For the most part, Earned Value Techniques (EVTs) on work packages are self explanatory; 50-50, 0-100, Milestone and so on. But a couple of them don’t behave as one might expect as far as how they calculate the earned value. Percent Complete, for instance, is typically used for work packages that span more than three reporting periods.
These longer % complete work packages lend themselves nicely to resource profiling, which is great for your FTE numbers, but surprisingly has absolutely no affect on the way value is earned. Now that statement may drum up some disagreements, but here’s what I got from testing profiled % complete work packages, and as yet, have not found a setting that will cause Deltek Cobra to calculate these any other way – which, if you read on makes a lot of sense, even if at first it seems wrong.
The calculation for % complete EVT work packages is a simple EV = BAC * % Complete. So if I have 1000 hours total for the work package, and the % complete value is 75%, my earned value is 750 hours; regardless of the period by period values in the resource profile. Check out this simple test example.
Above – I’ve loaded the resource over a ten month bell-curve on a % Complete work package. However, if I’m expecting to earn more value in the middle, I’m going to be disappointed. Look at what value I’ve earned after four months and am nicely on track at 40% complete.
If I was hoping to earn value based upon my resource profile, I should be at 500 hours. But going back to the math for % complete work packages (EV = BAC * % Complete) I’ve only racked up 400 hours of value. Of course that’s completely correct – Deltek Cobra worked as advertised. But it does illustrate the point that one’s expectations for value earned on a profiled % complete work package have nothing to do with the cumulative budget hours for the period you are in, and everything to do with the % complete value you’ve reported.
I also snuck in a Level of Effort (LOE) work package with an identical profile for comparison. That puppy has earned value based on the profile and is reporting 500 hours. Makes sense right – how else can LOE packages earn? It’s entirely based upon where the status date is and the cumulative planned value in those prior periods because there’s no underlying percentage value to calculate.
Profiling the resource in a % complete work package is terrific planning practice for your CPR format 4s and staffing reports, but is a complete waste of time if you’re wanting to control when you earn value. Your only hope is that having more people working on it during some specific period will accelerate your % complete during that time and thus drive your EV higher. Unfortunately if hope was a driving force in the universe, all our projects would come in on time and under budget and we wouldn’t need EV systems in the first place.
This brings me to my last point about % complete work packages, and one that warrants a blog all its own; you shouldn’t use too many of these in an EV system. In fact, none whatsoever would be optimum. Their inherent problem is that they are subjective. How do you know you’re 40% complete? Prove it! I’ll leave that discussion for another day.