The importance of estimates was reinforced recently by Ross Gittin's article on Defence force spending (basically how a lack of planning and a failure of governance means that those who initiate projects are not responsible for the inevitable budget blowout or maintenance nightmare that their decisions create). In short the majority of estimates were unreliable, yet billion dollar decisions are made on these estimates.
IT estimates are notoriously unreliable and would anyone bet the mortgage on any particular IT estimate no matter how good the track record of the person making the estimate?
Yet, we could not plan any IT project or Autralian defence force project without estimates.
Perhaps we need a Kinteic Theory of Projects similar to the Kinetic Theory of Gases. (Which says that that we can have reliable estimates based on the total behaviour of the system, and measuring the behaviour of individuals that constitute the system does not tell us anything particular meaningful. But we can have useful statistics such as temperature which is a statistic of the whole system). So the Kinteic Theory of Projects would say that looking at the estimates of a particular project is not particularly useful (except possibly for political purposes). What we need to do is recognise that we need a statistic akin to temperature -- what is the trend for estimates.
The problem with the defence force is that the excuse that a particular project can run into trouble is a safety net for mismanagment. Once is unlucky, twice is coincidence, three times is enemy action.
Having said that I once worked with two very different people who provided estimates and both were very good at it. One had more of a charming personality than the other and this person would look at a project, estimate it would probably take 95 units of work and then quote 100 units, explaing that unknowns had to be taken into account and we would do our best regardless. The other would also estimate 95 units and then explain to the project owners that if everything went OK we may be able to get away with 90 units but all we could do was try. Invariably things went to plan and 95 units were needed. The person who had dealt with my charming colleague was usually happy and relieved that they saved 5 units. The person who dealt with the other was usually cheesed off that they did not save the 5 units.
Very good reading. Peace until next time.
WaltDe