Gartner publiced it's five predictions on BI the coming year. Coming days I will try to give my opinion on a few of them. Let me begin with the second prediction of Gartner:
Gartner's second prediction is that by 2012 business units, rather than IT, will control at least 40 percent of the total budget for BI. The problem here, Gartner wrote, is that "business users have lost confidence in the ability of [IT] to deliver the information they need to make decisions."
So
business units will increase spending on packaged analytics, including
corporate performance management, predictive analytics, and online
marketing BI wares. In so doing, however, they "risk creating silos of
applications and information, which will limit cross-function analysis,
add complexity, and delay corporate planning and execution of changes,"
said Nigel Rayner, research vice president at Gartner, in a prepared
statement.
I support the analysis of the problem by Gartner. The business has lost confidence in the ability of IT to deliver.....And since we live in a 'every-problem-needs-a-quick-fix' - world, I also support the risc Gartner states of business units, that will go shop and wil find vendors that sell pretty pictures (see my fairytales blogpost). The industry teaches it wrong.......
So...what now?
Well, first of all I wonder whether this issue Gartner is rising, is really new. I mean, everywhere I come, decentralized BI packages/solutions have been implemented mainly because 'IT could not deliver'. Whether or not that's true is not really the issue. I just do not see how it's a prediction?? Can somebody help me here?
Now, let's embark on some kind of direction we - as an industry, as a trade - need to communicate. Business needs to be in control of the BI function. But...
A metaphor....Imagine the automotive industry (I know - bad one atm). Organisation's in that industry need to construct a (BI) factories with production lines (processes), robots, vendor (sources) agreements delivering parts (data), stock (CDW) etc..Now, this factory exists simply because it's efficient and cost effective, but most of all, because it is able to deliver on demand (coming from customers). Now, when these customers wanna order a car, do you think they wanna build a factory in their garden? Well, offcourse they don't....they are not even capable of doing that. Are they capable of formulating their demand - yes they are. How? Well, they have a catalog showing all kinds of characteristics (non-functionals) and options (functionality and content/metadata).
Setting up a centralized BI factory is an organisation's corporate concern. It's a basic service the organisation needs to build. Does this factory make pre-build cars? Maybe - but only when there's actual demand for it. Does this factory do custom builds? Well, to a certain extend where it remains efficient and cost effective.
Who payes what? Here the metaphor breaks a bit. Off course, customer's are paying for their 'cars and it's maintenance' - they know what kind of car they need for their particular purpose (work-related, private related, home-school etc..). It would be sooooo weird when the factory is paying for the cars and gives it to their customer saying.....'This is what you need'....
So - Business units are paying for their (BI) informational products. Nothing new I suppose? Duh....
But who is paying for the factory? Here the metaphor breaks a bit. All though - let's assume there are no car factories anymore (not unrealistic). Customers will still need car's. Like I said - it's not realistic that every customer will build there own car. So they will probably organise themselves (maybe in a cooperation or so) and let them be represented by a few of their own that are capable of setting up car factory. This 'cooperation' is going to build a factory with a budget the members have contributed.
You still with me?
What I am trying to say is that the BI central function, the BI factory, need to be financed by business. The execution and governance however, needs to be as centralized as possible to achieve certain economies of scale.
So - Business units need to finance the BI centralized function/factory.
And THIS IS THE REAL PROBLEM. Because, that is often not the case. Furthermore, the organisational structure of Business units with there own accountability is failing miserably at constructing/financing a centralized BI function. In the Netherlands a professor wrote very good book about this: from unit organisation to multi dimensional organisation (unfortunately in Dutch) by H.Strikwerda. Setting up centralized functions in a business unit oriented organisation is not possible.....Read this book!
So the problem is much deeper and more fundamentally grounded in organisational theory. It therefor extends BI big time!
Pfff....in writing this I embarked on something I should make a thesis on and go underground for some years. It's a huge subject.....I might do that....
Those who read this and still understand where I am going....let me know what you think. It's a complex subject, but sooo interesting to explore even more.
Hi Ronald,
I think it's a fantastic subject. Very deep thinking is required, and I believe their is no true solution at the end of the day - anyone who can figure out a full-solution to this problem may become very rich, very quickly if they can devise a manner in which to capitalize on it's knowledge.
Anyhow, it's a deep subject, and one that will be around for a very long time. As far as the Data Vault goes - methodology and all - this is something we try to address, but only in bite-sized chunks, not as an enterprise whole.
Remember this though: accountability needs to be in both IT and Business Units, all the way up to Business Execs & ownership, or - no solution will make sense.
Cheers for now, I must be tired.
Dan L
Posted by: Dan Linstedt | Thursday, January 29, 2009 at 07:35 PM