Practical Project Metrics II
|
from the President | I'm Going Social | Featured Article: Practical Project Metrics II |
|||
I begin with the premise that the purpose of metrics is improvement based on truth. Specifically, we need to accurately and truthfully determine where we are in the project, where we're going, what works and what we should improve.
We recognize that projects are made up of work. Project teams perform a series of tasks in order to build products or perform services. If we define the tasks correctly and execute them correctly, the project runs well. Fortunately, project management offers us a well-established tool to identify progress, predict future progress, and identify where problems exist for these tasks: Earned Value Management (EVM).
Also, we recognize that despite all our planning, changes will occur. Some changes are inevitable and unpredictable. We live in a dynamic universe and unanticipated changes occur both outside and inside the project. Other changes could have been avoided through better planning and execution. It is these changes that help us understand where we can improve. Fortunately, project management also offers us a well-established tool to monitor these changes: the change management system.
Our challenge, then, is to discover ways to exploit these two subjects to create a suite to metrics that achieve our goal: determine status, plan the future, and improve project management in ourselves and our organization.
If you’re not familiar with the basic concepts of EVM, please visit our web site and locate the on-line version of the this newsletter. You’ll find a brief tutorial on basic EVM.
If you've studied EVM, at this point I'm going to ask you to forget everything you've learned. Forget CPI and TCPI. Forget the “S” curve plot of variances. At best, they confuse your stakeholders.
EVM Control Chart
Instead, plot the percent variances of cost and schedule as shown below. It's simpler and more understandable to executives and other key stakeholders: Zero percent variance means on-time, on-budget. A vice president can understand being 10% behind schedule and 5% under budget. Telling your vice presidents that your SPI is 0.9 and CPI is 1.05 will only anger them.
Now, add control limits. I've put these at ± 10% and ± 18%. You'll need to determine where to put your own. The control limits determine if the project “okay” (within allowable limits), “Warning” (is getting into trouble), or “Danger” (in trouble).
If you're using the stop-light status color scheme, this also gives you quantifiable basis for green (okay), yellow (warning) and red (in trouble).
When the percent variance crosses the inner control limit, the project's status goes from okay (green) to warning (yellow). This forces the project management team to enact some problem-solving action and forces them to inform management that the project's overall status has changed. When one of the variances crosses the outer control limit, the project management team is now forced to escalate the issue to senior management.
Below is a simple plot of schedule and cost variance percent. Note that the project is doing well for a while, then takes a steep downturn.
Trend analysis calls this the “knee”. If you want to discover the root cause for the down turn, look just before the knee. The problem may not have occurred there, but it did manifest itself there. In the graph below, the knee occurs in period 6.
Trend “Knee”
One advantage to EVM is the ability to roll up multiple projects. The graph below is a histogram showing the status of all projects in a portfolio. The bar height is the number of projects within ±5% of the variance shown (the center bars show the number of projects from -5% schedule or cost variance and +5% schedule or cost variance. Note that the cost variance bars form a bell-shaped curve indicating a normal distribution. The schedule variance is weighted on the negative side indicating that the project managers are good at cost estimating and management but need improvement in schedule estimating and management.
Portfolio EVM Variances
Another advantage to EVM is the ability to decompose the numbers anyway you wish. The graph below shows EVM number for a project broken down by four categories: internal staff, contractors, outside vendors, and capital expenditures. The graph could should either schedule or cost variance.
When I developed the number for this graph, I made sure the composite EVM showed zero (0%) percent variance for all months. Note, however, that the internal staff's percent variance is negative, indicating that they are falling behind. Further note that the contractor's percent variance is positive to compensate. This means the internal staff is not doing their job and project manager has to depend on the contractors to get this project done.
Also note EVM allows you to roll up the decomposed numbers across a program or portfolio. The graph above could just as easily represent the entire portfolio as it could a single project.
Decomposed EVM
EVM need not be complicated nor confusing. A simple, practical implementation will offer excellent communication, root cause analysis, cultural analysis.
Keep in mind: