Crunch time

The view of traditional project management is that if every task is delivered on time then the project will be delivered on time. You can’t argue with that one – as long as the project has been properly set up with all the tasks identified and all the dependencies fully established.

But let’s be honest here. What is the probability that every task will be delivered on time? Even if you assume 99% certainty for each task, the cumulative effect of this across a project with, say, 150 tasks, is that it’s still highly unlikely that the project will be done and dusted on deadline.

In practice, the probabilities of tasks completing on time will be far less that 99%. As will be discussed below, we tend to plan with a probability in the order of 80%. We generally make up for this lower level of certainty by taking recovery action when things slip. We call that ‘fire fighting’, and as an industry we have become something of a dab hand at it.

However, a new approach to project management is beginning to emerge. It is being used increasingly in computer software development and defence contracting – two industries whose traditional performance in project management can make that of the construction industry look positively glowing.

No timewasters please

This approach is called Critical Chain Project Management. Its relevance to construction is huge and the DTI has funded a pilot study in the construction sector.

The Critical Chain approach has some very interesting things to say about the way we plan and manage projects. It recognises that we generally allow sufficient time for each individual task to be completed by including a safety margin or a ‘time buffer’ in our estimates. However, it’s not an exact science. Sometimes this safety margin is not enough and sometimes it proves to be unnecessary - and the effect of including a time buffer in each task is that much of this safety zone is often wasted.

This means that projects either overrun (despite having plenty of buffers) or they are not delivered as rapidly as they could be.

So how does this safety buffer get wasted? There are a number of factors in play here.

The first is the result of an asymmetric impact of dependencies between tasks. A task that is delayed, whose safety buffer has been used up, will pass this delay onto the next task. However, if a task completes early, this early completion is often wasted because the next trade is not ready to start before its planned date.

Another factor that can fritter away the time buffer is a phenomenon we may remember from our school days - the student syndrome. If you have more time than is needed to complete a task when are you likely to start on it? It’s just like doing your fourth-form French homework the night before it was due to be handed in!

Other buffer-wasting factors include starting work on a second task before the first is fully completed and the perennial Parkinson’s law, where work expands to fill the time available.

The Critical Chain approach sets out to neutralise these factors. It does this by stripping out much of the individual task’s safety buffer and aggregating it as a single large buffer to protect the project as a whole.

In pilot studies currently being undertaken by Kent-based Denne Construction, the concept of stripping out individual safety buffers was the most difficult thing to sell to site managers. This process involves reducing the time allowance for each task in the project to what is called ‘challenging but achievable’ durations. However, once the site managers could see that the time taken out of the individual tasks provided a single buffer which they could use to control the whole project, they became enthusiastic enough to give the system a go.

Fig 1 (above) is a probability distribution curve that provides the rationale for this approach. The task duration represents the programme time to which the subcontractor undertaking the task will commit. The greater the level of certainty required, the longer the task will take. Typically, we set durations with a reasonably high level of certainty of being able to complete the task in the time allowed, say an 80% level, equivalent to point C on the curve.

The critical chain creates a single ‘safety buffer’ to protect the project as a whole

Nicholas Fowler

The ‘challenging but achievable’ task duration can be considered as point B on the curve. Here, for example, there’s a 50% probability of completing the task in the time allocated.

The interesting thing to note about the Fig 1 curve is its shape. There clearly is a minimum time in which a task can be completed, represented here by Point A. However, there is no obvious maximum time in which the task will be completed. In effect, anything could happen. This is represented by the very long tail. If we insist on a very high level of certainty for the task completion, we can see why we are likely to get quoted a high duration.

So what changes with this new approach? Well, what was a 50-week programme remains 50 weeks overall. However, under the Critical Chain system, the project would consist of 40 weeks of planned work made up of the same tasks but with shorter individual durations and a total of 10 weeks of project buffer.

Once you get comfortable with using this system, the challenge is then to reduce both the durations further and also slash the size of the project buffer. This is where the ‘Buffer Management’ process comes in. Buffer Management is a system to monitor and control the project, which can also be used to identify ways of improving delivery on future projects.

This process involves monitoring the extent to which the project buffer is consumed as the project progresses.

Stop or go?

A ‘traffic light’ system flags up to the project manager how things are going. Any delay to the project will show up as consumption of the project buffer. If less of the buffer has been used up as a percentage of the progress of the job, the lights stay green.

However, if the rate of buffer consumption is greater than the progress of the project, the lights turn orange. This indicates that recovery action needs to be planned. If the lights turn red, then it’s time to get the extinguishers out.

A visual representation of this rate of project buffer consumption is provided by the ‘Fever’ Chart as shown in Fig 2 (left). A simple software programme provides all the necessary buffer management monitoring reports.

As part of this weekly monitoring process, reports are generated that highlight which tasks are consuming the project buffer. This allows continuous improvement to be a core element of the system. As well as highlighting where the fires are likely to be breaking out next, the weekly project review focuses attention on what measures should be implemented to ensure that fires can be prevented in future.

So impressed has one client, Housing 21, been with its pilot project that the firm has now adopted the Critical Chain process on an upcoming £8m sheltered home development. In addition, the team is using the Critical Chain approach on the pre-contract stage and will be involving the local authority and planning department in the process. This group involvement is seen as a key development in the application of this system and one that promises to deliver the greatest benefits. It is widely recognised that the most significant root cause of variability and disruption to construction projects is the lack of information available at the commencement of the construction phase.

So where next? Future areas for the research project include looking at how Critical Chain can be applied to tackle a number of the industry’s key ‘show stoppers’. These are big items that are usually out of the control of the project manager, and includes those tricky work elements that have long lead times requiring third party design, approvals and manufacture.

The next logical development will be the use of Critical Chain as a business management system to manage a contractor’s portfolio of projects. The ‘Fever’ Charts would be used to focus management attention where it is most needed. If it is green, leave well alone! cm