Fundamentals of Range Estimating
By William R. Duncan
Few project management terms are as misunderstood and misused as “estimate.” In the following paragraphs, we address some common misconceptions about this vital subject.
What exactly is an estimate?
A project estimate is an informed assessment of the likely project cost or duration. Informed means that you have an identified basis for the estimate. Likely focuses on the inherent uncertainty of the estimate: every estimate is but one of many possible outcomes. Cost or duration refers to the two major categories of estimates.
What should I use as the basis for my estimates?
The first step is to define the project scope. In the absence of a defined scope, you should not estimate anything other than the cost of defining the scope. Once you have the scope defined, you can rely upon the fact that similar scopes tend to have similar results, so most of the time you will use historical information as the basis for your estimate.
Where can I find historical information?
Historical information about previous projects should be available from one or more of the following sources:
- Project files. Your employer should maintain records of previous project results that are detailed enough to aid in developing estimates. If not, you must do so yourself.
Project team knowledge. The individual members of the project team may remember previous actuals or estimates. Such recollections can be useful, but they are generally far less reliable than documented results in the official project files.
Commercial databases. You may be fortunate enough to be involved in an application area such as construction or software development where you can purchase historical information.
But if every project is unique, what good is historical information?
The product of the project is unique, but most of the activities are not. By focusing on the management similarities rather than on the technical differences, you will usually find that you can generate valuable insights even when the product of the project is wholly different. We have seen project managers in fields as disparate as construction and biopharmaceuticals exchange estimating insights.
If there are “many possible outcomes,” how can I tell which one is likely?
You should not pick just one; you should identify the full range of possible outcomes. The figure below illustrates the problem with single point estimates: A and B are equally likely, but there is a far greater chance that A will be exceeded! In addition, single point estimates tend to become self-fulfilling prophecies. Both the project manager and the person responsible for the item being estimated need to know what the range is.
How do I know what the distribution looks like?
For the most part, the shape of the distribution isn’t nearly as important as recognizing that there is a range of possible results. In the late 1950s, the US Navy developed the Program Evaluation and Review Technique (PERT) using a beta distribution. In the absence of evidence to the contrary, we generally assume a triangular distribution because a triangular distribution produces more conservation results than a beta.
Seems to me like developing range estimates will take a lot of time…
Our experience is that it often takes less time by eliminating the posturing and gamesmanship that often accompanies the estimating process. We have had teams in our courses spend less than an hour to develop an estimate for a 4,000 hour project — an estimate that later proved highly accurate. A recent client spent less than a day to discover that a mission-critical, multi-million dollar new product development project was going to be eight months late. An investment of something less than $10,000 is expected to return in excess of $10,000,000. Time spent estimating is not a cost, it is cheap insurance.
What do I do when management cuts my estimate in half?
The first thing you have to do is find out why they are cutting. Most such management actions are created by one of the following misunderstandings:
- Senior managers often fail to distinguish between an estimate and a price. A price is the monetary value charged for a product or service. (This is true even for an internal project where no money actually changes hands.) The determination of price is a business decision that is related to, but separate from the development of an estimate. In this regard, you have an important responsibility as project manager: to ensure that the estimate is accurate in order to facilitate a good pricing decision.
Your manager may view your estimate as a negotiating position. Some managers assume that you have “padded” your estimate to provide negotiating room. If their assumption is true, then their cuts are appropriate in order eliminate the padding. If your estimate is truly an estimate (an informed assessment), you should be able to demonstrate — clearly and unequivocally — that the suggested cuts will either reduce quality or simply increase the probability of an overrun.
Others may use the estimate as a motivational tool. Most good managers understand that a more difficult task (where the target is aggressive) can serve to motivate the project team. But even good managers often fail to appreciate that an impossible task (where the target is so aggressive that it is clearly impossible) is a powerful demotivator.
OK, so I do all this work and come up with a range estimate for each activity on my project: how do I add up the ranges to get a project estimate?
The formulas for adding up ranges can be found in any basic statistics text.
I tried range estimates once before, but my project management software only has room for one estimate!
This is actually a variation on the price vs. estimate discussion above. After developing a range estimate, you still need to select a number from within the range to use for tracking and control—this is the number you should use in your project management software. Normally, you will use the mean of the distribution for this purpose. If we continue to assume a triangular distribution, the mean (and hence your budget) is nothing more than the arithmetic average of the three numbers that define the range.
Is that all there is to it?
Those are the fundamentals. More advanced topics include subjects such as who does the estimate and when, differences between cost or effort estimates and duration estimates, correlation, project level estimates vs. activity level estimates, and using range estimates as a risk management tool.
William R. Duncan is the principal of Project Management Partners of Lexington, MA USA. He currently chairs the Board of PMCert, the certification body of the American Society for Advancement of Project Management (asapm). He was the primary author of the original (1996) version of A Guide to the Project Management Body of Knowledge and was one of the founding members of the Global Alliance for Project Performance Standards (GAPPS) which has recently published a framework for performance-based competency standards for project managers.
© 2009 William R. Duncan – http://www.pmpartners.com/