Avoid overly optimistic estimates!
"Software types tend to be optimistic, and leaders tend to be in a hurry - a deadly combination that leads to planned schedules and budgets that never had may chance to be realized. - Gary A. Gack People tend to be overoptimistic when asked to make predictions. In the case of software estimation, developers tend to underestimate effort required to complete complex tasks. Overoptimism of estimators is their belief that the work being estimated will run under optimal (best case) conditions. Estimators should base effort prediction on objective project information and avoid an overoptimistic view from management or an unrealistic demand from the customer. Overoptimism can be avoided by a careful, reflective approach to estimation, as opposed to a "quick and dirty" or intuitive one."1
We should define what we want to achieve, what features we want to have and what work we have to do. (= project objectives) Divide the project objectives into many small work tasks. Perform the effort estimate for each work task separately. Take possible risks and task complications into account. Don't forget the test effort and the need for multiple iterations. Tasks priority, order, and dependencies must also be considered. If possible, make estimates based on quantitative, objective information from similar tasks performed in the past. Discuss the workload estimate with assigned people and also with independent experts. Make more estimates considering the worst-case, best-case and most likely scenarios.
At the beginning of a project there is the greatest uncertainty about estimates. There are things that you know are going to happen, things that you know can happen, and things that you never thought could happen. The area of estimation uncertainty diminishes while the project is getting more and more completed. The actual project effort is only known when the project has been completed and all project goals have been achieved. The ratio of the real effort to your estimated effort defines your estimation accuracy and your optimism bias. You should take this ratio (your optimism bias) into account when predicting the next project effort.
How many times has it happened, that you had to postpone a deadline due to an overoptimistic plan?
Reference:
- Adam Trendowicz, Ross Jeffery. (2014). Software Project Effort Estimation, Foundations and Best Pactice Guidelines for Success. Springer.
See more on: www.radim-automation.com
