Wayne Eckerson is a noted BI consultant who spoke recently to the Ottawa TDWI chapter. I’d call Wayne a guru, but someone once told me that guru was a polite word for charlatan. Wayne is the very opposite – he is a very down-to-earth speaker who delivered a direct, unpretentious and thoughtful presentation on the subject of BI organizational architecture.

One of Wayne’s interesting observations was that he sees the need for what he calls “purple people” for any successful BI organization. If we think of people on the business side as “blue” and the people on the IT side as “red”, then “purple people” are people that have a mix of skills that enable them to be effective at bridging the gap between the two worlds. I spoke to Wayne afterwards and he elaborated on the idea:

“Purple people are a blend of business and IT – not blue in business or red in IT but a combination of both. These are both senior and junior level folks. At the senior level, some start in the business and end up in IT and then usually come back to the business where they run a business technology group that acts as an interface between the business and IT. (In the BI world I call these teams BOBI – business-oriented BI teams.)  Some in IT become very conversant with the business and do a good job meeting business needs. These are directors of BI who interface with business executives more than their technical teams just about, to present budgets, roadmaps, funding requests, etc.

At the junior level, things are trickier, and not as effective. Most companies have business requirements analysts who interview business people, gather requirements, and translate those into specs for developers. I usually find there is a lot lost in translation with these junior level purple people.”

Another one of his key observations in the presentation was that from a BI architecture/organizational perspective, we can think of reporting as being a top-down process, with (we hope!) needs analysis, clearly defined specs, a process for building and moving data marts and reports into production, various controlling structures and so on.

Analysis, however, doesn’t really lend itself to this kind of approach – analysts may not know the questions that they want answered until they begin to delve into the data in a very ad-hoc kind of way. They want to quickly add data sources, join things together, and perform analysis that will lead to more questions, potentially the requirement for more data sources, and so on.

This leads to the business attempting to work around IT to get what they want, including bringing in tools that IT isn’t prepared to support. Analysis ends up being a volatile, bottom-up process, driven by the business, and the organization may struggle to keep it under control. IT fears chaos, but – to some degree – real analysis has a chaotic, or at least unpredictable, character. BI practice has to recognize the contrast in the very natures of reporting and analysis to be effective.

Wayne is a regular blogger and author of books and reports, such as Performance Dashboards: Measuring, Monitoring, and Managing Your Business. If you get the opportunity to hear Wayne speak take advantage of it – he delivers a lot of thought-provoking content that has application in the real world.