Roman Stanek, who is the founder & CEO of Good Data, had an interesting post where he declared: "Friends Don't Let Friends Overpay for BI."
"Business Intelligence projects are famous for low success rates, high costs and time overruns. The economics of BI are visibly broken, and have been for years."
"Nobody argues with the need for more Business Intelligence; BI is one of the few remaining IT initiatives that can make companies more competitive. But only the largest companies can live with the costs or the high failure rates. BI is a luxury."
I agree. I think that many BI projects do cost much more than originally scoped, as per Roman's blog, but the reasons are typically around the wrong expectations, and bad technology. But, perhaps that is what he's getting at.
Roman sites the reasons for the BI failures, including:
"1) They don't understand elastic scale"
"2) They try to control BI with a single version of the truth"
"3) They cannot measure success of BI"
To the first point, those who build and maintain BI systems, have no idea how to scale up their BI systems, in order to adjust to the needs of the business. Tossing hardware and software at the problem does not work, nor does limiting use. You have to architect to scale, most don't bother.
The second point is something I agree with wholeheartedly. In essence, we're dealing with "near time," or better said "some time" data, based on a past instance of time, and not on the real time aspects of the business. Thus, the data is old, erroneous, and thus not at all useful.
The final point is a systemic issue as well. Ask somebody to define the value of their BI solution, and few will be able to do it. What's the ROI, and what's considered a successful BI implementation? If it does not have value, we should not do it, and I suspect that many BI implementations are not earning their keep.