Concept: Over the past decade, Earned Schedule has been elaborated, empirically verified, and adopted by diverse industries. During the same period, the Agile framework has become widely used. Earned Schedule for Agile (AgileES) combines the speed and responsiveness of Agile with the accuracy and control of Earned Schedule.
Figure 1
Practice: A couple of years ago, I shared one of my articles* with an Agilist colleague. The article described how I had successfully applied ES to Agile projects. I thought he might want to use the approach on his project.
The paper was returned with the comment: “Even if you get the EVM numbers, is it not more important to actually produce ‘useful deliverables’?” After all, my colleague argued, Agile prefers people and interactions over processes and plans. Why think that EVM is even relevant to Agile?
I took the question seriously. The practical advantage I had observed might not apply broadly to Agile projects. The benefit might have been a happy coincidence. I wondered if there was proof that ES applies to all Agile projects.
Although tempting, the approach portrayed in Figure 1 was not my next step. Fortunately, I unearthed research that conclusively demonstrates the relevance of EVM to Agile projects.** The argument is ingenious: EVM cost metrics follow mathematically from Agile principles. So, EVM must be valid for Agile projects. The results I observed were not random. They were a corollary of Agile’s intrinsic logic.
Briefly, here is the proof posed in the research. (Brace yourself for a few equations.)
A project’s Release Date equals the Start Date plus an offset. The offset is the Sprint Length (L) times the Number of Sprints (N). That is,
RD = SD + L * N.
The Number of Sprints equals the current sprint number (n) times a performance factor. The performance factor is 1 divided by the Percent Complete***:
RD = SD + L *(n * (1/Percent Complete)).
Now, here is the tie to Agile: Percent Complete is defined in terms from Agile Burndown. It is the number of Release Points Completed (RPC) versus the number of Planned Release Points (PRP). That is,
Percent Complete = RPC/PRP.
Introducing EVM cost terms, the researchers deduce that the number of remaining sprints equals the current sprint number times a performance factor equivalent to the one above. This factor is the ratio between the Estimate at Completion and the Actual Cost:
RD = SD + L * (n * (EAC/AC)).
Thus, the EVM Release Date calculation follows mathematically from Agile principles and equals the Release Date from Agile Burndown.
The researchers then applied the Burndown and EVM Release Date equations to some test projects. Results from the two equations were identical, and the math was confirmed empirically.
Leveraging this approach, I demonstrated the relevance of ES to Agile. The full proof is lengthy (for details, click here), but a summary follows.
Start with the same base equation: the Release Date equals the Start Date plus an offset. Again, the offset is sprint length times the expected number of sprints, and the expected number of sprints equals the current sprint number times a performance factor.
The performance factor in this case is the ratio between the Planned Duration at the mean Velocity and the Earned Duration at the mean Velocity.**** That is,
RD = SD + L * (n * (PDv/EDv)).
Given a standard rate for Release Points, the velocity durations are equivalent to ES terms for Planned Duration and Earned Schedule:
RD = SD + L * (n * (PDes/ES)).
Using ES definitions of Estimate at Completion for time (EACt) and Actual Time (AT), it follows that the performance factor is the ratio between the Estimate at Completion for time and the Actual Time. The resulting equation’s similarity to the cost equation is striking:
RD = SD + L * (n * (EACt/AT)).
Thus, the calculation of Release Date is the same for Agile Burndown and ES metrics. To validate the math, I applied the Release Date equations to two Agile projects. The results from all equations were identical. To see graphs of these results, click here.
Conclusion: Practice is a matter of experience. Some Agilists claim that their experience invalidates practices like ES. Experience, though valuable, is not conclusive. Rather than countering one subjective claim with another, we need an objective approach.
To prove the validity of ES for Agile projects, therefore, Earned Schedule metrics were deduced from Agile metrics. If experience suggests that ES is not valid for Agile projects, that must be due to problems implementing the practice, not to conceptual gaps between ES and Agile.
References:
*Van De Velde, R. (2014). Earned Schedule for Agile Projects. The Measurable News, 1, 2935.
**Sulaiman, T., Barton, B., & Blackburn, T. (2006). AgileEVM—Earned Value Management in Scrum Projects. Agile '06: Proceedings of the Conference on AGILE 2006 (pp. 716). IEEE Computer Society.
Notes:
***This may be easier to understand if put as follows. The expected number of sprints equals the ratio between the current sprint number and the percentage of work completed. So, if the current sprint is number 5 and the work is 50% complete, the expected number of sprints is 10.
****As a performance factor, you might think EDv/PDv sounds better—the earned portion over the planned total. But, keep in mind that the Burndown equation used 1/Percent Complete as a performance factor. Because Percent Complete equals RPC/PRP, it is actually the inverse, PRP/RPC, that sounds better. For the derivation, see the full proof
