Browse DevX
Sign up for e-mail newsletters from DevX


Calculating Baseline Metrics for Business Intelligence

At first glance, calculating a baseline metric—an average of some measure over a specific period of time—appears to be a simple task. In reality, baseline metrics are slippery beasts, and not at all as simple as they appear.




Building the Right Environment to Support AI, Machine Learning and Deep Learning

ven a casual observer of today's IT scene has seen the buzzword storm that has developed around various forms of technology-enabled business management and monitoring: Business Performance Management (BPM), Business Process Management (also BPM), Corporate Performance Management (CPM), Business Activity Monitoring (BAM), and others. At first glance these acronyms all appear to be Business Intelligence (BI) by other names. It can be difficult to discern whether such designations simply describe the latest BI marketing concept or are new and distinct approaches. However, it is safe to say that all these approaches depend on the tools and techniques of BI; therefore, any discussion of basic BI techniques applies equally well to all these areas. This article discusses a type of business metric commonly needed for all these approaches: baselines. What Are Baselines?
Baselines are time-lagged calculations (usually averages of one sort or another) which provide a basis for making comparisons of past performance to current performance. A baseline may also be forward-looking, such as when you establish a goal and are seeking to determine whether the trends show you're likely to meet that goal—an essential piece of a Key Performance Indicator (KPI). Baselines have been around for as long as there have been analytic approaches to measuring execution and its results—in business, athletics, and medicine, to name a few.

Prior to the multidimensional cube technologies associated with BI, calculating useful baselines was difficult and expensive; analysts produced baselines infrequently and only for the most critical measures. Today, BI makes it practical to create a baseline for virtually any metric at any frequency. But that new simplicity is deceptive; calculating a meaningful baseline can still be complex, in ways that are very application-specific. In this article I'll illustrate this deceptive complexity using an example based on workflow analytics, in which the basic metrics and their baselines relate to transaction-processing volumes in a multi-national operational setting. Don't get hung up on this specific example though. The issues and ideas are applicable to a wide range of analytic applications.

Comment and Contribute






(Maximum characters: 1200). You have 1200 characters left.



Thanks for your registration, follow us on our social networks to keep up-to-date