The best approach to estimation is to use more than one technique!

 

A logical question when approaching a new project is “What is the best technique to estimate project effort or costs?”.

Anyone who has taken a foundation course in project management will have been exposed to a large number of estimation methods including analogous, parametric and three-point estimates.

On most projects, particular techniques simply aren’t applicable. For example, on a highly unique project, parametric estimation may not be feasible since there would have been no past history to develop rules of thumb.

However, on most projects more than one estimation method is viable, especially once planning activities are well underway and scope definition and decomposition are substantially complete.

A common choice is to utilize a single, bottom-up estimation method. Occasionally, this bottom-up method is performed using three-point estimates for those activities with which the team has limited experience or confidence.

This is certainly a justifiable approach since the overall project estimates will be derived from the detailed activity-level estimates, but it can also result in these final totals appearing quite bloated, especially if team members have consciously built padding in to their individual estimates.

If such estimates are presented to sponsors or customers, once they get over their shock they may be inclined to arbitrarily cut estimates – this is a lose-lose response which needs to be avoided at all costs.

A safer approach is to sanity-check the estimates before they are presented by using a different estimation method.

For example, combine expert judgment with Delphi method to evaluate the aggregate effort totals by role and assess that against the overall work being done. The relative ratio of effort estimated to be spent between roles may help to identify variances to focus on – for example, if development estimates are more than triple testing estimates for a brand-new product, one or the other may need further review. Assumptions analysis may also help to reveal specific estimates which need some further refinement.

Although the effort spent on estimation like with any other project management practices needs to be commensurate with the scale and complexity of a given project, the benefits of using multiple estimation methods will provide more than one source of validation while still retaining team member commitment to the work.

(Note: this article was originally written and published by me in September 2013 on my personal blog, kbondale.wordpress.com)