How Do You Use BI to Correctly Drive Future Strategies?

I’d say that driving without a look at the rear mirror can be quite dangerous. Rather than being holistic, we should consider BI as one technology which contributes to better management, rather than as an isolated panacea. If you can apply BI to on-going data in order to make near real-time decisions, then you are not just looking backwards. Say that you approach traffic lights, and the green blinks announcing that it would soon turn to red. Should you hit the accelerator or the breaks? If your BI can give this answer, taking into account past behaviour and data as well as present data (speed, location, ...) – then you have a good implementation. We see this kind of implementation more and more. In the context of Process Management, one of the fast growing products (Appian) actually stems from BI and applies this technology to many facets of process management. Other BPM vendors do similar things. BI is also increasingly integrated in the Office environment with add-on products such as Panorama. History is part of our present and certainly impacts our future, and those who have a good understanding and insight of history usually are able to better interpret the present and make sound decisions about the future

Avigdor has it right. One of the key elements in an performance management / business intelligence solution is "context". Is $100 million dollars in sales for the quarter good or bad - well that depends on the context.

The historical information contained in a BI application goes a long way to developing the right context for the metrics measured in the future. What makes a Key Performance Indicator effective - historical context. Context can be materialized in plans, historical patterns or management goals. While certain use cases need real time information, they tend to be "noisy" with a high degree of variance and fluctuation. Historical analysis provides perspective. Understanding trends allows variations to be -you guessed it - put into context. Normalizing data also allows disparate systems data to be more clearly understood as a cohesive information platform.

In the example above, if a retailer is two thirds though a quarter and their revenue is at less than a third of its goal for that quarter, is that an issue? If it is the retailers fiscal forth quarter, maybe not. But if it is the retailers fiscal first quarter, it is time for an aggressive Winter Sale! Tell the BI team not to be mesmerized by the real time POS Dashboard gadget - get the management team on phone - you have a problem.

Sometimes we we try to make the latest technology the case for for expanding the scope of our business intelligence initiatives - we've all done it. But the best organizations balance what can be done with what should be done.

In just about every effective implementation of a BI initiative I have seen over the last twenty years, some smart manager has asked this very important question. "If you had that information in our BI solution what would you do with it?" That question is normally then followed by "can we quantify any additional expense and complexity to a materially measurable outcome?"

These are the type of question that should drive decisions in a business intelligence system strategy. Then let the technologist decide if it is a real time web service, CEP logic or predictive algorithm or some other slick technology that is needed.