Handling High Uncertainty in Critical Chain Project Management

Some time ago, I was presenting to someone the concepts of Critical Chain — the project management methodology from the theory of constraints. At the end of the presentation, this person asked me: “I understand how this methodology protects against statistical and expected variation, but how can you apply this to high uncertainty projects?” I write this article to bring an answer to this question.

Critical Chain Project Management (CCPM) is an approach of project management which is a part of the framework of theory of constraints. It allows an improved protection of the delivery date, while enabling a more focus day-to-day management. People working on the project collaborate like in a relay race often allowing a quicker performance of the project. 

The best known tool of critical chain is a time buffer (called “project buffer”) placed between the last task of the critical chain and the delivery date. This buffer absorbs statistical variations of the performance of the tasks, compensates for expected risks and delivers signals to the project manager to prepare a recovery plan when necessary. It often represents about 25% of the planned duration of the critical chain.

But how this works when you manage a high uncertainty project?

Projects like the development of a vaccine or the construction of a spaceship are high uncertainty endeavor, for which the risks of deviation cannot be fully and precisely assessed initially. Unexpected events may happen that force the project to be re-evaluated, to cancel a part of the progress already made and restart at an earlier point. As a consequence, the project management methodology shall offer a mechanism to protect the project outcome when uncertainty strikes.

Every project is built on some hypotheses. While in usual case, these hypotheses are verified all along the project. In high uncertainty projects, some events happening may invalidate a part of these hypotheses, forcing a change on the project. This change can be a modification of scope to accommodate a performance that becomes impossible to reach, a dramatic increase in budget, or a significant delay because some clinical trials have to be retaken from the start. This change can even put in question the feasibility of the project and may result in its cancellation.

Dynamic Buffer Management for CCPM

The objective is to protect the outcome of the project materialized by two delivery dates: the original delivery date — based on the original hypothesis — and the “long date” — based on our best estimate of the uncertainty striking the project.

Having two delivery dates complicates the role of the project manager. He shall keep his team focused on delivering on the earlier date, the “target date,” while acknowledging the “long date” as a back-up in case of important changes in hypotheses. The team shall focus on executing the project like a relay race to deliver on the “target date,” and shall not aim to the “long date” to avoid transforming the “long date” into a self-fulfilling prophecy of failure. At the same time, the team shall recognize the risk of deviation for unexpected reasons to identify it when it happens.

A major threat to the execution of the project is regularly evaluating the risk of the project. Each risk analysis of the project can easily uncover events motivating a change in the plan (even if it is only to self-justifies the exercise of review). The plan is modified at each review and bring chaos in the execution. The methodology shall provide signals about when it is time to review the risks status of the project and when it is only time to execute as per the current plan. Such a mechanism protects the team and the execution of the project by bringing mandatory stability.

A solution is to add a second buffer between the “target date” and the “long date.” The project has two main buffers now: The “project buffer” and the “change buffer” (as this buffer shall protect the “long date” from changes in the context of the project). Besides this second buffer, the project needs specific rules to trigger risks analysis based on buffers penetrations.

Credits Unsplash — thisisengineering raeng

An Initial Risk Analysis

The project manager shall perform a comprehensive risk analysis at project initiation. This analysis clarifies and formalizes which risks were considered to determine the “target date” (we name them “expected risks”) and which ones were built into the “long date” (we name them “unexpected risks”). This analysis shall also stipulate how the remaining uncertainty (undefined risks — part of risk that were not identified yet, but can still strike the project) has been embedded into the “long date.”

The duration of the “project buffer” is determined as usual, but the duration of the “change buffer” is determined from the most probable impact of the “unexpected risks” plus an extra time to cover any not-yet-identified risks.

The project buffer

The penetration of the project buffer is evaluated as usual by the variation of the finish date of each task of the critical chain.

When the buffer penetration enters the yellow zone, it is the signal to prepare a recovery plan. But in the case of high uncertainty project, this situation triggers some additional actions.

The project manager shall prepare in parallel:

• An analysis of the causes of the penetration into the yellow zone to determine how much of this penetration is caused by variation of performance of tasks and “expected risks” and how much is caused by “unexpected risks”;
• The usual recovery plan to compensate for the variation of performance of the project due to “expected risks”;
• If the risk analysis determines that the project has been hit by “unexpected risk,” the project manager prepares an action plan to tackle the newly identified risks. This plan determines how much time he needs to transfer from the “change buffer” to the “project buffer” to take these new risks into account. This transfer creates a new “target date,” but still keeps the “long date” protected.

When the buffer penetration enters in the red zone, it is the signal that the recovery plan and the additional risk mitigation plan executions are launched. If this is the case, buffer time is transferred from the “change buffer” to the “project buffer.”

The change buffer

The cumulated time transferred from the “change buffer” to the “project buffer” is the “change buffer” penetration.

While this penetration remains in the green zone, nothing is to be done. This corresponds to the uncertainty that is expected to hit due to the nature of the project. The situation doesn’t need yet any restructuring of the project and reacting at that stage is only bringing instability and chaos.

If the penetration enters the yellow zone, a significant share of the original hypotheses of the project is invalidated. It is the signal to the project manager to start an analysis to determinate if the project, as it is defined, can still be executed and what new arrangement may have to be set to cover the new risks. 

This analysis has the potential to change significantly the scope, delivery term or budget of the project. It may happen that major hypotheses underlying the original schedule or a major criterion of success (scope, budget or time) get challenged. If this is the case, the plan may require a full rescheduling of the project with new “target date” and “long date,” budget and potentially new adapted scope. Therefore, the report of conclusions and corresponding action plan shall be signed off by the major stakeholders of the project.

If the penetration of the change buffer enters the red zone, the plan is activated to protect the “long date.”

The synchronized management of the two buffers allows protecting the “target date” and keep the team motivated, while acknowledging the high uncertainty nature of the project and adapting it dynamically to a new context. It gives the needed signals to re-evaluate the risks at critical moments avoiding the exaggerated and disturbing burden of regularly assessing risk. This prevents high-level expediting and imposing corrective or mitigation action when it is not yet necessary.

The publication of the penetration of the buffers to the whole project team is a great motivational tool. A well-informed team spontaneously works on solutions to bring back the project in the green zones without the need for building or executing recovery plan.

Didier Varlot has practiced the Theory of Constraints for the last 25 years in several industries from railways industry to healthcare services, from chemical industries to green energy supply. 

Find his other theory of constraints article on his Medium blog. His webinars and articles are also accessible from his website.