Saturday, September 17, 2011

Focus and Leverage Part 54

In my last posting I gave you some insight into problems that surface when using Critical Path Project Management (CPM). Specifically, I laid out four common behaviors that all contribute to delayed projects in one way or another.  The four behaviors are:

1. The Student Syndrome

2. Parkinson’s Law

3. Multi-tasking

4. Cherry Picking

So how does Critical Chain Project Management (CCPM) address and reduce/eliminate the impact of these four behavioral issues? There are six things that can be done to help mitigate Parkinson’s Law as follows:

1. Eliminate task due dates and durations.

2. Eliminate multi-tasking by focusing on one task at a time without unnecessary distractions.

3. Significantly reduce the built-in task safeties/buffers and create project buffers, feeding buffers and resource buffers.
4. Encourage the relay race mentality so that as soon as something is ready to pass on, notify the recipient in advance that it’s coming.

5. Change the way progress is measured.

6. Hold regular “escalation” meetings to provide timely assistance to workers.

Let’ look at these six things in a bit more detail and see how CCPM works to lessen their impact, but more specifically how to reduce the behavioral issues.

Unlike CPM, schedule building in CCPM requires that the schedule be constructed with only the time required to do the work, thus eliminating the safety time. This action leads to and supports complete elimination of task due dates and durations. Now when someone is notified in advance that work is coming, we can now take advantage of early finishes. Basically we are saying become a relay race runner. Without dates in place, we now know when a resource needs to be available.

One other significant difference between a CPM schedule and a CCPM schedule is that while CPM recognizes the longest path of dependent events (i.e. the critical path), it fails to recognize resource contentions. CCPM also recognizes the longest path, but also considers resource contentions and develops what it calls the critical chain. This is a very significant difference in how the project is managed.

To reduce the behaviors we experience with wasting time because of too much safety embedded, in CCPM we typically cut the actual estimated time by 50%. We actually say that this new time estimate is our 50% confidence interval. Because the task’s targets are 50% confidence estimates, what we are actually saying is that half of the time the task completions will come in early and half will be late. The theory is that the early finishes will offset the late ones so we don’t need all of the protection that was spread throughout the project. Instead we take the 50% removed time and create a project buffer which is inserted at the end of the project. The buffer acts very much like a savings account in that if more time is needed, a withdrawal of time is made from the buffer. Conversely, if a task is finished early, a deposit is made to the buffer bank. Realistically, the buffer is there to guard against Murphy’s inevitable impact on the project.

In addition to the project buffer, there are two other types of buffer that we establish. The first kind is referred to as a feeding buffer which protects task paths that feed into the critical chain. If a feeder chain is late, it will delay the tasks on the critical chain. Like the project buffer, the feeding buffer is typically half the size of the safety time removed from the feeder path. The other buffer is referred to as a resource buffer. These type buffers are created to ensure that the appropriate people and skills are available to work on the critical chain tasks exactly when they are needed.

Earlier I mentioned that CPM uses a simple % complete metric to measure project status and we were concerned about this method because not all tasks take the same amount of time. Typically, with CPM, the first 90% of the project gets completed fairly quickly, but the remaining 10% seem to take forever. One of the primary reasons is this metric. The key to managing a CCPM project is through buffer management. Unlike CPM, CCPM monitors the % of the critical chain completed compared to the % of project buffer consumed. As tasks are completed, we know how much buffer has been withdrawn or deposited in our project buffer. As long as the critical chain is completed without using more than 100% of the project buffer, we know the project is on schedule.

Two are final points to consider when running projects. The first is how CCPM eliminates cherry picking. When using CCPM software, each day a prioritized listing of tasks is presented to the project manager and are color-coded as red tasks. This is the order in which the tasks must be done, thus avoiding the cherry picking that goes on with CPM. The other point to consider is what type of meetings the project manager should have as he/she manages the project. The first meeting is a daily meeting to ask predecessor resources working on a critical chain task how much longer they need to complete the task so that they can inform the successor resource when to expect the new work. This gives the successor resource time to clear his/her workload in time to receive the critical chain task. The second meeting is a project review meeting whereby the team meets to review the status of the project. This meeting usually involves the use of fever charts that graphically display the % of critical chain complete compared to how much buffer has been consumed for each individual project. The project manager usually starts the meeting with a portfolio fever chart which displays all of the projects in work and then reviews each individual project fever chart. This meeting is intended to accentuate any problems with accelerated buffer consumption compared to completing the critical chain tasks. If these meetings are handled properly, the success of projects being completed on time and on budget/scope will increase dramatically.

Bob Sproull

No comments: