The framework described in the previous section was built upon the premise captured in equation (1), that the chief factors influencing the final duration estimate relate to the size of the task (or its ‘volume[1]’ as we shall now call it), the nature of the resources and the external working environment.   A general process for evaluating and justifying the duration estimate is best based upon examining each of these three factors independently.   Let’s examine each in more detail.

Task Factors

We have seen that there are two types of factors emanating from the analysis of the task itself.  The first, which we called the ‘effort driven’ effect, comes from the need for human resources to expend effort, perhaps augmented by material resources, in order to get work done.   Durations derived from this factor will depend upon the size of the task, the degree of its complexity and any distinctive properties or characteristics required.  For example in the case of the painting task discussed earlier the duration was affected by the wall space to be covered, the number of coats but also perhaps certain demanding attributes relating to the cornices or other interfaces with adjoining objects.   We make the point here that ideally, the volume is a quantity which is independent of the resources that will be performing the job.  In practice this allows productivities to be set in relation to the skills and experience of the resources likely to be assigned to the work.

The second factor, which we called ‘time effects’ involves influences that can be measured and estimated directly in units of time.  These arise both from

• cost-incurring sub-tasks not warranting consideration of effort and productivity and
• pauses, delays and lead or waiting times.

This is shown schematically in the following diagram.

Resource Factors

There are two chief sub-factors that influence the performance of the human resources that will be working on the task.  These are:

• Productivity rates
• Attendance rates

We saw how the skill and experience of the painters and the effect of a piece of equipment together defined a collective productivity which converted the quantity of task volume into one of effort, expressed in units of work-hours.  Productivity is therefore expressed as the number of output units per unit work-hour spent completing them.   It is a function of skill and experience which are pure attributes of resource capability.  However, if the task requires a particularly stringent degree of quality not already expressed as part of its volume then this can be incorporated into the productivity factor.  Productivity can also be adjusted to account for the effects of any materials or other non-human resources that may be applied to the job.  This can be expressed implicitly by means of an adjustment to the productivity rate or explicitly by means of a Material Resource Factor (MRF) which is then multipled to the productivity.  These methods are clearly equivalent and differ only the visibility of the effect of the MRF.  We saw an example of this in a previous post in the form of a long-handle painter which was deemed to provide a productivity boost to the painting team.

The second resource factor, the attendance rate, when dividied into the effort calculation provides an estimate of 'effort-driven' duration - that part of the duration that is directly dependent upon the quanity of effort expected.  The attendance rate is the average number of work-hours per unit hour of duration that can be applied to the task and is therefore measured in units of FTEs (full-time equivalents) as we saw.  Therefore, if we are able to supply 1.5 work-hours per unit hour of duration, this must mean that we have the equivalent of 1.5 people available to attend to the work.  On the other hand, an FTE of 0.5 would indicate that on average, only half of any unit of duration would be expended on the task at hand.

This means that the more people we add to the work, the less duration will be required.  That is, if we double the number of painters, we will halve the duration.  Mathematically, this can go on forever but of course as we alluded to earlier, in the real world things are different.   A famous law of economics, the ‘law of diminishing returns’ tells us that as more people are added, the duration reduction might initially speed up but then slows and if pushed too far, could even go into reverse.  The reason is that as the size of the team grows, it will eventually meet constraints of some form, possibly a lack of space or competition for tools, which then affects productivity.  In other words, productivity can become dependent upon attendance, and can increase with growth of attendance at low levels (increasing returns) but eventually fall away as these become too large.  We shall examine this effect in a future article.

To establish the likely attendance levels, we need to think in terms of ‘effective’ availability which is a realistic assessment of the amount of effort that can be expected of the human resources assigned.   The effective availability of a resource is defined to be the total time a person spends working per week less the time required to honour other commitments made or likely to be made.  The following should be recognised and factored into the result:

• The expected demands made from other projects are not necessarily evenly distributed over the weeks covered by the project and
• unexpected demands on a person’s time might be made by sources outside of our project.

Therefore a conservative assessment of effective availability should be used, i.e. one that is less than the average expected.

External Factors

To focus on task and resource factors exclusively would be to assume that the work takes place within an ideal set of conditions.  In reality there are usually effects from the work environment that will add to the time taken to complete the job.  These are also time-driven effects and so can be represented by adding directly to the duration calculated above.  These would include interruptions, delays, service breakdowns and a contingency component.

These are the three major influences that affect the estimate of duration and each has a presence in the general formula we derived in a previous article - reproduced here.



The role of these influences and their factors can be seen in the following diagram.

The two diagrams in this article provide both a process and a summary of the general framework that can govern the generation of all duration estimates.  The big challenge, as we have flagged several times in this series, is quantifying the task size.  This will be particularly acute when dealing with 'knowledge work' where physical measures such as square meters are in short supplly.  We shall deal with this next.