Can’t anybody here play this game? – Casey Stengel

IBM has a new paper out on that age-old question – why do BI projects fail, and what can be done about it? The paper is entitled “Bridge The Gap Between BI Best Practices and Successful Real World Solutions“. The first few pages are the usual marketing fluff, and they generally contradict the “meat” of the paper, which begins a little further in. That is, once again, we see a particular technical/product solution proposed to solve what is not a technical problem. This is accomplished by simply asserting that this particular technical solution maps neatly over the business problems Gartner has uncovered. If you are brave enough to hack your way through to where the Gartner material begins there are some interesting discoveries to be made. By “interesting” I mean “depressing”. Taken as a whole the paper can be thought of as a fine example of what the Gartner research itself reveals.

The paper begins with a set of now-common observations: that BI programs need a business sponsor, that IT ends up “selling” BI to the business (and doing it badly), that BI tends to get “stuck in reporting”, and that “Technology is rarely the culprit if the BI project is considered a failure”. All well and good. And then at page 2 we read, in bold, all-caps:

IBM COGNOS EXPRESS: THE KEY TO A WINNING BI STRATEGY

I see. IBM’s technology will be the “key”. That’s a relief. Gap closed! Close the document and move along.

But if I keep reading, I discover that the folks at Gartner have done some research on the practice of BI programs, most of which are not particularly related to technology (on the contrary.) The results aren’t good. That doesn’t mean they are surprising, of course.

The Gartner section of the paper is called “The BI(G) Discrepancy: Theory and Practice of Business Intelligence”. They break out 9 aspects of BI implementation, and discuss what should be done in each aspect, versus what their research indicates is actually taking place in the real world. The results are a confirmation of what most of us “in the trenches” feel intuitively: there seems to be little correspondence between what should be done, and what actually is done. And technology isn’t going to change that.

The whole thing is worth a read, but the most eye-popping turns out to be the discussion of -  BI strategy! That thing that the latest IBM product will provide a “key” for! Turns out only 2% of organizations informally surveyed in mature markets had anything called a BI strategy. That’s not a typo. 2%. And this is among Gartner clients. Let that sink in for a second, and then consider this quote from the paper:

“Nearly shocking results are obtained when reviewing the so-called BI strategy documents. Almost never would those qualify as strategy in Gartner’s opinion. Quite often a strategy is merely a statement like “We have a Microsoft BI strategy” or “Our BI strategy is SAP” indicating what products the organization is using or planning to implement. Other times the “strategy” is merely an architecture diagram… This is as if the Ferrari Formula 1 team described its racing strategy as “using Bridgestone tires, Shell fuel, a V8 engine and red paint.”

I like the use of “Nearly” to suggest seen-it-all unflappability on the part of the author.

The analyst goes on to describe the initial 2% number as “rather optimistic” (raw-ther, old sport!), blows some dust off the dictionary definition of “strategy”, and then (perhaps beginning to get a little exasperated, and reaching for the bottle) muses that:

“The question could be expanded to: Do executives even understand what constitutes a strategy?”

Yes! It does appears that the question could indeed be expanded to that!

Everyone, and I mean everyone, I have ever encountered in this industry who works above the level of writing reports struggles with the problems outlined in the Gartner material every day. And yet here we are, decades now into the world of BI, and it doesn’t appear to be getting any better. BI still seems to be mired in confusion as to what it is – what is its identity within the organization. The default position seems to be: it’s a technology. IBM et. al. seems ok with this, and I can’t blame them. As long as the discussion can be returned to “BI is a product (and our product is the best!)” they seem to be happy, as they have a tangible thing to sell. My own feeling  (obviously) is whenever the real answer to this question is found, it won’t be “Cognos” or “Microsoft Analysis Services” or any other piece of software, and I say this as someone who spends his days with the products in front of him.

If executives don’t have a grasp of the rudiments of BI strategy (or perhaps strategy in general), it seems that the best anyone can do is try to keep pushing technology. At least that seems to be what IBM’s “strategy” is with this document – provide a high-level summary, name the product and map it to what is “supposed” to happen in an organization, and hope for the best – and that no-one keeps reading. Or what the Gartner analyst, in the section on what goes on in the real world when it comes to the business case for BI, characterizes as a “leap of faith”. I’m not kidding, they actually use those words to describe what Gartner clients are doing to justify their BI investments.

Check the paper out, it’s worth a read.