to run our businesses. And this is a good thing since it means there's a rich and lively road ahead for those of us in the business intelligence field.
Wayne Eckerson is director of research at The Data Warehousing Institute, a provider of in-depth, high-quality training and education in the data warehousing and business intelligence fields.