The failure rate of our IT projects is insane. So why do we accept it?
Having had the benefit of a long — and somewhat unusual career — observing enterprise IT from pretty much every conceivable angle, I’ve begun to have certain realizations. It seems that after watching the same story play out two to three hundred times in companies of all shapes and sizes, even I start to detect a pattern.
I find one pattern particularly troubling. When it comes to executing enterprise IT projects, I fear we have a problem.
In my experience almost everyone in our industry has, at least some inkling, that this problem exists. But as with other unpleasantries, we don’t like to talk about it. It makes us uncomfortable.
In our defense, even if we were to get past our aversion to the awkward, we still have a problem with this problem. We don’t know how to describe it.
We’re not really sure what’s causing it. We know our projects don’t work out as well as they should. We know that this situation has endured for decades. We know that others have the same problem. But we don’t know what to do about it.
By we, I’m referring to those of us who charter, manage, work on, are affiliated with, use, are impacted by, pay for, sell, or implement enterprise IT projects and any stockholder in any business that does. If you fall within that spectrum, I’m afraid that we, includes you. And the problem I’m referring to is our abysmal record of being unable to execute these projects successfully.
There are several aspects to this story. Over the course of the next few weeks — and perhaps months — I intend to address many of them. However, for the immediate future — meaning the next week or so — I’d like to start by framing the issue from three perspectives: frequency, cost and cause.
Today’s topic will be the frequency of enterprise IT project failures and what the term “failure” actually means.
If you listen to the luminaries of the IT universe (Gartner, IDC, Standish, HBR, Capterra, et al), you get fairly consistent estimates of project failure rates. These authorities report that using even the most forgiving criteria, only twenty-nine percent (29%) of enterprise IT projects are successful, i.e., everything promised is delivered, on time, on budget. For the sake of this and future discussions, I’ll be generous and round up to thirty percent (30%).
I will also note two things. First, seventy percent (70%) of all IT projects — seven out of ten — fail. Second, these failure rates have remained consistent over many years.
Now, in the interest of fairness, seventy percent (70%) isn’t quite as bad as it appears. Not all failures are created equal.
Fifty percent (50%), one half, of IT projects just experience some significant, or in accounting parlance, some material level of failure. In other words, they’re way (30–300%) over budget, way behind schedule or way under-deliver. The system, or a very big piece thereof, never does what it was supposed to do or what the user community wished it did. It never lives up to the promise and hype.
Only twenty percent (20%), one out of five, are complete and utter disasters. Total failures. Smoking craters that leave management no recourse but to write everything off, and participants no recourse but to find other jobs.
It gets better!
If a project costs over one million dollars, the added complexity causes the odds of failure to jump by fifty percent (50%). That means that the odds of a total failure (smoking crater) for larger projects are thirty percent (30%) — roughly one out of three. The odds of a significant (or material) failure rise to about sixty-five percent (65%) — almost two out of three.
The odds of a true success for an enterprise IT project in excess of one million dollars — one that fully delivers as promised, on-time and on-budget — fall to about five percent (5%). That is only one success out of twenty (20) attempts. Would you get on a plane that had only a 1 in 20 chance of landing you safely at your destination?
McKinsey & Company working in conjunction with Oxford University found that seventeen percent (17%) of such large enterprise IT projects, one out of six, go so badly, they threaten the very existence of the organization.
Let’s put this situation into graphic context.
Playing Russian Roulette with five bullets in the gun (five out of six) offers better odds of success than running a large IT project. And the risks of even undertaking such a project are equivalent to playing a “regular” game of Russian Roulette (only one out of six bullets in the revolver) with the life of your business.
Maybe I’m just risk adverse, but in my humble opinion you’d have to be insane to play Russian Roulette under any circumstances. The idea of playing that game with five out of six bullets in the gun is pure idiocy. And yet, we do it every day.
I have to ask why?
Perhaps we allow this condition to continue because we know that avoiding technological innovation altogether guarantees death by other means — the market and competition slowly eat you alive. Maybe we don’t think we have a choice. Maybe we don’t care.
Whatever the reason, whether we realize it or not, we are making a choice to allow this situation to continue. I, for one, don’t think our current choice is a good one. I think there are better options than perpetuating the status quo. Over the coming weeks I hope to introduce a few.
I’ll be back soon with more.