The Real Deal: David Axson on why most BI projects disappoint

This continues a series of guest columns from practitioners and bloggers I respect. The category - The Real Deal describes them well.

I first met David Axson in 1996 when he was pioneering concepts around process benchmarking at The Hackett Group and have stayed in touch. He is noted author and speaker in business management and technology. He is currently working with a number of organizations on the translation of technology promise into tangible results – which should keep him busy for many years. He was also a former Head of Corporate Planning for Bank of America. 

"Remember when just a few years ago it appeared that SAP stood for Solves All Problems? Well now the focus is squarely on the power of business intelligence to drive increased competitiveness, value creation and any other appropriate buzzword you care to mention – and SAP’s right there with their acquisitions of Business Objects, Outlooksoft and Cartesis to its portfolio. Can their cure for cancer be far behind?

Don’t get me wrong, I fully appreciate the potential of business intelligence tools and applications to dramatically contribute to an organization’s success,but I also see numerous initiatives disappoint. Why the gap between expectation and realization?

Having observed and participated in numerous “BI” initiatives over the last twenty years I think the problem is easy to understand. Take a look at the exhibit below.

 

This is the David Axson view of all IT projects. It’s incredibly simplistic and devoid of acronyms – not an OLAP or XML in sight; but it does encapsulate the main elements of every BI program. First we collect data from somewhere; then the structure it; before putting into an efficient and accessible storage vehicle;the data can then be accessed in order to perform certain transformations or analysis that can be automatic or user-defined; and finally we use it to make decisions. Pretty simple really. However the disconnect comes when one considers the relationship between cost/effort and value. In any BI project, the first four steps consume 95% or more of the effort/cost. We spend lots of time and money sourcing data, cleansing data, organizing data and transforming data. However, when you adopt a value perspective close to 100% of the value occurs in the fifth step when the resulting intelligence is actually used to make a decision. Now in the average project plan and budget, consider how much money or how many tasks are devoted to the first four steps versus the fifth step. All too often the “training” line is minuscule and consists of teaching users how to print and providing them with the number of the help desk which is usually staffed by folks who know little or nothing about how to use information to make better decisions. We deliver incredibly rich applications with little or nig guidance on how to use them effectively. It’s the business equivalent of giving your car keys to your sixteen year old and the only guidance you offer is to, “Go have fun.”

Until we give at least equal weight to the use of information and the associated skills, authority and confidence necessary to make better decisions as we do to the “cool” technology we will continue to lament the disappointing ROI on BI."

David  can be contacted at david@davidaxson.com or at www.davidaxson.com.

Comments

[John Lopez]

I believe another factor contributing to the failure of BI initiatives lies in the long latency between inception and the first deliverable that end users can understand. As you point out, 95% of the effort is up front.

For this reason I believe it is valuable to take the key metrics and deliver a bare bones deployment where training can be done and there is enough interesting data to show the fundamental idea (which has probably been nothing but an abstraction up to that point). Even with provisos that the data is incomplete and as such preliminary findings may be invalidated by the completed system, such a partial delivery can create excitement and support for the project to be carried to completion.

Without such a step it is easy for the BI project to end up on the cutting room floor due to a lack of understanding by those most benefited by the project's completion.

[Gregory Y]

I have to concur. Management does not have appropriate skills in "connecting the dots". There is no culture of constructive, proactive, analytical management.

[Timo Elliott]

As somebody who's passionately interested in BI failure (and what to do about it), I agree wholeheartedly with your points (except maybe about the training -- I'd say it's more like giving the keys to a powerful motorcycle, with a much higher risk of accidents to the driver!)

I would say, however, that one common IT failure is to assume that nothing useful can be delivered without the pain of going through steps 1 through 4.

Smaller projects that can make better use of existing information, or get it into the hands of more people, can be a good way to provide "cover" for the larger, more expensive projects, but this option often isn't considered -- I suspect because of the prevailing "engineering" approach to BI, rather than treating it as a business initiative first and foremost...

Your readers may be interested in more on the same theme:

Why BI Projects Fail: http://www.timoelliott.com/blog/2008/06/why_business_intelligence_proj.html

The Five Fatal Flaws of BI:
http://www.timoelliott.com/blog/2007/03/the_five_fatal_flaws_of_bi.html

[Brendan]

"There is no culture of constructive, proactive, analytical management."

Amen to that.