To see.

If you store your products in a large warehouse and you want to know what’s there, you install rows and rows of lights so you can easily see what’s there. If there are dark corners you want to peer into, you shine a flashlight/torch into the shadows.

BI shines light on the enterprise.

I think the difficulty we’re having in this discussion is a matter of two things: the first is that generalizations generally don’t work well (my fault); and the second is simply a matter of magnitude or perspective.

I work with a lot of technologists who make the switch from departmental to enterprise class projects. And it’s not always an easy switch.

One of the issues that cause this is the very nature of BI and not because of any tools or methodologies. BI requires access to vast quantities of metadata that is not always smoothly integrated between disparate source systems.

In a ficitonal perfect world, a business would start operations only after building a coherent technology foundation that would gracefully evolve along with the business. New software and hardware would not be allowed into the enterprise if it did not plug right into the existing infrastructure. The unfortunate part of that is that the business is then limited by the technology, which is never a successful arrangement.

In reality, businesses make business decisions, damn the technology. Even Microsoft and Oracle work this way. They’re better at using their own products today the way they’ve been telling thier clients to use them for years, but there are still exceptions. They are certainly not going to let software get in the way of business.

All of this is where the IT promise breaks down–we don’t really deliver on everything we promised we would. You don’t have a cohesive system of integrated sources that executive management can exploit. You have a series of systems that are used departmentally because those systems do what the departments need them to (IT must serve the business). Are you going to make everybody stop using systems just because one uses COBOL and one uses Java, etc.? Just look at how many reports are pumped into Excel spreadsheets and you’ll have your answer.

BI (of which Analytics is a piece) is a tool/skill/exercise to abstract all of the source systems and present a coherent, homogenized business view of the data that business owners can use to make on-the-spot command decisions as well as strategic decisions after doing some analysis.

BI is IT. It’s just from a different perspective and magnitude. If a form validator is a part of an ERP application, the ERP application is a part of the enterprise BI (wish I could do graphics here).

Think of a cellphone. Is it a single example of IT or is it a BI tool? Certainly it is a defined product resulting from a defined project. But, it can ring (alerting me of an issue); I answer it (drilling down with my questions); I hang up and call someone else to ask him questions (drill through); call a third person, tell him my decision and authorize him to execute it. How does the phone work differently if I’m a departmental manager or the CEO?

I like Cognos as a BI tool because it’s easy to use and I can force it to do what I want it to do. But it’s just a tool and is only as good as I make it. JDBC is a powerful tool, but only in the hands of a competent Java developer.

I guess what I’m saying is that parts are parts. The only real difference between BI/Analytics and non-BI/Analytics is that non-BI/Analytics technologists are fearful, leary, unsure of how BI/Analytics is going to affect them and their hard built projects.

IT is necessarily entropic. IT departments are either rigid or flexible. The flexible ones are able to adapt to business needs. Rigid ones often aren’t. BI is not simply a business inquiry system, it is a facilitator for IT change (IT can use BI in so many ways the business users rarely care about or understand). And it is up to the BI team to prove it (BI must serve two communities:IT and the business).

In summary, BI/Analytics has the same roles and responsibilities as any other IT function and the same partnership with business. The perspective and the magnitude are larger at the enterprise level, but the mechanics are more or less the same. And if it’s not in response to a business need or opportunity, it’s worthless anyway.

What a coincidence! I was grappling with exactly the same thoughts just two days back while trying to explain this to a client of mine.

Like you, both of us felt that BI/Analytics was more than just ‘a lights-on or dialtone-on’ kind of thing. And yet how do you describe that elusive difference, because analytics is far more than just deploying a couple of IT products (as a lot of BI software vendors would want you to believe).

Your point is very valid, it is unfortunate that IT and analytics are often dealt with at an arms distance from business strategy, and it is even more unfortunate that improving revenue, reducing cost, improving physical ops and reducing logical ops are things you rarely see technologists talk about.

P.S.: I wrote a rather long reply to your comment and then decided to make it a post instead. Okay, so my Christmas post is dedicated to you. Cheers!

P.P.S: I will go ahead and post your question for dratz to comment upon. I am sure he will have a few great ideas on the matter.

P.P.P.S.: You have a cool blog. Wish I had discovered it earlier! Keep up the good work!