**Estimating Duration of Activities**

Once we have identified the activities and their sequences, in order to put calendar dates we would need to know their durations. Activity duration depends on two things – effort and resources. Effort is the “man days” or “man hours” required to complete an activity. It is the amount of effort that needs to be put on an activity. After estimating the effort the duration can be computed based on the number of resources working on the activity. For example, if the effort is 8 man hours then the duration would be 1 day if one resource works on the activity. It will be about 0.5 day if two resources work on it and so on. Below is the formula for calculating the duration:

Effort = Duration x Number of Resources

Or, stated in a different way, Duration = Effort/Number of Resources

There are several ways of estimating the effort of an activity. Each of these is listed below.

**Parametric Estimate**

If an activity is about painting a room then one of the most common estimates is to use a thumb rule of how much time it takes to paint one square foot of wall. Next we take the total measurement of all the walls to be painted. We get the final estimate by multiplying the thumb rule with the total square feet. This is called a Parametric Estimate and is generally the most accurate because the thumb rule is made using data from past projects and expert judgement. Another example would be when trying to estimate the time taken to excavate (dig) for building construction. Thumb rule for excavating one cubic meter would be used and depending upon the total volume of the excavation the final figure can be arrived at. In software development projects we regularly use Function Point (FP) based estimation. In this technique we calculate the number of function points and use a standard productivity chart to arrive at the final “man days” estimate of activities. This is also a parametric estimate.

Parametric Estimate = Thumb Rule per unit x Number of units

**One-time Estimate**

There are times when the project manager asks the concerned team member or a subject matter expert to estimate the time required to do an activity. The team member would give a figure based on expert judgement and this estimate would generally contain a buffer. For example, if the project manger asks the team member to estimate the duration for painting a wall, based on the person’s expertise he believes that the work can be completed in 5 days but would rather say 6 days just in case something goes wrong. This buffer is not known to the project manager. Hence, this kind of estimate could be filled with buffers for each activity and all buffers would add up to give a longer than actual schedule. This would also inflate the costs as they depend on the efforts. The situation is worse if there are multiple levels of hierarchy involved in time estimation. If in the above case the team leader asks the team member and gets an estimate of 5 days, he would probably report 6 days as estimate to the project manager as he is unaware of the buffer already been kept by the team member. These buffers add up to create a very inflated schedule. It is not a good project management practice to have buffers. Instead, there should be reserves at the end of the schedule. Reserves are known to everybody and kept for the entire project based on project risks.

**PERT (Program Evaluation and Review Technique)**

In the previous two estimation types, we used historical data from past projects and/or expert judgement to arrive at an estimate. This is only possible when such work has been performed in the past. However, if the activity is a totally new one or differs enough from the one done in the past, then we may not be able to come up with an accurate figure for the effort estimate. In such cases, we can use PERT. This technique suggests the use of three estimates – Optimistic (O), Most Likely (M) and Pessimistic (P).

An Optimistic estimate is the earliest possible completion of the activity

A Pessimistic estimate is the latest possible completion of the activity

A Most Likely estimate is the most probable completion of the activity

We then apply the PERT formula to calculate the PERT duration of the estimate. It is also called weighted average duration and is given by the formula below:

PERT Estimate O + 4M + P)/6

For example, if an activity has the following estimates:

O = 5 days

M = 8 days

P = 10 days

Then, PERT Estimate 5 + (4x8) + 10)/6 = 47/6 = 7.833 days

As “Most Likely” has been given four times weight, the PERT estimate would be very close to it.

When using a PERT estimate, we are also interested to know how much risk is involved in the activity’s estimate as that helps us determine the probability of completing the project on time. This is done by calculating the Standard Deviation of a PERT estimate. Formula for calculating standard deviation is give below:

Standard Deviation P – O)/6

In the above example, Standard Deviation 10 – 5)/6 = 5/6 = 0.833

It means that the activity can be completed within 7.833 +/- 0.833 days or anywhere between 7 and 8.666 days.

We shall see in the next section how we can calculate the standard deviation of the project when we are using PERT estimates for all activities in the project.

Once we estimate the duration of each activity using one of the above mentioned techniques, we then estimate the number of resources working on each activity. This gives us the activity durations that would be used for scheduling. It must be noted that law of diminishing returns applies to the conversion of effort to duration. For example, if an activity can be completed in 1 day by 1 person, then its duration is 1 day and effort is also 1 day. If we add one more person to the same activity then the duration may not go down to 0.5 days. It will take a little longer than that. This is because there will be an overhead of communication between the two resources. Hence, one needs to use expert judgement when converting effort to duration as it may not be a simple arithmetic calculation.

Project Management, Project Planning, Project Initiation, Project Execution, Project monitoring, Project Control, Project Closure