Business Intelligence Network - The Vision for BI and Beyond

The Matrix – Now at a Computer Near You

Published: October 3, 2006

With a matrix, data can be grouped and structured in ways that are not possible in a standard report table, but does this result in too much of a good thing?

The most widespread business intelligence (BI) activity seems to be reporting. Not multidimensional analyses (online analytical processing/OLAP), not data mining, not data marts, not anything else, but good old-fashioned reports. Furthermore, these reports may not necessarily be based on data coming from a dedicated database (e.g., a data mart or a data warehouse). Instead, and in order to avoid waiting for the implementation of such a dedicated database, many reports are created using data coming directly from operational systems. Reports are that important to the business. 

From a business intelligence perspective, there is also a big difference between reports and other business intelligence output. Whereas tools such as OLAP and data mining demand more proficient power users, most business users are capable of reading reports with their facts, figures and measures. Furthermore, the creation of these reports is most often automated. Consequently, reports may correspond to 80 to 90% of business intelligence output within an organization. 

And what do these reports communicate? Potentially anything and everything about the business. This can be a lot of facts and figures – all put in a spreadsheet that easily becomes cluttered with information.  

Enter the matrix (a.k.a., cross table or nested spreadsheet – but these terms sound a lot less cool). With a matrix, data can be grouped and structured in ways that are not possible in a standard report table. In a matrix, each row and column can have different levels of information, often grouped in categories. As a result, information can be nested into an unlimited number of data dimensions. There are still two dimensions in the sense that we are talking about rows and columns. However, from a data point of view, the matrix is multidimensional as many dimensions of data can be displayed. Inevitably, it risks becoming unreadable, but who is going to say that? Matrices have become a prerequisite to use by any self-respecting (and perhaps self-serving) analyst. 

And given how common matrices have become, all self-respecting report tools now support matrices. No more reports limited to just the classical table format. Add levels, nest levels, and present it all in a matrix! Really advanced matrices can also present several measures for the data dimensions.  

So, who is going to challenge the matrix? The more information that can be presented, the better it must be! And information overload is not a problem. It is just the brain that is too small. After all, given the popularity of multi-nested, overly complex matrices, it must be the right thing, right? 

Without a good matrix, how can, for example, a business maturity model be devised? Business is complex, and only a matrix can reflect this truthfully by being just as complex. It is like when you have a fire in your house and you need to open the window in order to get the smoke out so that you can play chess. It all gets very complex and difficult to understand. 

As matrices are so commonly used as a reporting layout, they have effectively become an important part of business intelligence. Even though matrices are nothing more than reports, their potential complexity should be kept in mind. The usage of matrices can certainly be beneficial. However – and as always – there can be too much of a good thing.  

What is worse is that some businesspeople think that information just has to be presented in a matrix; otherwise, it will not look professional enough. Maybe it is a tactic to always use matrices that quickly become difficult and confusing to understand. This way, it also becomes more difficult to challenge whoever presents the information in the matrix. Questions are rarely asked when things are so confusing that it is not even possible to figure out where to start the questioning. At the same time, matrices seem so obvious. After all, they are just rows and columns. 

Thus, the matrix can be a perfect weapon to report business activities with relatively little risk. Properly utilized, matrices can even serve the business. 

And why not be visionary and go to the next level of the reporting matrix? The drawback of today’s matrices is that they still only have an x-axis and a y-axis, even though these axes can contain many levels of data. What is missing is the z-axis in the matrix! This way, even more information can be presented – and still in a single sheet of paper. (This is not always the case with only x- and y-axes where the number of levels can be too many to be shown in one sheet.) Where the x-axis is the width and the y-axis the height, the z-axis will add depth to the matrix. This will allow for some really deep thinking about what the heck is being presented and the business reality we are in.

If you found this article helpful and would like to receive the latest insights each week from the experts featured on the Business Intelligence Network, please subscribe to the Business Intelligence Network Newsletter.