The illusion of certainty?
Stephen Lowsley, Delay Expert compares Critical Path Analysis as a project management tool and its use in delay analysis.
A brief search of the internet will provide a plethora of academic articles relating to the project modelling technique known as Critical Path Analysis (CPA). In this article I am not going to discuss the operational details and the various nuances of the technique, I am going to compare the use of Critical Path Analysis as a project management tool and its use in delay analysis to quantify delay and extension of time (EoT).
The development of CPA in the 1950’s is largely credited to the US company DuPont with the technique being referred to as the Critical Path Method (CPM). Although there have been some refinements, the basic principle of CPM is employed in the modern programming software of today such as Primavera and the like.
At the same time that CPM was being developed, the US Navy Special Projects Office embarked on the Polaris Intercontinental Ballistic Missile program. This was a major project involving research and development of missile launch systems, government organisations, academic institutions and contractors. In order to monitor and control this vast project a system referred to as PERT (Program Evaluation and Review Technique) was developed.
PERT is similar to CPM, however rather than a single duration for each activity it allows for a ‘three-point estimate’ consisting of a most likely time, a shorter optimistic and a longer pessimistic time. These durations are applied to the activities statistically and in the case of the Polaris project allowed effective time management of the very many participants, the critical path and float and it is credited with a saving of two years in the overall project period.
The PERT three-point estimate of activity duration is used today in ‘Monte Carlo Simulation’ risk analysis techniques to estimate the probability of completion at specific times, recognising that there is not a single specific outcome and a single specific projected date for completion.
In the UK CPA programming software is extensively used for the planning and control of construction projects, unfortunately it is often probably not used as effectively as it could be. Many construction projects can be considered as being similar, however all projects are to a certain extent unique, involving many participants, a high level of complexity and uncertainty with plenty of scope for things to go wrong, probably not dissimilar to the US Navy’s Polaris project.
I have been involved with the programming and management of construction projects for over 45 years and in this time have been involved with CPA from hand drawn networks in the 1970’s through to the modern software of today. In my opinion CPA is an invaluable project management tool, however it provides a precise answer and as such requires precise information. Such detailed information is unlikely to be available and CPA provides an answer based on a very prescriptive set of circumstances, usually one of a very extensive range of possibilities.
In project management circles this programme uncertainty is clearly recognised with CPA being utilised to simulate the likely outcome rather than to predict the future with absolute certainty. Due to this uncertainty a CPA based programme is not static and operationally there is likely to be a need for an ongoing amendment of computer logic and activity durations to reflect the changing circumstances and to ensure that the programme reflects a realistic forward projection.
The PERT system developed over 60 years ago recognised uncertainty in respect of the management of time and in my opinion, programming involves looking at possibilities rather than a search for absolute certainty.
In the world of delay analysis, the term ‘Critical Path’ can be very emotive and I have found that things can be somewhat different from project management with a static CPA being relied on to calculate and quantify a precise answer.
The introduction of the PC and the development of programming software was seen as a bit of panacea in respect of the analysis of delay with the credibility of expert opinion being dependent on a computer-based CPA. Over time, however this approach has attracted some criticism including adverse comments from the courts.
A more pragmatic approach based on experience of the construction process is now usually considered as more appropriate. This aligns more with the use of CPA from the perspective of project management rather than its use to quantify precise delay for delay analysis.
As a simple example of the differing approaches, on a two-and-a-half-year project a delay to the critical path of two days occurs at an early stage.
From a project management perspective, it is reasonable to suggest that despite the delay being considered as critical the delay could be overcome in the future two-year period, required to complete the works without any additional effort or input. Although this is the case the risk of not meeting the completion date and the probability of a later completion will have increased.
In a typical delay analysis utilising the planned programme and CPA a precise measure of a two-day delay to completion will be calculated.
A two-day critical delay occurring in the last month of the project or an initial delay much greater than two days is more likely to result in a direct delay to completion in both scenarios.
This being the case the impact of a critical delay on completion may be dependent on the timing of the event and its extent rather than a blind calculation of the delay to the critical path.
One of my favourite analogies relating to the use of CPA for delay analysis was made by lawyer Doug Masson who when discussing problems with logic and causation said “CPA becomes as stable as a house of cards”1
A similar comment was made by the courts in the case of City Inn v Shepherd where Lord Drummond Young considered that a major difficulty “is that in the type of programme used to carry out a critical path analysis any significant error in the information that is fed into the programme is liable to invalidate the entire analysis.”2
I concur with this statement, however in my opinion to provide a precise answer any CPA requires precise information and therefore rather than just error, as considered by Lord Drummond Young, any uncertainty is also likely to invalidate the entire analysis.
The development of PERT recognised uncertainty and in the use of CPA for project management this uncertainty is recognised with CPA being used to evaluate likely outcomes, whereas with delay analysis CPA is often blindly used to calculate delay without consideration of the facts and difficulties and provides a false illusion of certainty.
I have some concern in respect of the use of CPA for the retrospective analysis of delay, particularly if it has not been used during the course of the project. This is not to say that it should not be employed, however any use should be undertaken with care with the emphasis being on experience and the evaluation of the factual evidence rather than a precise answer calculated by a computer.
1Masson, D. Following the Critical Path, Contract Journal 16 Feb 1995 P34-35
2City Inn Limited v Shepherd Construction Limited [2007] CSOH 190