One of the big risks for a Scrum project is uncertainty in the Product Backlog for scheduling project. One of the well-known tools for controlling uncertainty is the cone of uncertainty. As Margaret Rouse defined, cone of uncertainty is “a graphic depiction of the increasing accuracy that is possible for estimates as the details of a project become more known over time.” The rule of the cone of uncertainty tells us that as time progresses our estimates get better. What we also know from the cone of uncertainty is that in the very beginning of a project our estimates are lousy. They can be wrong by a factor of four. So, to review, at the very beginning of our project if we estimated that the project’s duration would be two years, most likely we are off by a factor of four, meaning that project could be as short as six months or as long as eight years, a pretty difficult range to ask the business to allocate resources, plan marketing events, training events and all the other things that go around the support of launching a product. Knowing that we get better over time what do we do next? The cone of uncertainty teaches us that our estimates get better over time. This is due to more knowledge about the project.
The requirements in the business case, what the actual product is going to do become clearer. If you would like to make the cone of uncertainty even more accurate, iterations and have your subsequent estimations at the end of each iteration even more accurate, you can put a little reality into your estimation, meaning you can start taking some metrics by collecting important data.
For example, let’s look at the following photos.
These are photos from Tehran in 2017 after a big snowstorm when my wife and my kids were visiting for Christmas. Living in Auckland, a city that has never had snow, it was a real shock to wake up the day after Christmas to approximately 22 inches of snow. I just made an estimate. I said there is approximately 25 to 30 cm of snow on the ground. However, is that accurate? I happen to think I’m a pretty good estimator because I watched the news in the morning, and the news said that there were 30cm of snow. However, while arguing over the snowdrifts, my wife said each area is much different, and this seemed to be more than 30cm of snow. I won’t tell you who was right (Although my estimate was correct), but let’s just say the person on the field usually wins (has hard data).
Once you start with a project you still will have to estimate at the end of every sprint.
Take a minute to follow me on Twitter if you enjoyed this post and want to see more.
Most commented