Comments for Critical Lesson in Estimation at http://www.stevenmsmith.com , comment 1 to 4 out of 4 comments http://www.stevenmsmith.com Fri, 10 Oct 2008 18:34:32 +0100 FeedCreator 1.7.2 http://www.stevenmsmith.com/my-blogs/management/critical-lesson-in-estimation.html#comment-73 People who will do the work should do the estimation as well. - Esther Estimate Sun, 20 Apr 2008 11:35:24 +0100 http://www.stevenmsmith.com/my-blogs/management/critical-lesson-in-estimation.html#comment-16 Arnon, I looked at your website, which I now wish I would have looked at before my first reply. I see you are very experienced. I suspect you have experience about happens when inexperienced people make estimates. My experience is that inexperienced people do their best when making estimates. Unfortunately, they tend to underestimate. I don't think anyone should be surprised. Warm regards, -Steve - Steve Smith Tue, 26 Jun 2007 13:28:38 +0100 http://www.stevenmsmith.com/my-blogs/management/critical-lesson-in-estimation.html#comment-15 Hi Arnon, The forecast from an inexperienced IT person are typically worse than the forecasts of an inexperienced meteorologist. Why? The forecasts of meteorologists are grounded in probability. They talk about the chance of rain rather than saying it will rain tomorrow. Inexperienced IT people typically use point estimates, such as, "We will deliver on August 2nd," which prevents their customers from seeing the inherent variability of an estimate. Meteorologists are also trained on how to forecast while IT people, in my experience, typically receive little training on estimation. And if they do receive training, their company's culture may prevent them from practicing what they learn. Meteorologists also have data and models to aid their predictions. IT people, in my experience, rarely have data about previous projects and typically have zero models for making estimates. I believe estimation is an art as well as a science. People need experience in order to make solid estimates. If inexperienced people must make the estimate, I suggest avoiding point estimates entirely. Like a meteorologist, create a probability based forecast. I like a graph with the x-axis being time and the cumulative probability of delivery on the y-axis. The graph is dynamic rather than static. As you learn more, the probability change. I've notice that many groups are afraid to share their uncertainty with their customers until they must. I don't recommend that approach. But if, for whatever reason, you must hide the estimate's variability from your customer, that doesn't stop you from making it visible internally. You learn to be a better estimator by practicing. If you hide the variability from yourself, it's poor practice. Thank you for the question. And I wish you and your group success with your efforts. Best regards, -Steve - Steve Smith Tue, 26 Jun 2007 07:52:20 +0100 http://www.stevenmsmith.com/my-blogs/management/critical-lesson-in-estimation.html#comment-14 What happens when the people who would actually have to perform the task are those inexperienced folks - what does that mean for the educated estimates?... - Arnon Rotem-Gal-Oz Tue, 26 Jun 2007 02:17:04 +0100