Tuesday, June 11, 2013

Focus and Leverage Part 225

In this posting we'll review an additional TOC methodology used for improving the completion rate of projects.  We'll do this by comparing TOC's version of project management to one of the most commonly used project methods.
3.  Critical Chain Project Management (CCPM) – Most companies today are using the more traditional Critical Path Method (CPM) for executing projects.  The problems experienced using this method are typically projects being completed later than the promised due date, and when they do get completed, a high percentage of them are  over budget with the original scope reduced.  CPM uses estimated task durations with well-defined individual task start and completion dates.  Typically progress is measured by a percentage of tasks completed against the projected delivery date.  The problem with tracking a project like this is that not all tasks have the same duration and the tasks resources have a tendency to choose easier tasks to work on so that their tracking metric looks good (i.e. cherry-picking).  Using this metric to track progress, it’s not uncommon to see 90 % of the project completed fairly quickly, only to see the remaining 10 % dragging-on endlessly.

In addition, when the project is being planned, resources responsible for completing tasks add excessive amounts of protective safety to the estimated durations to guard against variation and uncertainty. When this happens, the final delivery date is a reflection of these inflated task durations.

Another problem with CPM is that resource contentions are not considered when calculating the critical path of the project.  The critical path in CPM is defined as the longest path of dependent tasks meaning that a task can’t be started until the previous one has been completed.  As a result, when given multiple projects to work on, the resource is forced to use a behavior known as "bad multi-tasking."  Bad multi-tasking occurs because the resource is forced to split time between the multiple projects that have been assigned in order to show progress on each of the assigned projects.  That is, because the resource wants to demonstrate progress on all of the assigned projects, they will work on one project for a while, then switch to the next assigned project and so on.  Bad multi-tasking leads to significantly longer project cycle times.

One of the principal problems associated with CPM is what happens when a task is completed.  Because estimated task durations, with a well-defined task start and completion dates,  are used to develop the CPM schedule, there is no provision for accommodating tasks that are finished ahead of schedule.  Because of this, a behavioral phenomenon known as Parkinson’s Law rears its ugly head.  Parkinson’s Law states that work expands to fill the available time, so the resource will use all of the time allotted to the task.  They do this because their credibility is on the line, meaning that if they said 5 days and the task was completed in only 3 days, they would be expected to duplicate this on subsequent projects.  So even if a task is completed early, it is not reported as such.

Another behavior that comes into play is what Goldratt dubbed the Student Syndrome.  Because the task estimator knows that he or she has loaded the task duration estimate with a protective safety buffer (as much as 50% of the task duration), there is a tendency to procrastinate the start of the task.  And when Murphy strikes, and we know he will, the task can be delayed which causes delays in the project.  The consequence of these two behaviors is that early finishes are not passed on, but delays are.

Once again, TOC offers a different approach to project management known as Critical Chain Project Management (CCPM) and the differences between CCPM and CPM are enormous.  Like CPM, sequential project tasks are laid out and durations for each are estimated.  However, because we know that these duration estimates are many times in the neighborhood of 50% overstated, CCPM removes the task durations from the plan and then pools a portion of the withdrawn safeties to create a project buffer.  This buffer is calculated based upon a certain percentage of the removed safety time (i.e. approximately half of the task duration time estimates).  This project buffer is similar to how a bank account works with deposits and withdrawals.  By that I mean, if a task takes longer than planned, time can be withdrawn from the project buffer.  Likewise, if the task is completed faster than the planned time, the time saved is added back to the project buffer.  CCPM also includes a feeding buffer which protects the planned completion dates for tasks not  located on the critical chain.

CCPM also has a different way of tracking projects than the way CPM does.  CCPM uses a graphical display called a fever chart to provide an early warning when the project is stalled.  The fever chart is much like a run chart in that along the x-axis we see the percent of the critical chain completed while on the y-axis we use the percent of the project buffer consumed.  The fever chart has three different color zones to indirectly indicate the rate of buffer consumption.  If you divide the percent of critical chain completed into the percent of project buffer consumed you will have an index that indicates how the project is progressing.  If this ratio is less than or equal to 1.0 (green zone), the project will either be completed on time or early.  If the ratio is greater than 1.0 (yellow or red zone depending upon how high the index value is), then a plan must be developed and executed to recoup the lost time or the project will definitely be late.

Another stark difference between CPM and CCPM is how resource contentions are handled.  CCPM considers these contentions when it calculates its own version of the Critical Path called the Critical Chain.  Unlike CPM, CCPM considers both task dependencies and resource contentions when developing the project plan.  This acknowledgement of resource contentions helps negate the multi-tasking problem we see with CPM.

CCPM also addresses the two behavioral phenomena spoken about earlier, Parkinson’s Law and the Student Syndrome.  CCPM uses a relay race mentality whereby as soon as a task is completed on the critical chain, the baton is immediately passed on to the next resource.  Because of this relay race mentality, unlike CPM, early finishes are passed on with the saved time deposited back into the project buffer.  Delays to the project are minimized because of the existence of the project buffer.

So what kind of results should you expect to see with CCPM?  When using CCPM it is not uncommon to observe a 30 to 50% reduction in the time required to complete a project compared to that same project using CPM.  On-time completion rates are typically 90% or greater and projects typically finish on scope and on or below budget.
Bob Sproull

No comments: