Flyvbjerg: IT projects lack sufficient risk assessment
According to VKR Professor and Head of the Danish Institute for IT Program Management at the IT University, Bent Flyvbjerg, cost overruns on large-scale IT projects can be attributed to faulty risk assessment. In a new scientific article, the researcher and his colleagues present a power law to combat the problem.
The entire organisation suffers when large-scale IT projects incur unexpected cost overruns. In worst case scenarios, as was the case for American retailer Kmart and the British automobile glass company Auto Windscreens, a company can go bankrupt as the result of a catastrophic IT project. On average, 45 percent of all large-scale IT projects experience cost overruns and in many instances to the tune of more than 200 percent. So, why do large-scale IT projects blow up?
According to a new research article by VKR Professor at ITU, Bent Flyvbjerg, titled The Empirical Reality of IT Project Cost Overruns: Discovering A Power-Law Distribution, published in Journal of Management Information Systems, current IT project risk assessment practices are severely flawed. When assessing the risks involved in realizing IT projects, project managers cannot rely on a statistically limited list of possible problems that may arise. According to the professor, IT projects are subject to infinite variance. This means the list of possible risks to a project’s successful completion on budget is literally endless.
The fact that IT projects have infinite variance and are subject to an endless list of potential risks determines how they should be risk assessed. According to Bent Flyvbjerg, the problem is that risk assessment of IT projects is typically based on an assumption that IT projects follow a normal distribution plot – the so-called bell-curve – which labels extraordinary events as rare deviations from the norm.
However, in reality IT projects follow a power law that reflects the unpredictable nature of large-scale IT projects.
“In reality, we’re talking about the exact opposite of the statistical concept of regression to the mean. We’re talking about regression to the tail. No matter how great the magnitude of cost overruns on a given project are, odds are even greater cost overruns will occur in a future project,” says Bent Flyvbjerg.
“We need to dispel the illusion of adequate predictions. IT projects are impossible to predict. Instead, IT project managers should focus their efforts on prevention.”
According to Bent Flyvbjerg’s research project managers need to study and pay close attention to the projects that have failed miserably and resulted in cost overruns and loss of revenue. In other words, project managers must take heed of black swans – the unpredictable events that derail IT projects.
“As a project manager, it is of utmost importance to learn from mistakes – your own as well as mistakes of others. You have to find out where things went wrong and what prevented successful completion of a project,” says Bent Flyvbjerg. “But the most important lesson is to forget everything about predictions. You cannot make predictions when you are operating with infinite variance and risk.” Theis Duelund Jensen, Press Officer, tel: 2555 0447, email: email@example.com