Scheduling tools are only as good as the activity duration and cost estimates entered into them. Accurate estimating of these elements is therefore imperative for creating a schedule that is realistic.
This article discusses the principles and approaches to aid the project manager’s attempt to accurately estimate the duration and/or cost of a project.
One of the major challenges of scheduling projects is coming up with accurate estimates for the schedule. This is such an important and difficult component of scheduling that whole jobs have been created around this function. Estimators, as they are titled, work eight hours a day and forty hours a week estimating the duration and/or cost of projects. Estimators attempt to apply scientific approaches to a skill that is primarily an art form. One useful estimator resource, used in the engineering and construction industries, is the RS Means book.
It provides valuable and detailed labor, equipment, and material estimates for construction projects. The RS Means construction estimates are based on historical data. They provide the estimator with the closest approach to a scientific estimating method. However, the RS Means data is limited to the construction industry. Estimators in fields like the research and development industry do not have this tabulated cost data. And will have to rely on less scientific approaches to estimating.
The difficulty in estimating is inherent in the characteristics of projects. Projects by definition are unique. They are not repetitive processes. Therefore, it is difficult to find similar historical projects for analogous estimating efforts. There will always be something distinct about a project to distinguish it or separate it. Despite this set back analogous estimating is one of the best approaches to project estimating.
When analogous estimating is not a viable option you may want to consider parametric estimating that examines relationships between variables on an activity to calculate time and/or cost estimates. In parametric estimating you determine the unit duration or cost and scale this value for the number of units in the project. It’s important for the measurement to be scalable for accuracy. For example if it took two days to code 500 lines of computer code it will take eight days to code 2000 lines of computer code. Parametric estimating is another approach that attempts bring a scientific method to what is primarily an art form.
Many times the most accurate way to estimate a project is to procure expert judgement. This is someone with direct knowledge of the field in question. Who would be better to provide this knowledge than the people actually doing the work? Yes, resources performing the work always know best how to estimate their work, provided they have some related experience.
There is also another benefit to obtaining team member estimates: they take ownership of the estimate. If the estimate comes from upper level management or perhaps, a sales representative, the team members performing the work do not have ownership of the estimate, and are less likely to produce the effort required to achieve the estimate goals. Ownership matters! It has a significant impact on team member productivity.
So we obtain our estimates from team members performing the work. The question remains, what kind of estimate do we get from team members? Do we ask them for a pessimistic estimate? This will provide the most conservative estimate. For that matter is a one-point or single-point estimate sufficient? Or would a three-point estimate considering the pessimistic (P), most likely (M), and optimistic (O) outcome be better. With the three-point estimate we introduce some statistics to the estimate that reduces some of the subjective nature of the estimate. The formula for a three-point estimate is as follows:
This equation is a very simplistic average. The Program Evaluation and Review Technique, developed by the U.S. Navy in the 1950s, employs a weighted average. It weights the average to the most likely estimate, which should be the most accurate. The weighted average duration or cost estimate is below:
Thus, a little bit of statistics is added to the estimate that provides a more scientific and structured approach. Also, studies have also shown that team members that have difficulty providing a one-point estimate are more readily able to provide a three-point estimate. A three-point weighted average estimate is the best expert knowledge approach. You can even provide a standard deviation to the estimate as follows:
The simple average, weighted average, and standard deviation can be computed in a spreadsheet and then the duration and/or cost estimates can be transferred to the schedule.
Summary
Creating a believable schedule that can be performed in the real world requires accurate duration and/or cost estimates. Because projects are not operational processes but unique temporary endeavors estimating schedule durations and costs from similar or analogous past projects is not always feasible. Consider the scientific parametric approach when projects are dissimilar, but variables from one project to another are scalable.
Another recommended approach is expert or team member three-point judgement on durations and/or costs inserted in a weighted average equation estimate. Estimators can also provide an initial top-down estimate on deliverables, and then follow up with a more detailed and accurate bottom-up estimate starting with the individual tasks (activities).
References:
Rita Mulcahy’s PMP Exam Prep, Eighth Edition, Rita Mulcahy, PMP
Forecast Scheduling with Microsoft Project 2010, Eric Uyttewaal, PMP