Erik Bernhardsson on project delay

Erik Bernhardsson wrote an interesting article on better ways to estimate how long a (software) project is going to take. (“Why software projects take longer than you think: a statistical model”)

He states that project duration estimates are typically rather well calibrated for the median, but not for the mean. The reason is the slower half of projects take (much) more time to complete than the faster half take less time. He discusses a log-normal distribution of the “blowup factor” (actual time divided by estimated time) as a reasonable model. That may all sound very cryptic and difficult but he explains everything very well, go have a read.

Leave a Reply