Monday, February 7, 2011

People always finish projects behind schedule and over budget - but here is some help

The United Kingdom has a proud tradition of delivering projects way behind schedule and way over budget. For example, the new Wembley stadium in London, which opened in 2007, was originally scheduled to open in 2003. It was also about $450 million over budget. Similarly, the Jubilee Line extension to the London Underground system cost $5.3 billion instead of the estimated $3.2 billion, and was almost two years late; likewise for the prestigious Millennium Bridge covering the Thames; Sadler’s Wells theatre in Islington; the list goes on and on.
But at least it is not as bad as the (in)famous mother of all planning disasters: The Sydney Opera House. This was scheduled to open in 1963 at a cost of $7 million. Eventually, it opened in 1973 and cost $102 million. But I guess that’s simply because all Australians are really British descendants with a gene for criminality.

Anyway… of course there is nothing British about all this. The same happens in many countries, and to most individuals. We all tend to be rather unrealistic in our project planning and for instance consistently underestimate our completion times. Strangely enough, research shows that we only display this so-called “planning fallacy” for our own work. When it comes to estimating the completion time of someone else’s project, we are actually quite accurate. That is why we all snigger at our friends’ project plannings, declaring “they will never make that”, to subsequently make the exact same planning errors for our own work.

Aiding accuracy?

Professors Roger Buehler from Simon Fraser University and colleagues designed a series of experiments to see if he could help people to improve their planning ability. He enlisted a bunch of psychology students and asked them to estimate their completion time as accurately as possible for a particular thesis. At the end, he checked how many of them had been late. Perhaps not surprisingly (if you are also human) more than 70 percent were late. He then ran an experiment in which he asked people to make two plannings: one “if everything went as well as it possibly could” and one “if everything went as poorly as possibly could”, to see if that would make them a bit more realistic. Unfortunately, it didn’t. The majority of people still were way late, even in comparison to their most pessimistic forecast!


Roger then designed another study. He thought, “perhaps giving them a deadline will help?” He asked a large number of respondents to think of a project they were going to undertake in the near future – the projects ranged from all sorts of academic assignments to fixing one’s bicycle or clean their apartment. Then he checked whether there was a real deadline for their project (which was true for about half of them) and asked them for their estimate of their completion time. At the end, he checked how many of them had finished their project before their estimated time of completion. As before, about two thirds of people had been overly optimistic, and finished their project late, regardless of the nature of the project. And there was no difference between the groups with or without a deadline. Apparently, setting a deadline does not help a single bit; we do not become more accurate in our planning forecasts.

Recalling past experience

Then Roger thought of yet another experiment. He had noticed that when people explained the logic for their planning, they almost always only referred to what might happen in the future, during their project; they never seemed to think about relevant past experiences with similar projects, where surely they had experienced that they tended to be overly optimistic and usually finished late. So he thought, “perhaps that might help; asking them to think about similar past experiences before making their project planning?” And so he did. He ran a similar experiment but first asked people to remember their relevant past experiences. Didn’t help a single bit… His bemused respondents first told him at length about their past planning errors and then, during the subsequent minute, made the exact same planning errors all over again.

Then Roger took a deep breath, and designed one final experiment. Now he did not only ask participants to recall their past experiences before making their project planning but now he also asked them explicitly to actually incorporate these past experiences in their planning. And – beware and behold – that helped. When Roger explicitly forced people to connect their past experiences to their new planning, they finally came up with some quite realistic forecasts.
Hence, the planning fallacy is a pervasive and persistent phenomenon that is not easy to tackle. We are quite good at forecasting someone else’s project, but we are stubbornly overoptimistic about our own. The only way to make us see some sense is to be grabbed by the ear by someone like Roger, and be forced to explicitly take into account the past errors of our ways. Then there is no denying it anymore, not even to ourselves.