All You Need to Know About Critical Chain Project Management

When you start a project, you plan to succeed, but few projects are without their problems. Eliyahu Goldratt’s Critical Chain Project Management (CCPM) was added to the Project Management Body of Knowledge (PMBOK) in 1996, and it aims to do away with frequently-encountered issues that can lead to a poorly performing or failed project.

CCPM methods help project managers to focus on the project schedule, but they also help to reduce scope creep once it’s underway. After all, it’s harder to shoot at a moving target or accomplish a goal when the goalposts keep moving. The most successful projects are completed within the planned timeframe with the planned resources. When they don’t, cost overruns are almost inevitable.

So far so good, but you’re probably wondering, is the Critical Chain Project Management (CCPM) approach actually used? After all, beautiful-sounding theories don’t necessarily work well in practice.

Well, the long roll of companies that have implemented CCPM and experienced improved project performance is impressive. The results speak for themselves. Big names like Texas Instruments, Harris Semiconductor, and Lucent Technologies adopted CCPM and now report that they spend half, or even less than half the time they once needed for similar projects.

Critical Chain Project Management - the theory

CCPM consists of three theories. Before we look at the practical application of CCPM, we’ve got to understand the thinking that goes into it.

1. Theory of constraints

This theory may seem all-too-obvious, but sometimes, the obvious becomes easy to overlook. Goldratt’s theory of constraints says that any system will have constraints that govern its output. To focus on constraints, we follow five steps:

1. Identify the relevant constraint
2. Exploit it (get it working at 100% capacity)
3. Subordinate all other elements of the system to the constraint
4. Elevate the constraint (add resources to reduce the constraint, but only if it isn’t broken in steps one to three)
5. Return to step one and repeat the cycle for every new constraint

2. Common cause variation

There are two types of variation. Common Cause Variation is part of the system as a whole. Special Cause Variation has a specific source such as a team, an employee, a machine, or a circumstance.

Common Cause Variation applies to projects because it affects the time needed to perform part of any project. Let’s say you’ve got several teams. On the surface, they work independently, but one team can’t start working until the other team has finished its work. The obvious effect will be variations that stem from the interdependence of activities.

To make this clearer, let’s say you flip a coin. You might expect to have a 50/50 distribution of heads vs tails, but the fewer the times you flip the coin, the more likely you are to get a skewed result. Cumulative probability effects mean that the more tasks there are, the more likely we are to get a median result.

3. Statistical laws govern common cause variation

Let’s suppose we’ve got a chain of activities. Each activity in the chain has a 50 percent likelihood of finishing within a day.

But there’s a 90 percent probability that the activities will take two days. Supposing that there are four activities in the chain, scheduling them to take eight days makes sense on the surface. However, there’s a strong possibility that we could finish sooner.

If we use the statistical law of aggregation, we can achieve a 90 percent probability by scheduling the activities at 50 percent of the worst-case estimate of eight days. Just to be on the safe side, we can add a two-day buffer. Thus, instead of scheduling eight days for the chain of activities, we can schedule them for six days.

We also need to remember the central limit theorem. In simple terms, this means that the larger the sample, the closer it gets to a normal distribution. In projects, we find many tasks that have a minimum possible time needed and then the potential to run much longer than that. We saw that in the graph we looked at a little earlier on. But the central limit theorem says that the combination of many activities would have a more symmetrical distribution than we saw in our graph.

To put CCPM into practice, you need software that can track task dependencies, manage buffers, and provide real-time visibility into project progress. Traditional spreadsheets simply can’t handle the dynamic nature of critical chain scheduling. You need something that updates in real time as buffer consumption shifts, and that’s where dedicated workflow tools earn their keep.

What problems does CCPM solve that other methods miss?

Using the central theories we just discussed, Critical Chain Project Management strives to eliminate undesired effects that make projects longer and costlier than we planned. OK, ‘eliminate’ is probably too strong, but the improvements are real. We know the theories. Now, let’s look at what CCPM addresses when compared to a traditional approach. Armed with this knowledge, we’re ready to walk through the process.

1. Eliminate overly long estimates of how long things take

Project managers are already very well aware that the minimum time things take and the actual time it takes to get them done can vary greatly. As a result, they tend to play it safe when it comes to time estimates.

They get their information from people who already know the job that must be done. What they don’t give, though, is the probability that the estimate is correct. If you return to the graph of probability vs. cumulative probability, you’ll see that a low-risk estimate could be twice as long as the one with a 50 percent probability of accuracy.

With project participants and project managers tacking on a contingency time to be sure work will finish on schedule, overall projects end up covering a whole lot more time than they really need to. In our conversations with COOs at mid-size wellness and nutrition companies managing global expansion, we’ve observed that approval workflows often add 30-50% to total project timelines because each person pads their estimates independently.

2. Wasting time waiting for things to get urgent

With lots of time to get things done, most people involved in a project will report timely completion. But about ten percent of them will report late completion. Given that the project has played safe with a 99 percent probability of timely completion, 99 percent of activities should finish on time or even earlier than expected. Why does this happen?

Let’s think back to the time when you were a student. You knew that you’d be writing exams at the end of the semester.

During the semester, you probably did about one-third of what you needed to do to prepare for the exam. But in the last third of the semester, when D-Day drew nearer, you tried to get two-thirds of your studying done. “I didn’t have enough time to prepare for the exam!” you said.

The truth is, you did have enough time.

You just tried to get everything done at the last minute.

The phenomenon of planned procrastination is often referred to as “student syndrome,” and it’s human nature. Turns out, people tend to drag their feet until things get urgent, and the same is true of projects. You’ve probably heard C. Northcote Parkinson’s saying “Work expands to fill available time,” and it’s true!

Did you follow this pattern when you were a student? Are you project teams following it too?

3. Failing to pass on time or resources saved

Failing to pass on savings is a painful problem in projects. Again, human nature has a role to play. If you’re part of a project, you’ll be rewarded for finishing on time, and you’ll be punished if you finish late. If you finish early, on the other hand, you don’t get anything. In fact, you could end up with less than you would’ve received if you finished on time because you spent less time on the task. The same is true of materials or resource savings.

If you use less than was first supposed, there’s nothing in it for you. On the contrary, you might lose out. So, what do you do? You down-prioritize resource savings and time savings. You overwork steps that could’ve been reported as finished long ago, and so on.

4. Delays as a result of activity paths merging

When you’re busy with a project, be it simple or complex, several things are usually happening simultaneously. At some point, usually near the end of a project phase, everything starts coming together.

There’s just one problem: if the merging activity paths consist of three paths, and one of them is early, the second on time, and the last late, the phase isn’t complete till the late activity is finished. That makes all three processes as late as the one that got completed last of all. Which is a bit maddening.

Older analysis techniques like Critical Path Analysis don’t allow for path convergence delays.

5. Delays because of multitasking

We usually think of multitasking as being a “good thing,” but in the context of projects, it causes delays. It’s easy to understand why.

Let’s suppose that you’re busy with four activities and that each of them takes a week to finish. Because you’re multitasking, you don’t finish any of them in a week. You split your day up into segments, and each task ends up taking four weeks to complete.

If this is how you usually work, you’ll end up budgeting four weeks for a task that should take only one week.

The result?

A project that takes longer than necessary to finish. We got this wrong at first too. We let people juggle tasks in parallel until we learned that forced sequential task completion dramatically improves both speed and quality of work. Interestingly, research has shown that we just aren’t wired for multitasking. So, we can add the fact that the total time needed for each task is longer than it ought to be too.

6. Losing focus in areas on which you should be focusing

As a project manager, there are a lot of things to distract you from the priorities where your focus is most urgently required. These could include:

• All activity paths starting simultaneously
• Changes that occur during project performance
• Using dollar value instead of schedule performance as a measurement
• Spending time addressing variances that fall within acceptable parameters

Can a project manager handle all of these at once? Not effectively.

Putting CCPM into practice

Now that we’ve got a (very basic) understanding of the theory behind Critical Chain Project Management (CCPM), and the problems it tries to eliminate, we can examine just how to do it. Here’s how we can finish projects sooner and avoid the pitfalls of traditional project management.

1. Exploit the constraint

Up till now, your project management activities have been based on constraint after constraint. This might happen late, or that might happen late, so you budget more time than you really need just so that you can get finished on time.

The first step is to get the people who estimate the time needed for a task on the same page as you are. First, they need to know what you are trying to achieve.

No, you don’t want the “safest” estimate alone. Instead, you want the average time in which an activity can be completed if everything goes well. You’ll also want the low-risk estimate they’d usually give.

Now that you have that info, you can start constructing your critical chain using the average time, plus a buffer that is midway between best and worst-case scenarios.

2. Subordinate chain paths that aren’t critical

Most project managers allow noncritical paths to start early. They reason that by doing so, these paths have enough leeway to get finished in time. Use this approach, and the noncritical path participants know they’ve got lots of slack to play with. They’ll use it to the full.

With CCPM, you’ll use late start schedules, and you’ll build in a reasonable buffer, but you won’t necessarily plan for a worst-case scenario.

3. Use buffer times effectively

One misconception we see constantly is that buffers should be sprinkled into every individual deadline. In Critical Chain Project Management, you don’t do that. Instead, you place a single shared buffer at the end of the critical chain. By doing so, you can exploit the law of aggregation which we discussed a little earlier. The buffer is listed as an activity itself, but you don’t specify what work might fill it. Sounds simple because it kind of is.

The simplest way to calculate buffers is to add up all the activity time-frames. Half of that total represents your buffer. This buffer is shared among all activities, so if one is late and another can’t start because the previous one was overdue, there’s time to set things to rights.

Planning templates for project management

Example Procedure

Quarterly Strategic Planning & Goal Setting Workflow

1Revisit annual plan goals

2Break down goals into smaller chunks

3Review budget and benchmarks

4Create action steps and benchmarks

5Set expectations and timelines

+2 more steps

View template

Example Procedure

Annual Planning

1Define your goals using SMART criteria

2Build your budget and financial projections

3Set timelines and quarterly checkpoints

4Create contingency plans for when things go wrong

5Review and finalize the annual plan

View template

4. Buffers for subordinate paths

A critical chain consists of subordinate paths that feed into the critical chain. Project managers must protect the critical chain by providing potential buffer time to use at the point where each subordinate chain feeds into the chain.

By doing this, you can protect your critical path from delays in subordinate paths. And if those delays don’t materialize, your project speeds ahead towards an early delivery date.

5. Resource buffers

Now, it’s time to do away with overruns on resources. You’ll only apply resource buffers to the critical chain, not the subordinate ones. When you’re working on a project that carries a lot of risk, or if you’re using subcontractors, financial incentives could be among the resource buffers you apply.

That means that you can incentivize early delivery instead of penalizing it, and you can build in penalties for late delivery. The feeding chains don’t get these buffers because you’ve already added time buffers at the end of each one.

The critical chain and people

We’ve already seen that typical human behavior patterns affect project performance. CCPM overcomes these behaviors. Here’s how:

1. Ditch “student syndrome” date-driven delays

Instead of giving dates for individual activities, you only provide dates for completion of the activity chains as well as the buffer time. Now, the teams engaged in the project aren’t focused on deadlines they can delay. Instead, they focus on finishing as soon as possible.

Start dates for the rest of the critical chain are approximations and aren’t cast in stone. Because you’re planning according to best-case completion times, you don’t impose penalties for being late if work started when the resources were available, people and teams aren’t multitasking, and they pass on the completed activity as soon as they’re finished.

2. Eliminate multitasking

As a project manager implementing CCPM, you’ll expect 100 percent focus on the task at hand from every individual and team.

Multitasking would be fine if you could do several things at once and still focus on each one to the full. But that’s not humanly possible, so multitasking goes out the window.

3. Buffers and managers

We saw that project managers often lose focus on a project because they are giving attention to processes or tasks that still fall within an acceptable variance. In CCPM, the buffer time is the indicator of when trouble might be ahead.

You’re expecting to use at least some of the buffer time to tie up loose ends, so the first third of it doesn’t yet represent a time when the project manager needs to intervene.

Once the delay enters the second third of the buffer period, it’s time to start examining the situation a little more closely. Is there a problem, how much of a problem is it, and what should be done?

The last third of the buffer indicates the time when managers initiate contingency plans to get things back on track. By applying this approach to management by exception, project managers can skip unnecessary interventions entirely.

4. What if you’re managing several projects at once?

Although it’d be wonderful to focus all your effort on one project at a time, the truth is that you, as a project manager, are probably taking care of several projects at once. What do you do now?

The answer is to determine what the multitasking constraint is. Where’s capacity likely to be constrained?

Is it the availability of qualified personnel? If these constraints can’t be overcome in a multi-project environment, you’ll set the rhythm for the management of your project based on the limitations you can’t eliminate. Simplifying matters could be as easy as adding capacity buffers between several simultaneous, synchronized projects.

If the constraining factor doesn’t overrun the buffer time, everything’s still on track, and if all the buffer time isn’t needed, the project can be moved forward and completed early.

5. Critical Chain Project Management (CCPM) and resources

Since resources can never be infinite, resource allocation must be prioritized. In discussions with operations managers at growing e-commerce companies, we’ve heard that resource contention becomes the primary constraint once teams scale past 10-15 people. The following criteria work best:

• Critical chain activities take precedence over no-critical activities
• Activities that have the greatest potential to penetrate the project buffer over ones that haven’t.
• Activities that penetrate feeding buffers to the greatest extent over ones that have lower buffer penetration.

Why CCPM is simpler than you think

Compared to other project management techniques like PERT, Monte Carlo Methods, and earned value methods, CCPM is relatively straightforward. When project managers can monitor buffer penetration in real-time, they’ll have enough time to help the process along so that it can finish before the buffer time is over. They know when to start examining delays more closely, and they know when it’s time to intervene.

The planning process is also simplified and doesn’t require any specialized software tools. However, software remains useful in monitoring project progress. Tallyfy provides a clean, simple, and effective means of monitoring project progress in real time.

A quick summary of CCPM

When working on single projects

• Capture the critical chain with an eye to resource constraints.
• Reduce the times you would ordinarily use to the point where they have a 50/50 chance of being complete (aggregation will help these to be realistic).
• Place a buffer time at the end of the critical chain.
• Add buffers to subordinate chains that feed into the critical chain.
• Add resource buffers that will ensure you have the resources you need when you need them.

When working on multiple projects simultaneously

• Know which resource will constrain progress
• Create a schedule for that resource
• Adjust project sequence to match the constraining factor
• Add buffers based on resource constraints or capacity limitations

Measurements and controls

• Buffer management will be the area of focus for project managers.
• Assign resources based on buffer information.

Finally, ensure that you’re not incentivizing or encouraging late completion because you didn’t take the human element into account.

Why adopt CCPM? If you do so, you stand to complete projects faster and more cheaply. The Project Management Institute (PMI) reports that Critical Chain Project Management (CCPM) can reduce the time needed to complete projects by 50 percent or more. Now, that’s something to celebrate!