The problem here is that EVM is based on a detailed and accurate “plan” at the start of the initiative. And Agile avoids that detailed upfront planning because it usually represents waste and throwaway work.
In an agile project, the backlog and release plan may be a best guess based on high-level estimates of the work using imprecise measures like story points. So your actual results are highly likely to differ from the baseline you created. Remember the Agile Value statement, Responding to Change vs. Following a Plan?
If your results fall short of the plan, how can you possibly know if it is because the team did not perform well or that your original estimates were wrong? This comparison is back to reliance on plan-driven methods and assumes or is predicated on the ability to accurately forecast the future. Focusing on the variance is navel-gazing at best.
That was a nice piece by Anthony Mersino, “Enough Already, There is No “Value” in Earned Value Management.” I would go further to say that the planning work done ahead of a project is worse than “waste” or “throwaway work.” Indeed, the plan can become the objective, and can make it much harder for teams to pivot to the correct solution because that would entail all sorts of baseline change requests and other administrative effort. Here are some of my blog posts on EVMS vs. agile (here and here).
One important point that’s often overlooked is the value piece. Just because a team did what was planned and spent money doesn’t mean anything valuable was achieved. This is the entire point of Fredrick Brooks’ The Mythical Man-Month. While the amount of effort on an assembly line may be strongly correlated the amount of output (think of widgets), software and knowledge work don’t operate in that way in the slightest. The amount of progress and labor hours need not correlate at all due to training and communication. Brooks wrote:
“Cost does indeed vary as the product of the number of men and the number of months. Progress does not. Hence the man-month as a unit for measuring the size of a job is a dangerous and deceptive myth. It implies that men and months are interchangeable.”
Here’s an everyday illustration of the concept from Mersino:
Let’s assume that our initiative is to buy Christmas presents for my wife. I have a budget of $200 and I am going to get her 4 presents, one per week for the weeks leading up to Christmas. Each present will cost $50.
- Sweater ($50)
- Yoga Mat ($50)
- Gift Card ($50)
- Framed photo ($50)
… When my wife opens all the presents on Christmas, the “value” she places on the first 3 is almost nothing. The first one (sweater) actually had to be returned since it was the wrong size (my bad). The other two are just meh. But the 4th gift, the framed photo of her and I, this gift makes her cry, and cry. She hugs me. She takes the photo and puts it on a prominent place on her dresser and she looks at it every day.
The EVM calculations are about costs and waste, and they have absolutely nothing to do with value. Value is measured by the customer, and it is detached from the cost.
Very informative and well documented. Included the Section 809 Panel Recommendation on EVM and Agile.
https://discover.dtic.mil/wp-content/uploads/809-Panel-2019/Volume1/Recommendation_19.pdf
Thanks Larry! Here’s a thought I just put on Twitter… If EVMS is useful, and no commercial company uses it, then there is a HUGE profit opportunity. Start a tech company w/ EVMS built-in and then make a killing. The fact no one does this means the marketplace has resoundingly rejected EVMS.
Is that the ultimate trump card? What’s the best response to it? Perhaps that commercial companies aren’t as ambitious in deep tech developments as the government? But then why is the government saying it’s being more than outpaced by commercial industry?