Glen B. Alleman was kind enough to respond to one of my recent posts, taking exception to my assertion that a simple, calculated estimate at completion (EAC) is probably the most powerful underused tool at the project manager's disposal.
He laid out a scenario where a varied time-phased budget (Budgeted Cost of Work Scheduled, or BCWS) would introduce a sufficient level of error into the calculated EAC so as to significantly reduce its effectiveness.
I must disagree, if, for no other reason, than the time-phased budget isn't part of the calculation.
As I discussed in that earlier post, the classic cost performance index divided into the budget at completion formula can be algebraically simplified to dividing cumulative actuals by the estimated percent complete. With those two data points, remarkably accurate management information can be gleaned and used to avoid project catastrophe.
To support my argument, I would like to ask: What other options are out there? What other methods can rival the simple earned value management (EVM) version?
After suffering through semesters and semesters of accounting, and semesters and semesters of finance, I can say with a fair degree of certainty that there's nothing in the accountant's toolbox that can even come close to the accuracy of the simple EVM method.
As counter-intuitive as this may sound, the person who has been through a 40-hour EVM class is in a far superior position to relay accurate at-completion project costs than is the accountant, who, of course, needed years and years of post-secondary education.
I intend to speak on this at the EVM World 2009 conference being put on by the PMI College of Performance Management this week. If you're in attendance, look me up. Otherwise, leave a comment. I'll see it, I promise.
He laid out a scenario where a varied time-phased budget (Budgeted Cost of Work Scheduled, or BCWS) would introduce a sufficient level of error into the calculated EAC so as to significantly reduce its effectiveness.
I must disagree, if, for no other reason, than the time-phased budget isn't part of the calculation.
As I discussed in that earlier post, the classic cost performance index divided into the budget at completion formula can be algebraically simplified to dividing cumulative actuals by the estimated percent complete. With those two data points, remarkably accurate management information can be gleaned and used to avoid project catastrophe.
To support my argument, I would like to ask: What other options are out there? What other methods can rival the simple earned value management (EVM) version?
After suffering through semesters and semesters of accounting, and semesters and semesters of finance, I can say with a fair degree of certainty that there's nothing in the accountant's toolbox that can even come close to the accuracy of the simple EVM method.
As counter-intuitive as this may sound, the person who has been through a 40-hour EVM class is in a far superior position to relay accurate at-completion project costs than is the accountant, who, of course, needed years and years of post-secondary education.
I intend to speak on this at the EVM World 2009 conference being put on by the PMI College of Performance Management this week. If you're in attendance, look me up. Otherwise, leave a comment. I'll see it, I promise.
EVM is an excellent methodology available for control and like every other method, but I am trying to find out what the limitations are? I will appreciate any information on this.
Some areas where there are limitations that I have faced:
1. If the final budget is unknown, as mentioned by Octavio, but there is an added scope.
2. In terms of sensitivity—is an SPI of 0.8 any better than an SPI of 0.75?
3.If the expectation of accuracy is +/- 10% is it reasonable to forecast a EAC when CPI is 0.8
Hi, friends,
I´ll hope obtain the PMP Certification on July. I´m working in a defence company in this moment, and i´m the project plannig and controlling of IR&D Projects.This projects have a lot of characteristics (Sorry, I´m Spanish and it´s possible that I write some letter wrong), and it´s very difficult to applicate the EVM Method because:
1- Normaly, this projects have a pluriannual duration
2- The Spanish Dod, have a peculiar form to contract the projects. The total budget of the contract is sign in the same year, and normally, the schedule of the contracts is for many years. On contracts exist a price scalation formula but it´s very difficult to applicate this on the EVM method, because predice the individual rates of the price scalation formula (Energy,steel and labor)it´s impossible.
3-to evitate this problem, I deflat the Actual costs respect the sing year base, but this is very difficult to Know, if the results are corrects.
Do you understand me? If yes I´ll follow write other day.
Many thanks
I tend to agree with Glenn's assertion that the time for EV light has come. I manage commercial software implementations and routinely develop custom code during the project. The traditional EV requires a level of maintenance that my customers are unwilling to pay for but maintain the desire to understand the EAC and current Estimate to Complete (ETC).
As a project manager for a software vendor, I am only responsible for costs on my resources which simplifies the problem so I typically use the BCWS - ACWP to calculate the EAC, then maintain schedule. This enables me to simplify the concepts of EV to my customers into a scheduling exercise. If we stay on schedule, we stay on budget.
In years past, I did use traditional EVM effectively on 8 and 9 figure Government projects and believe the appropriate place for traditional EVM are these very large projects.
Michael,
I sense we're talking about two sides of different coins. The "quick estimate" of EAC you suggest in principle is not a problem. But it is just that a "quick estimate."
Executing the program on a month to month basis, and in our domain weekly Earned Value, with a mid-flash report formally sent to the client (NASA) - quick estimates are not what is needed for contract execution.
After attending EVM World as well and listening to those speaking about simplifying the EV discipline, I was struck that we're headed to two different directions.
Simplification on simple projects is great for making those back of the envelope assessments of "will this project ever come in on time or on budget." Great stuff. Keep it going.
For the DCMA compliant programs - we are working 2 at the moment - the formality of EV and associated probabilistic asessment are in fact "heavy lifting." And as it should be. When spending billions of tax payer money, the PP&C activities should be scaled appropriatley.
This reminds me of the "agile" development process early on. Where "on size fits all," and it's the agile size. Maybe the "light weight EV," is a concept who's time has come.
But like the agile discussion, a context and domain are needed before any meanigful conversation can take place.
EVM is just one method, used to predict EAC, and, depending of our "thouhgts" about "possible" future workforce behaviour, shows 4 different "possible" outcomes. In that case, we can say that EVM is the method that shows more "range" of possible future outcomes for EAC, that any deterministic value - as all other methods, it is only prediction.
That predicted range is preety accurate when used in conjuction with Critical Path Method, and in case of project that have only one Critical Path that can't be altered (one long critical path without integration points and braces integrating to that critical path).
In case of projects that have integration points along critical path, or, worse, that have integration points on final milestone (parallel braces), or in case of integration of multiple projects in Portfolio, EVM and related EAC fails badly.
In accordance with my experience and reality, my opinion is that CPM and EVM are old, incorrectly based and questionable - usefull tehniques.