Estimations for the effort involved in work to be done are poor or absent
Estimations for the effort involved in work to be done are poor or absent.
It should be accepted that there can only be low accuracy in early estimates, increasing as more becomes known about the project. The earliest point at which truly accurate estimates might be possible is after the architecture has been decided and proven. Yet even at this stage, the need for accurate estimates for all but the most immediate work (for the current iteration and possibly for the next iteration) should be questioned. Estimation is an art; some people are very good at it, while others manage only with difficulty. A process that works well for one person may be totally useless for another, so it might not be productive to define a process by which estimation should be done. There are a few rules of thumb to improve estimation abilities. Do not penalize people for getting estimates wrong, or you will be motivating them to give high estimates and then ensure they become accurate. For a similar reason, rewards for accuracy should not be very lavish, but sufficient to motivate people to make the attempt to improve their abilities. Give feedback on how close the estimates were to the actual measured effort, and be prepared to give a breakdown of what was included, in case the estimators need a feel for where effort is incurred.