There is an art and a science to estimating that you should be familiar with for the PMP Certification Exam. The art comes from the expert judgment that team members and estimators bring to the process. Their experience from past projects is valuable in developing estimates, determining the best method to use, and being able to look at an estimate or the assumptions behind the estimate to assess the validity.

In addition to the artful contribution of experts, team members, and estimators, there are a number of methods that make up the science of estimating.

## Analogous estimation

*Analogous estimating* is the most common method of estimating. The aforementioned experts normally conduct this form of estimating. In its most basic form, analogous estimating compares past projects with the current project, determines the areas of similarity and the areas of difference, and then develops an estimate based on that.

A more robust application determines the duration drivers and analyzes the relationship between past similar projects with the current project. This can include size, complexity, risk, and number of resources, weight, or whatever other aspects of the project influence duration.

To use this method effectively, projects must be similar in fact, not just in appearance.

Analogous estimates are most commonly used early in the project at a high level. There is an expectation that there will be a range of estimates and that the range will be progressively elaborated into more detail as information is uncovered.

The benefits of using analogous estimates are that they are relatively quick to develop, and they’re not very costly to develop. However, because they’re usually done at a high level, they’re not the most accurate method.

Say you need to develop training materials for the childcare center staff. The training will take two days. You developed a similar training program a few months ago, but that program was three days long. Therefore, you reduce your estimate by 33%. However, this new training program is relatively complex compared with the earlier one; it includes first aid, food prep, and safety training.

Therefore, you increase the duration by 10% for a complexity factor. The first training class required four hands-on demonstrations, but this one needs only three. Given this information, you can access the prior project to find the durations and modify them based on current information.

Last Class | Effort | This Class | Modifier | Modified Estimate |
---|---|---|---|---|

3 days | 200 hours | 2 days | –33% | 133 |

Easy | Included | Complex | +10% | 20 |

4 demonstrations | 40 hours | 3 demonstrations | –25% | 30 |

Total | 240 hours | 183 hours |

## Parametric estimating

*Parametric estimating* uses a mathematical model to determine durations. Not all work can be estimated this way, but when you can, it’s fast and easy. You multiply the quantity of work by the number of hours it takes to accomplish it.

For example, if a painter can paint 100 square feet per hour and you have 6,000 square feet to paint, you can assume 60 hours of effort. And if you have three people painting (60 / 3), then it should take 20 hours, or the equivalent of 2.5 days.

## Three-point estimates

When a lot of uncertainty, risk, or unknowns surround an activity or a work package, you can use *three-point estimating* to give you a range and an expected duration. You collect three estimates based on the following scenarios.

**Best case:**The best case scenario means that you have all your required resources, nothing goes wrong, everything works the first time, and so on. This is represented as an*o,*for “optimistic.” Or you might see it represented as**t**_{o}, for “time optimistic.”**Most likely:**The most likely scenario takes into account the realities of project life, such as someone being called away for an extended period, work interruptions, things not going exactly as planned, and so forth. This is represented as an*m,*for “most likely.” You might see it represented as**t**_{m}, for “time most likely.”**Worst case:**The worst case (pessimistic) estimate assumes unskilled resources, or not enough resources, much rework, and delays in work getting accomplished. This is represented as a*p,*or**t**_{p}, for “time pessimistic.”

The simplest way to develop the expected duration, or **t**_{e }(“time expected”), is to sum the three estimates and divide by three. You might see this referred to as a *triangular distribution.* However, this isn’t the most accurate way because it assumes an equal probability that the best case, most likely, and worst case scenarios would occur — and that’s not realistic.

In reality, the most likely estimate has a greater chance of occurring than either the best case or worst case scenario. Therefore, weight the most likely scenario and take a weighted average. The most common way of calculating a weighted average is

**t**_{e} = (**t**_{o} + 4**t**_{m} + **t**_{p}) / 6

The PMP exam might refer to this as a Beta distribution or *PERT estimating.* (PERT stands for Program Evaluation and Review Technique.)

## Reserve analysis

*Reserve analysis* is used for work packages, control accounts, or any level of the WBS. It looks at the complexity and riskiness of the duration estimate and determines whether reserve is needed to account for the uncertainty, risk, or complexity. Sometimes, you will see reserve referred to as *contingency reserve* or *buffer.*

You might see reserve as a percentage of time based on the phase of the project or as a set number of hours. It’s a good practice to revisit the reserve and see how much you have used compared with how far along you are in the project. When reserve is used, the reason for it should be documented so you can include the information in “lessons learned.”

Reserve is not padding. Pad is bad. Reserve is developed based on analysis and thoughtful deliberation. Padding is throwing extra time in the schedule because you didn’t take time to do the work to develop a well thought-out estimate.