A few weeks ago I ran one of my project management courses. Part of the week is about estimating. One approach I cover is the Delphi Technique. (See last month’s post on what this is) Usually I just talk about it in theory, but this time round I’d decided, with some trepidation, to see if I could run an exercise to prove it worked. Even I was amazed at how accurate an estimating tool it was.
I’d seen other demonstrations of the Delphi Technique. Usually they involve predicting something that isn’t too hard to work out, like the average weight of a group of people. I wanted to predict something difficult, something that unless you were an expert in that field, you would probably have little idea what the answer was. I decided on the distance between the earth and moon. Personally, before I looked up the answer, I would have put it anywhere in the region of 50,000 miles to 5 million miles.
I firstly asked the group if anyone knew the answer. Only one person out of twenty did. He was excluded. Everyone else wrote their best guess with a one-sentence justification on a post-it which I collected. I then drew a long horizontal line on a white board, scaled roughly in miles and started sticking each post-it above the relevant place on the scaling line. The first estimates ranged from around 6,000 miles to one million light years! I read out all the justifications to the group. Some were rational like “It took roughly three days for the Apollo mission to get to the moon travelling at around 5,000 miles per hour so distance must be around 300,000 miles” Others were less logical such as “I think Mr Spock said this on Star Trek!”
We then repeated the process. The range of estimates tightened from around 60,000 miles to 800,000 miles. Most where in the same order of magnitude as the previous round’s more logical figures. People started building on the previous round’s more logical justifications or began to think of new ways of working out the figure. For example one person remembered some speed and duration figures from the Apollo 13 film and another remembered the diameter of the earth and used it to pro-rate out an estimate.
We did three rounds altogether. I was quite astounded with the result. I would have been happy with something in the right order of magnitude, but the average estimate from the group was 225,000 miles. The actual average distance is 238,000 miles. So the Delphi Techniques was accurate within 5%. Not bad!
What do you think? Have you ever tried the Delphi technique in a work situation to estimate figures like project times and costs? How accurate was it? I’d be interested to hear your comments.
Trackback URL for this post: http://www.orgtopia.com/2009/11/19/estimating-the-distance-to-the-moon-using-the-delphi-technique/trackback/
{ 4 comments… read them below or add one }
I was part of the group participated in the experiment. Beside the fun we had while doing it, it was an excellent lesson for the future. I’m definitely going to try it in my next project
Thanks Shally – Glad you enjoyed the experiment!
David
I was also there, and while initially sceptical, I was totally won over and am looking forward to the next big project where I can try this out – great to have a documented example of this working in action!
I think this is the part of the course I remember best. It was a solid introduction to the Delphi technique – one that is tough to forget (especially given that one sticky note was somewhere near Camden Market!). Fantastic tool and an excellent way to achieve a consensus.