Do We Need to Know Exact Cost of Failed IT Projects?

Ann All

My post from last week on IT project failures gave a prominent shout-out to Roger Sessions, chief technology officer at ObjectWatch Inc., and author of "Simple Architectures for Complex Enterprises." He's gotten a lot of recent attention in the blogosphere with a white paper in which he estimates the annual cost of failed IT projects at an astronomical $6.2 trillion.

 

Among those who disagree with Sessions' methodology is IT Business Edge contributor Dennis Byron, who came up with a far lower $500 billion estimate for the annual cost of failed IT projects when he crunched the numbers. You can make your own judgment -- and see those of others, including bloggers Michael Krigsman, Leo de Sousa and Todd Williams -- by downloading Sessions' white paper, which is titled "The IT Complexity Crisis: Danger and Opportunity."

 

Writing on ZDNet, Michael Krigsman this morning highlights another critique of Sessions' numbers, this one by IT consultant Bruce Webster. Krigsman shares Webster's key points and knocks him a bit for not providing numbers of his own. Reworking Sessions' original numbers using different assumptions -- as Byron did in his post for IT Business Edge -- would be "a worthwhile project," writes Krigsman.

 

Not being a natural numbers person, I'm a little frustrated by all this wrangling over the precise cost of IT failures. As Sessions wrote in a comment on my post:

 

I hope readers don't get overly focused on the exact cost of the IT failure. The much more important message is that what we are doing today is not working. We can either continue doing what we are doing even though it isn't working (the traditional definition of insanity!) or start doing something else.

 


One thing I think we can all agree on, the cost (whatever it is) is too high. Focusing narrowly on the numbers is an interesting mathematical exercise, but I worry it could divert attention from actually solving some of the problems that commonly wreck IT projects. Sessions certainly hits on a serious one: projects that are just too large and complex to succeed.

 

A main impetus of Agile software development is to reduce complexity, which is why more companies are using some form of Agile for their tech implementations. IT Business Edge parent Narrowcast Group recently switched from a traditional waterfall development process to a modified Agile one. We posted our new project development process in our Knowledge Network earlier this month.

 

Krigsman makes a fine point, however. Coming up with a model that would more accurately quantify the cost of IT failures should reveal some of the more common causes of IT failure and help companies make smarter IT investments.



Add Comment      Leave a comment on this blog post
Dec 29, 2009 1:10 AM Michael Krigsman Michael Krigsman  says:

Ann, thanks for continued focus on this important issue. I agree with your title -- knowing the precise scope of failure is less important than being aware it is a very large number.

Reply
Dec 30, 2009 1:40 AM Leo de Sousa Leo de Sousa  says:

Ann,  thank you for the mention and putting a focus on this important topic.

For me, complexity has a huge cost to my publicly funded organization.  As an example, we reviewed our application delivery platforms and made a conscious decision to simplify from 7 to 3 platforms. 

The benefits are much simpler systems, facilitating our staff to remain current with their skills and faster solution architecture decisions (fewer platforms to choose from = minimize the "technology religious wars").  The end result is more projects get delivered successfully (fewer failures).

So, I agree the exact cost of failures is much less important than recognizing that the number is large and we need to do things differently.

Best regards, Leo

Reply
Dec 30, 2009 1:59 AM Aleks Buterman Aleks Buterman  says: in response to Leo de Sousa

Ann -

   I'm glad to see this getting more spotlight.  Roger and I have been going back and forth on the actual numbers (on both our blogs), and the $500b number is close to what my analysis was before recognizing the cost of benefits lost.  Another critique vector is about failure of project management metrics to actually provide enough insight to judge whether an effort failed or succeeded - and even those metrics are not generally well maintained by the organizations.

   One more thought on your post - Agile development practices in their pure form (i.e. without proper oversight mechanisms) tend to increase complexity of the overall environment, rather than reverse.  The practice of re-use has not been ingrained in the application development community to forestall the duplication (one major driver of complexity) that naturally occurs in a decentralized or federated environments.  So I'm glad to see that it's a modified Agile that Narrowcast has switched to, and it may be of interest to readers to know what other project management structural steel has implemented to deter complexity. 

Regards,

Aleks

Reply
Dec 30, 2009 7:34 AM Roger Sessions Roger Sessions  says:

Hi Ann,

Thanks for the discussion. I have responded to Webster's criticism at Michael's blog. Rather than repeat that here, let me refer readers over there.

Hoping for a simpler 2010!

- Roger

Reply
Feb 1, 2010 1:54 AM Robert Dempsey Robert Dempsey  says: in response to Aleks Buterman

Hi Aleks,

Being a set of principles, it is not the purpose of Agile to tell software teams how to develop software. While some Agile methods such as Extreme Programming do prescribe certain engineering practices such as TDD and continuous integration, most do not as that is not their purview. Also, I must disagree when you say that Agile increases complexity. I suppose it does if you are in an environment where there is no process in place; otherwise, in most cases I've seen it make things much simpler.

Reply

Post a comment

 

 

 

 


(Maximum characters: 1200). You have 1200 characters left.

 

null
null

 

Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.