Critical Chain Project Management & Feeder Buffers

Chain_red_critical_link_1Published in the Project Post-Gazette, February 2014

by Paul Lohnes,. MBA, PMP

February Project Post-Gazette

As we promised, this article in the series of understanding the differences between the Program Evaluation & Review Technique (PERT) / Critical Path Method (CPM) and Critical Chain Project Management (CCPM) when it comes to how a schedule deals with resource assumptions and duration estimation. While this may be something for a scheduling article, the PPG has already dealt with the schedule aspects of CPM versus Critical Chain Method (CCM) (See the PPG Feb 2013 and May 2013 issues – Schedule Dispatch); this discussion is about the management prerogatives and decisions that using these approaches to the resource management and duration determinations require.

A quick review of both the PERT/CPM and CCPM is in order to prepare us for our more detailed discussions of how the CCPM use of feeder buffers builds on our last article (See Jan 2014 Well-Skilled PM) on how CCPM deals with duration estimates problems through its use of a project buffer.

In PERT/CPM, the tasks are designed into a network of interdependencies along with their duration estimates and resource assignments under the assumptions that resources are available to the project at 100% for the life of the project. Also, CPM tends to produce duration estimates that are victims of both Parkinson’s Law and the Urgent Deadline Syndrome (sometimes called the Student Studying Syndrome) that tend to increase the amount of safeties added to the duration estimate. This means that those asked to provide duration estimates tend to expand their projection with a value that will allow them to be right about 95% (use of the 95% confidence level) of the time given the consequences for coming in behind schedule. This also firms up the inability for CPM managed schedules to take advantage of tasks that finish early since the task responsible party is chided for providing a ‘padded estimate’ instead of being congratulated. Since PM are not desiring to put subsequent tasks into execution early since it will cause the budget impact to shifted to the left – temporally.

CCPM on the other hand takes the approach that estimates are padded given the training of the current generation of PM and BA, so it uses a 50% confidence level of duration by cutting all duration estimates in half. This simply means that 50% of the tasks will complete early and 50% late. Those that complete early can give up their ‘padding’ to those that do not – providing an amount of estimation slippage in both directions. The real advantage of CCPM comes from taking the 50% of ‘reduction’ of estimate duration (this is 50% of 50% reduction meaning 25%) and using it as a project completion buffer (PCB) at the very end of the project to act as a ‘surge protector’ for tasks that need it when estimations are really off the mark. Thus, the project deadline begins at 75% of the normal CPM deadline – 50% durations plus the 25% PCB.

Now, the network of tasks that form the longest chain is now the critical chain, but there are other solutions that can assist with task chains that are not on the critical network. These are called feeder chains and the buffers that are created from their 50% confidence factor (CF) reductions are placed at the end of these chains BEFORE they rejoin the critical chain. Feeder buffers are sized in similar fashion to the overall critical chain PCB in that using the normal estimates of those tasks comprising the feeder chain (for example chain CD in Figure 1), cutting them to a 50% CF where 50% of the tasks will completed early, and 50% late. These buffers help to prevent these feeder chains from slipping to the point where they begin to impact the critical chain.

figure1 buffer

Figure 1 PCB and Feeder Buffers

Figure 1 shows both the Project Completion Buffer (PCB), and the feeder buffers that help to protect the critical chain (in red) from schedule delays that originate in the feeder chains (non-critical chains) such as CD fb1 -> K, and CE[GI] fb2 -> END. Notice that the feeder buffers are placed at the end of the feeder chain before it rejoins the critical chain or the project completion node. The PCB protects the entire project by supporting those delays on the critical chain that are not absorbed by what we call the ‘50% over/under’ estimation process. This is where the Well-Skilled PM (WSPM) took the ‘normal estimates’ that are the result of decades of training to which your PM and schedulers were subjected that involved the application of Parkinson’s Law, Urgent Deadline Syndrome, and the Bad Multitasking Syndrome. These biases cause most estimates to contain largely inflation safeties in order to protect behavior that penalizes early finishes, late, and missed estimates on the part of your PM and schedulers. The WSPM must learn to manage projects without these outdated and destructive management practices if you want to improve your project success rates.

Using the feeder chain CD where for our example, the normal estimates are 6 and 8 units respectively (See Figure 2 below). This would mean that for argument’s sake the normal estimates are at 95% (close enough to 100%) so we can create our 50% CF values of 3 and 4 units respectively resulting in a feeder buffer of 50% of 7 units (the remainder of the normal duration estimate) or 3.5 units. Since the normal estimates are not precisely 100%, we usually round down to a whole number of 3 units for the CD feeder buffer. You can develop your own best practices (working practices) given the context of your projects and organizations.

figure2 buffers

Figure 2 Feeder Buffer Sizing

Why does CCPM and buffering work better than the traditional project management perspective based on the both the CPM and Program Evaluation and Reporting Technique (PERT) principles? It is not so much the advantages of the CCPM solution, but how it addresses the limitations and assumptions of the PERT/CPM approach that we have listed in previous articles, but will list here for coverage:

  1. Not starting a task until the last moment, but scheduling it on a earliest start basis (Urgent Deadline),
  2. Taking the entire estimated time for task, or not finishing early regardless (Parkinson’s Law), and
  3. Management by consequences where estimates become commitments

The CCPM solution addresses these limitations of the PERT/CPM approach by being resource-objective, human-nature friendly, and management smart. What this means simply is that the CCPM solution does not support one of the biggest issues with the PERT/CPM approach that resources are infinitely available to project and that they are often assigned to multiple activities (multitasking) on the same or other projects. CCPM determines the critical chain by determining and investigating the resource limiting aspects of task assignment and interdependence by discouraging multitasking and the bloating of duration estimates that these actions create. Through the use of the aggressive duration estimation process detailed above, the CCPM supports and utilizes early finishing tasks and does not penalize task responsible parties for such actions. Also, by removing the traditional project management principle of holding task owners to their estimates as commitments or not making their estimates (within a manageable time frame), the results of poor estimates can be reduced significantly.

Now, having detailed how CCPM deals with these limitations, we caution readers in what may be their wish to implement these principles immediately. PLEASE, do not! Take time to do your own research and increase your understanding of the CCPM solution. Our sister training institute, the MCLMG Training Institute (MTI) provides online seminars on these topics. You can email their proctor for more information.

In this article we have covered the purpose for using feeder buffers and how they are sized, the next step is how to manage the feeder buffers and in that case the PCB as well. However, we have already run over our time and space and thus must leave that topics for our next article in March 2014. We suggest that you do your own research and knowledge improvement on these topics.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s