Estimation Models

I was recently tasked with creating a model to help determine duration for a particular kind of work. Working through making the model a reality I began to appreciate its value.

A targeted estimation model is a good tool to try and understand the impact of specific factors on the work at hand; also a useful tool for what-if analysis. It can be particularly handy in helping stakeholders understand the various factors considered and their role in arriving at the predicted duration of work.

The more specific a model, the more useful it is likely to be. One-size-fits-all models result in extremely iffy results that often fail to account for the unique factors and constraints of particular kinds of work.

Useful as models are, it's important to strike a note of caution about them. Data emerging from a model is, in the end, a prediction (guess) and NOT the gospel truth. If the initial understanding of the models factors and constraints turns out to be close to the way they play out in reality, data emerging from the model will serve as a useful guideline. At no point should the data emerging from a model be considered absolute.