One Expert Pegs Yearly Cost of IT Failure At $6.2 Trillion

blognoggle writes "Roger Sessions, a noted author and expert on complexity, developed a model for calculating the total global cost of IT failure. Roger describes his approach in a white paper titled The IT Complexity Crisis: Danger and Opportunity. He concludes that IT failure costs the global economy a staggering $6.2 trillion per year."

Comments

incompetence

[X10]

Does this mean that IT people are generally incompetent? Or is it just the IT managers who are incompetent? Or, just maybe, it's all IT people who don't read /. who are. Hmm..

Re:incompetence

[ta bu shi da yu]

I'd look at it differently. I would firstly work out exactly how much money is generated through effective IT services and projects, and then I'd work out how much money is saved through effective IT services and projects, and then work out how much is lost through projects that go wrong. I think this sort of analysis would give a more true picture of the benefits and risks of IT projects.

Re:incompetence

[pnewhook]

The amount of time / effort / money I've lost over the years due to buggy and crashing computer software is staggering. And I solely blame this on incompetent software developers. I'm talking of both commercial software (I'm surprised they let some of this crap out the door - do they know what testing is?), and also my own experiences working with development teams.

I've had developers work for me that think they know everything there is to know, refuse to listen to any advice, and basically try to write software only in the way they believe it should be done, completely ignoring the needs and requirements of the system lead and the customer. Throw in to the mix some elitism and a complete lack of ability to communicate without insulting an derogatory statements, and you've got a profile of a large percentage of current software developers. I'm still working to undue to colossal mess of my last ex-software lead that I ended up kicking off the program because he fundamentally didn't know what he was doing (despite thinking he was the best developer on the planet). I've also worked with some amazingly brilliant software developers, but unfortunately they are few and far between. The sheer arrogance of some software developers is astounding.

Re:incompetence

[ozmanjusri]

Company, Annual turnover, $mln. 2004

1. Symantec 1364
2. McAfee (NAI) 597
3. Trend Micro 508

Which OS is costing this?

Which company just blocked the best efforts of the rest of the world to develop an interoperable set of document formats?

Microsoft has repeatedly prevented progress in computing. the opportunity costs of that alone are incalculable.

Re:incompetence

[Bucc5062]

The amount of time / effort / money I've lost over the years due to buggy and crashing computer software is staggering

The amount of time / effort / money lost over the years due to poor management, bad analysis, and improbable times lines is staggering.

There, fixed it for you. You do see that your own statement is about as arrogant and condescending as the programmers you want to insult. Buggy code, crashing software is not just the responsibility of the programmer, it is the responsibility of the leadership as well. Why was it buggy? Bad design specs, no code reviews, tight time lines with large interruptions? Why did it crash? Poor QA and review by business owners? ridiculous deadlines, poor working conditions, low morale?

There is more there then just "bad programming" as if programming exists in some bubble. Developing is not assembly line work, it is a complex art and yet over decades management has viewed it from an industrial age mentality. Work from x to y, produce x lines of code, stop what you are doing and look at something else no matter where you are at. Certainly there are arrogant programmers, just like there are arrogant managers. I challenge you though to see that both need each other to reduce the number of bugs, the minimizing of crashes (really "crashing computer software? Not Abending or exception failures?) When a positive work environment is set that people tend to work better, with less error. That is the job of management and yes, even leads. For the record, I have been in lead and oversight positions. The best role I played was to get out of the way and let my people do their job. Along the way I would just ensure that we maintained a high quality of effort and we kept on focus to the requirements provided.

Re:incompetence

[A_Lost_Frenchman]

No ! It just means the guy who wrote the white paper, and the guy who comments on it, are both incompetent.

A large number of these will eventually fail. I assume the failure rate of an "at risk" project is between 50% and 80%. For this analysis, I'll use the average: 65%.

Using the same kind of bullshit reasoning here is what I found: A large number of human beings will eventually die. I assume that human beings live between 0 and 100 years. For this analysis, I'll use the average: 50 years. Except that the average life expectancy is not 50 years but actually much higher. Taking the mean of the minimum and the maximum is not at all the same as taking an average, you may as well be pulling the numbers right out of your ass.

To find the predicted cost of annual IT failure, we then multiply these numbers together: .0275 (fraction of GDP on IT) X .66 (fraction of IT at risk) X .65 (failure rate of at risk projects) X 7.5 (indirect costs) = .089. To predict the cost of IT failure on any country, multiply its GDP by .089.

You're trying to introduce a global economic indicator using only 1st grade calculus, that's certainly an interesting approach. So the basic reasoning is that 65% of all IT projects fail, and when they fail, not only do we lose everything that was invested in this particular project, but because of the indirect costs, we are actually going to lose 7.5 times more money ! There is so much bullshit in this sentence I don't even know where to start ! First of all, is the project a failure because it was delivered late, because it is not completely satisfactory, because there are bugs ? In any case, there is almost no chance that the project is such a failure that we can't get anything out of it. What's more there is no way it is going to cost 7.5 times more money than that, which leads me to all the stupid assumptions.

1. explicit assumption: 66% of all Federal IT dollars are invested in projects that are "at risk". I assume this number is representative of the rest of the world
   => It's not. The US is not even remotely representative of the rest of the world
2. explicit assumption: I assume the failure rate of an "at risk" project is between 50% and 80%.
   => Maybe you could have looked up the real number included in the definition of an "at risk" project. For all we know it could be 10% of 90%, assuming you know the number when you actually don't doesn't make it right.
3. implicit assumption: I assume that the average of the minimum and the maximum is the same thing as the average over all projects.
   => It's not, come back when you understand basic statistics.
4. explicit assumption: I will assume that the ratio of indirect to direct costs is between 5:1 and 10:1. For this analysis, I'll take the average: 7.5:1
   => Same thing as above, you don't actually know the number, it could be anything. Plus you make an average on minimum and maximum values which makes no sense at all.

Now the worst part is that Michael Krigsman seems to find the study interesting:

Although not precise, the numbers demonstrate the seriousness of IT failure around the world.

No, they don't ! We don't have a clue how precise they are, which means we don't have a clue how far they are from the truth. All the assumptions are completely wrong, and not just a little.

Michael Krigsman is CEO of Asuret, Inc., a software and consulting company dedicated to reducing software implementation failures.

I propose we make a study on how much money is lost to software and consulting companies dedicated to reducing software implementation failures. Assuming one fifth are incompetent frauds like Krigsman, and the number of projects involving consulting companies is between 20% and 70% (we take the "average" 45%), and making the same du

Re:incompetence

[Nerdposeur]

Also, this happens too often:

Manager: We need to add Feature Y.
Coder: But that builds on Feature X, which is still buggy.
Manager: I don't care. The customer wants it.
---
(A month later)
Coder: Can we take some time to fix the bugs in Feature X and Y?
Manager: No, we have to make Feature Z, which builds on X and Y. We can fix them later.
Coder: If we'd known you wanted Feature Z, we would have done X and Y completely differently.
Manager: Hmmm. Well, it needs to work by next Tuesday.
Coder: (very quiet expletive)

Re:incompetence

[pnewhook]

So I'm not a programmer because I don't accept crappy buggy code? Get real. I've 20 years of real-time mission critical coding experience; I know what constitutes good code and architecture, and what is bad.

Software developers need to learn they have to ARCHITECT code. You cannot just hop in and start writing code without a plan. It would be the same as trying to build a house without blueprints, just nailing up wood at one of the corners.. It will kind of look like a house when done, but will have lots of problems. Code is the same thing.

Simple solution, put it into the cloud

[alen]

and it will magically work the way it's supposed to work

Does your company lose 10% to IT failure?

[BadAnalogyGuy]

The global domestic product is approximately 60 trillion USD. If 6 trillion is lost to IT failure, then on average, every company is losing 10% productivity to IT failures.

This is simply not credible, and this guy should be strung up by his pinkie toes and flogged with ostrich feathers until he admits he eats eggs benedict on Tuesday mornings.

Re:Does your company lose 10% to IT failure?

[postbigbang]

The credulity problem gets worse when one considers how much more productive people have become when using various applications. Yeah, some of them are probably counter-productive, but others (office apps, line-of-business apps) have transformed how we do business for the better. The number seems terribly grandiose even if you push all of the negatives to one side of the equation.

Re:Does your company lose 10% to IT failure?

[Kintanon]

We see this in our clients relatively frequently. Primarily because small to medium sized businesses are some how allergic to backups. No matter how hard we push for them to actually spend money on a backup system that is appropriate to the size of their business a lot of them end up cheaping out on either no backup, or a backup that isn't the right fit for them.
The resulting failure a year or two down the line can cost then a huge piece of their annual revenue.

Other places we see this are when clients try to put their own (Windows) servers in and screw something up that requires the OS to be reinstalled to undo.
In my experience a lot of these "IT Failures" are actually management/client/accounting failures that happen to overlap the IT spectrum. If you can't get the proper budget to do your job, that's an accounting failure that shows up in your area. If management refuses to abide by their own usage guidelines on the network and constantly are passing around infected files that's going to increase your infection rate. And if a client adamantly refuses to change their tapes then when they have a flood in their server room and it gets toasted that's going to translate into longer recovery times, longer down time, and lost revenue.

They should have tried

[vandelais]

turning it off and on again.

Absurd.

[Chicken04GTO]

Crap article with no way to substantiate the actual dollar amounts. How much money would be lost if a large company simply had no IT department at all?

A sad day.

[fuzzyfuzzyfungus]

His failure to title the report "The IT Complexity Crisis: Epic Fail" saddens the hearts of all good men and patriots.

Something's wrong with adblock

[belthize]

There was an ad masquerading as an article by Michael Krigsman the CEO of Asuret, Inc., a software and consulting company dedicated to reducing software implementation failures.

Re:Something's wrong with adblock

[fuzzyfuzzyfungus]

Sounds like somebody needs to hack together a new form of adblocker.

Integrating data from social media networks(get your checkbooks ready and form an orderly line, venture capitalists), particularly those, like linkedin that handle professional affiliations; along with influence peddling/lobbying data from the likes of opensecrets, this tool could automatically grade the trustworthiness and cheap-hackticity of a given article's author, saving you the trouble of manually ignoring the astroturf and marketing fluff.

It'd be a lot trickier, and less precise, than just using regex and blocking known-evil domains; but, in principle, it should actually be possible to use "social media" normally a stalwart friend of subhuman marketing scum, as a source of the information necessary to thwart the same in their vile designs...

Asleep at the Switch

[professorguy]

So let me get this straight... You spend a dollar trying to improve your business process. It doesn't work out. So you're out a dollar. I get it. But then you're out a further $10 because if it HAD worked out, that's what you WOULD have saved. Puh-lease. That's assuming your idea was worth a fuck. NOT ALL IDEAS ARE.

I can easily prove that you personally have lost millions of dollars because there were plenty of things you COULD have done to earn those millions. Why didn't you start a search engine? Why didn't you write the twitter application? Not skilled enough? Heck, you should have bought that winning lottery ticket! And while we're at it, why did you waste your money on fixing your car when it just got wrecked a month later?

My god, you've cost yourself millions of dollars due to your incompetence!

BS Rolls downhil

[Bucc5062]

What nonsense. One of the foundations of project falure is built from the top down. Executive leaderships say "make this work" what ever "this" may be. Top leadership runs around then looking for a solution. Many times they go to a vendor and of course the vendor says "Why yes, our product will solve "this" problem". So instead of so good due diligence on the part of analysts to truly see what the specific needs are, the company purchases this cost saving solution; perhaps it is a service, perhaps it is a soup to nuts enterprise system, perhaps it is off the shelf, out of the box software.

Soon into implementation or pilot the upper levels managers finally begin to see what their own IT staff and their customers were trying to tell them
1 - We don't need "this"
2 - "This" does not fit our needs
3 - "We have to use "this?", the current system works.

Even worse, while the company has a qualified in house staff that understands the specific needs, they will hire consultants to tell them how "this" can work for them. It could be that certain decision makers were favored by the vendor to "try it out" only to find later that the trail cost more in lost time, money, effort while the vendor pockets the dough.

Cynical? Not really. Over my long time in the business I have seen this time and time again. Even though there is a good staff structure in place to handle company IT needs top corporate leaders will buy from a vendor because the perception is that the goal will come quicker. Never mind that the product may not fit, IT will make it fit. Never mind the internal customers that need retraining, we'll hire new people...and on and on. All to try and save time. The bottom line is that any failure of an IT project begins with the top leadership not doing their job. The first question they should ask and answer before dropping a dime is "Do we really really need "this". The second, "Is it an emergency?". The third, "Do we have staff to create "this"?, the fourth "how will this effect our internal customers?. In a world where the attitude is "We need it yesterday" there will be more failure, but do not fault just IT, fault corporate leadership.

(yes I rtfa and it was fluff, stupid and providing no insight to why)

Re:BS Rolls downhil

[gillbates]

with the top leadership not doing their job

As much as I hate to say it, you are dead wrong about this. Many people think the job of top leadership is to make the company run smoothly. It's not. Their job, specifically, is to:

• Improve the overall impression of the company in the mind of the shareholders, i.e. "increasing shareholder value" or stock price. Notice that this activity has nothing to do with the solvency or efficiency of day-to-day operations. It is why companies blow billions of dollars on unworkable "solutions" (i.e. outsourcing) and hare-brained ideas (i.e. Web 2.0...)
• Reflect the bias and preconceptions of the shareholders. Again, this explains why CIO's typically choose vendors and products which have no relevance to day-to-day operations: the shareholders know little to nothing of operations, and when every other company is using SAP or Oracle, you had better make sure you do as well. Even if all of the company data could fit on a floppy disk.
• Finally, most importantly, the job of the CxO is to finish projects. The shareholders and CEO has no clue what the company actually needs, so the CIO must finish *some* project, regardless of how irrelevant it turns out to be. He'll pave the way for later CIO's by installing a ticking-time-bomb of a system which eventually gets so bad that it must be replaced. No matter how badly the project turns out, no matter how useless or counterproductive, the CIO gets paid based on the size and complexity of the project. The only way a CIO can fail is if he is so concerned about "getting it right the first time" and "making a smooth transition" that he wears out the CEO's/shareholder's patience and fails to finish a project in what they consider a reasonable timeframe. It doesn't matter if it is complete junk; the CEO won't ever hear the problems.

I found it much easier to get along in Corporate America once I discovered the quality of the job is less important than the careers of management. Nobody got fired for shipping a buggy product, but people have been fired for not meeting deadlines.

This article is self promotion bate.

[upuv]

I've seen so many of these sorts of articles lately on /.

It's really devaluing /.

It would be nice to have some mod facility to get these nuked. It's disappointing that such a long running resource like /. is now being infected with self promotion. One of the best self promotion FAILS was the one about face book switching to some C++ frame work from php in order to save 10s of thousands of servers resources. I'm still laughing about that one.

A few reasons ( in my opinion)

[KDN]

1. Vendors grossly over selling what they can do. How many times has your company bet on a future product of a vendor that is the best thing since sliced bread and will be available in 3 months, and then 3 months after that, and then 6 months after that, and then a year after that, and then 3 months after that, etc. And when it finally comes out most of the pie in the sky features you were depending on don't really work. But they say it will work in the next version. Real Soon Now.
2. Star Trek style management: Managers who think their crew are Scotty who pulls off a miracle every week. Its never been done, we don't have time to do it right, but its got to work right the first time given not enough resources. Sure it works on Star Trek, its in the script. FYI: I love the Star Trek series, but I also know the difference between fiction and reality.
3. Changing requirements: tell me, who could build a house if you were changing the design every week? One week its a ranch, next week its an apartment building, next week its solar power, next week its wind power, next week it has 5 bathrooms instead of one, next week the bathrooms get moved to different areas of the house, next week the water supply gets moved to the other end of the house. And by the way, we need to cut your budget and move up the deployment date. Doesn't that sound like what happened to Duke Nukem Forever?
4. Big Bang deployments. Designs where a completely new design replaces an old one. No system wide testing (remember the Hubble? The system wide test was deleted to save money.). The old system is torn out, the new system is thrown in, and everything has to work the first time because you can't go back. And there are no facilities for debugging or diagnostics or changes because of course the programmers got everything right the first shot.
5. Ignoring your own staff. Staff does a detailed bakeoff of competing products and chooses the clear winner. Manager goes with the looser because he owns stock in that company. Company deploys product, deployment goes badly, manager blames staff.

Note: these are composite examples from many sources I have gotten over the years. They are not against any one company. But I think they are indicative of the industry as a whole. And that is sad.

Many causes of failure

[ErichTheRed]

"IT failure" is a very broad term and can happen for a lot of reasons:

• Poor requirements definition
• Poor project management (not keeping scope creep in check, not being semi-flexible and delivering a useless piece of software, etc.)
• Execs or IT management desperate to make schedules and forcing release dates
• System failure due to poor planning (no redundancy, poor-quality outsourced hosting, etc.)
• Lack of, or incomplete testing
• Bad code quality

My take on this is that the main cause of failure is the fact that IT still hasn't settled on a set of engineering principles to deliver projects. Things change way too fast still -- over the life of a 2-year project, your hardware platform may be changed out from under you, for example. PHP, .NET or Java may be swapped out for YetAnotherCoolLanguage0.1alpha4. This is made worse by unscrupulous vendors, poorly-trained consultants, and lack of acceptance by the user base of the software.

I think the author is referring to the direct cost of a failure. Every few months, the technical publications run an article or two about a large company or government agency writing off millions of dollars for a failed SAP/Oracle Financials/similar package deployment. Whenever I see one of these, it's interesting to see what happened. Usually it has something to do with one or more of the causes I listed above. Generally, the more expensive, tranformational and long a project is, the worse the results are. It's not just vendors either - I've seen in-house projects spiral down the same way. The other thing that comes to my mind when I read articles like this is why they didn't see it coming. Don't IT executives talk to each other over golf or something and say, "Yeah, SAP screwed us out of $100M in consulting fees. I'd watch them if I were you..."?

Other branches of engineering aren't immune to this though. Construction and infrastructure projects often run over time and budget. The difference is that a construction project gets finished one way or another. A software project failure means throwing away two years of work and putting the hardware on eBay.

Pathetic

[oldhack]

Mgmt blames the devs, and devs blame the mgmt, and both get modded "insightful".

It's like watching hot naked babes wrestling in the mud, except that it's the exact opposite.