Beyond Typical Business Intelligence

NWC GUEST COLUMN

We're so busy maintaining our users' databases and reacting to the vagaries of the business (sometimes called "alignment of IT with the business") that many of us haven't realized we've automated most of the organization's operations and we're now sitting on a data gold mine.

This has led to a mini-growth industry in business intelligence tools--those that create pie charts, dashboards, cubes and the like. But, at best, the knowledge delivered by these tools can be shallow; at worst, the results can be misleading because these tools don't usually take into account the interaction of different factors.

And these BI tools put an unfair burden on users. Although the business operations staff is knowledgeable about what they're doing in their own domains, their knowledge is imperfect. They know what they and others in their field have done. But some of what they believe may not be validated by the data--conventional wisdom isn't always right.

Moreover, some of what they know may not be refined enough. They may know that A, B and C are important factors affecting revenue, but not know the relative importance of each, so they could end up focusing on the weakest of the three.We must go beyond the conventional approach to BI. A whole world of statistical analysis has been built up over the years to reliably analyze the very types of data we maintain. These statistical tools can help refine and deepen the knowledge of everyone in our organization. But how does this kind of effort get started?

Like many of you, I work in information technology and hold the title of chief information officer. These terms were created with at least a vague understanding that we all have responsibility for the meaning as well as the handling of information. So it is perfectly natural that IT would initiate an effort to help people learn from their data.

And after all, the business units are buried in their day-to-day work. (Yes, I know we're buried in our work, too. The difference is, we're supposed to be championing innovation and delivering new value all the time.)

Of course, this works best when IT and the business experts investigate the data together, so we don't go far in the analysis without input from the business units. The joint dialogue and "voyage of discovery" help achieve something else--a true alignment.

Too often, the phrase "aligning IT with the business" implies that IT must breathlessly run to catch up with the business as it goes in whatever direction someone else has determined. True alignment means IT and the business units together define the best direction for the organization to go--and IT shouldn't be afraid to take the lead.IT leadership is also necessary because analytical projects require lots of programming work, even when setting up simple BI tools. (There are no magic bots that go out and find just the right data our users are looking for.)

If we're going to invest in this work anyway, let's use more sophisticated statistical tools that will provide high value results--not just a set of pie charts.

Unlike conventional BI, this kind of data analysis also can become more integrated in everyday systems. After you discover what works best to achieve enterprise goals, the next step is to incorporate this knowledge into the systems that business units use. IT no longer provides just an electronic inbox for transactions, but also some knowledge about how to best handle those transactions--knowledge based on the analysis of all that data we've built up. n

Dr. Norman J. Jacknis is the first CIO of Westchester County, N.Y. He has had extensive experience in the private sector, most recently dealing with enterprise middleware products. Write to him at [email protected].