5th Edition PMBOK® Guide—Chapter 6: Comparison of Estimating Techniques


1.  Introduction

The last post covered the time management process 6.5 Estimate Activity Durations, and among the tools & techniques, which are listed below for reference, are the various techniques for estimating activity durations.

6.5 ESTIMATE ACTIVITY DURATIONS
TOOLS & TECHNIQUES
1. Expert judgment Expert judgment can be used by using historical information to give duration estimates from similar projects.  It can also used to reconcile different estimating methods.
2. Analogous estimating Uses a measure from a previous similar project to estimate the duration or cost of the current project in a top-down approach.
3. Parametric estimating Uses an algorithm to estimate the duration or cost of the current project based on historical data from previous similar projects.
4. Three-point estimating A single-point activity duration estimate can be improved by using the most likely, the optimistic (best-case), and pessimistic (worst-case) estimate for each activity.  These three estimates can be combined by using the triangular or beta distribution formulas.
5. Group Decision-Making Techniques Team-based approaches can be useful for improving duration estimates.
6. Reserve analysis Duration estimates can use “contingency reserves” for risks in the risk register that the “known-unknowns” of the project.

 

Estimates made by individual experts use the technique of expert judgment (tool #1), and opposed to this are group decision-making techniques (tool #6), which create estimates in a brainstorming fashion by a group.  This post will contrast the three estimating techniques of analogous (tool #2), parametric (tool #3), and three-point estimating (tool #4).  The remaining reserve analysis technique (tool #6) is particularly useful in conjunction with three-point estimating.

2.  Analogous, parametric, and three-point estimating

The reason why I am creating this post is because the PMBOK® Guide definitions of analogous and parametric estimating both use the words “parameters” and I thought, well, there’s a potential source of confusion right there.  So I am going to compare these definitions below to help the reader distinguish between them.

  Estimating

Technique

Explanation Kind of projects used for
1. Analogous Uses actual duration of previous, similar projects as basis for estimate of duration of current project. Previous activities are similar in fact.
2. Parametric Uses statistical relationship between historical data and other variables. Previous activities in similar in appearance.
3. Three-point Improves single-point estimates by using three estimates:  most likely, optimistic, and pessimistic. Can be used for projects that are new.

Analogous and parametric estimates are examples of top-down estimates, the first being on the basis of the entire project and the second being on the basis of some sort of unit measure (dollars per square foot, hours per line of code, etc.).  They both use historical data, with parametric estimates taking longer, but being more accurate than, analogous estimates.

Three-point estimates are example of bottom-up estimates, which are obtained from analyzing individual work packages.  The single-point estimates are enhanced by using risk analysis to figure out how much each estimate would take in a best-case (optimistic) and worst-case basis (pessimistic).

They are the most accurate of the three techniques above, but they also take the most amount of time to complete.  It is entirely possible that you could use all three techniques in different stages of planning, the analogous and then parametric estimates to get a preliminary estimate, and then the three-point estimates to get a more refined estimate.

3.  Three-point estimates

The three-point estimate relies on three different estimates:

    Estimate Explanation
1. tO Optimistic Based on realistic analysis of the resources likely to be assigned, their availability, and their productivity.
2. tM Most Likely Based on analysis of the best-case scenario.
3. tP Pessimistic Based on analysis of the worst-case scenario

However, they can be combined into the three-point estimate in two different ways:

  Estimation distribution type Expected Duration (tE)
1. Triangular distribution (tO +   tM + tP)/3
2. Beta distribution (PERT technique) (tO + 4tM + tP)/6

The triangular distribution is simply the average or mean of the three separate estimates.  The beta distribution, based on PERT (Program Evaluation and Review Technique), assumes a great deal of confidence that the most likely estimate tM is accurate.  This allows the distribution to give 4 times more weight to tM than to either tO or tP.  The way to remember the denominator of this formula is to realize that there are the equivalent of six terms rather than three in the numerator, if you consider the 4tM to be the equivalent of four terms, tM + tM + tM + tM.

After a brief interlude this weekend, I will continue next week with a summary of the last planning process, 6.6 Develop Schedule, which puts together all of the other five planning processes in the Time Management knowledge area to form the project schedule.

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: