I recently fielded a forum question about the cost-creation versus value-adding capabilities of information technology (IT) and advanced (i.e., bespoke) analytics in enterprise. Here is how I responded:
Regarding the linkages between information technology (IT), advanced analytics, and value, I would gently suggest that IT is a cost center, and advanced analytics are the value-adding proposition. In other words, don't go to the IT department if you are seeking to activate value-adding analytics (though I will concede that IT does have an effective role in business intelligence [BI] production, which is very different from advanced analytics in my view).
Unfortunately, IT solution providers know full well that advanced analytics is what creates value, and so IT firms will typically "bundle" various analytic offerings with a proposed IT solution in an effort to bamboozle the client into believing that scarce IT dollars can buy both transaction management and advanced analytical services together in one "big" IT installation deal. Buyers of IT solutions should therefore beware.
What is needed today is for IT managers to yield the analytics space to subject matter experts with analytical solutions that stand separate from the data warehousing infrastructure, while seeking to reduce costs in IT by exploiting the economies of scale that IT solutions typically contribute to the cost analysis.
Again, IT is a cost center, while advanced analytics (separate from BI) are the value-adding activity.
1 comment:
I perfectly agree: The IT department is a cost. The advanced, predictive analytic is where the ROI resides!
And I also agree when you are saying that solution providers are doing all they can to "bamboozle the client"!
That's even worse than that! Let's take an example: you buy an analytical suite to create some "ranking" or "scoring" for your enterprise. Depending on you analytical suite (and on the competence of the user), you can obtain a revenue of 40.000 euros or a loss of 13.000 euros. (These numbers are not choose randomly: these are real numbers coming from the KDD98). All depends on the analytical software! And, once again, solution providers are doing all they can to "bamboozle the client", to hide this fact!
The other day, I got an interesting question: "Let's assume that I use an analytical software to produce a ranking. There exists situations where it's not possible to "rank" anything because there are not enough data. Will the analytical software crash in such situation?" The answer is, of course, no. Analytical software never produce "false" predictions but they only produce better (or worse) predictions/rankings. Everything would be a lot simpler if some analytical software simply failed to produce anything. If that would be the case, we could easily "see" what's the right analytical tool to use!
To summarize, the IT department is a cost. The advanced analytic will generate large ROI if you have the right analytical software (and certainly not SAS or SPSS).
Frank - CEO Business-Insight
Post a Comment