Voices on Project Management offers insights, tips, advice and personal stories from project managers in different regions and industries. The goal is to get you thinking, and spark a discussion. So, if you read something that you agree with--or even disagree with--leave a comment.
Experts on the assassination of U.S. President John F. Kennedy have come to a couple of conclusions: 1) the Warren Commission got it right and 2) many people have a hard time accepting that such a monumental, history-changing act wasn't the result of a massive, expensive, difficult-to-execute conspiracy.
So, when I started writing about how earned value management systems (EVMS) can accurately predict the future with some simple calculations, I received some responses that expressed varying levels of incredulity. It simply goes against intuition that an information element as important as at-completion project costs could quickly and easily fall out of an EVMS.
But since I've already firmly ensconced myself at the end of this limb, let me take it a step further: The estimate at completion (EAC)--the brass ring of management information systems--can be calculated without a baseline, a work breakdown structure or a formal change-control process--none of what we've been told are essential parts of an acceptable EVMS. None. Nada. Zilch.
Now, I'm well-aware that the previous sentence is the metaphorical equivalent of pulling the pin on a grenade and rolling onto the floor of a conference room full of EVMS experts, but I can explain.
The traditional formula for calculating an EAC (EAC = BAC/CPI) can be algebraically reduced to dividing cumulative actual costs by cumulative percent complete. That's right, we're talking two date elements, easily collected. And the resulting information is far, far more accurate than anything that the general ledger can produce. It's also much more accurate (and faster and easier) than re-estimating the remaining work and adding that to cumulative actual costs.
In fact, it's so much more accurate, faster and easier than any competing information stream, that I'm frankly flummoxed that the calculated EAC isn't the centerpiece of EVMS use everywhere.
I got your point; EAC=AC/%complete.?. Why not?..
But from my own experience, calculating the EAC is not as complecated as that in any cases... the most complecated part is to estimate that bloody %Complete!
So we come back at the generic pitfall of EVM, having a robust plan, including defined ways of measurement for that %complete. There's so many kind of activities, and so many ways to measure it... In order to collecte that information, you will have to properly establish your deliverables, then your WBS, and then the way you will measure your progress for each deliverable or activity among the others in the global framework of your project which means having defined how that framework looks like (network and timelines)... In other terms, you will need a strong Planning Process... so hard to obtain, I agree...
No way to do without, if you wanna have an idea of your project status, EVM or not, you have to plan...
You may remember few years ago a quotation from a paper on planning "Failing to plan is planning to fail"
In effect you appear to be saying that you can calculate an estimate of the projects completion costs by assuming the same level of "efficiency" going forward as has been achieved to date - fair enough. And this can be done without a baseline, WBS or change control process - yep. But why?
Irrespective of whether a project would benefit from formal EVMS processes, the use of a baseline, WBS and change control process is surely good practice on every project.
I don't know about others, but I wouldn't even contemplate running a project for most customers I deal with unless a formal change control process was in place and agreed, and that relies on a baseline to change away from. As for a WBS, I don't believe I'm infallible enough to develop the full project definition without the structure and cross-checking that this provides. So yes, I agree, you could calculate an EAC without the project elements you mention, but the chances of having a successful project without those same elements approaches zero (but at least you'd have an indication of how much you might overspend by).
incredulity - The state or quality of being incredulous; disbelief.
ensconced - To settle (oneself) securely or comfortably
flummoxed - To confuse; perplex.
Mr. Hatfield's point is right on. My experience has been that once you're in the middle of a project, you will track reasonably close to your CPI through completion. Based on this, his EAC calculation is both quick and accurate.
I've seen bottom-up EAC calculations based on all sorts of "change" assumptions, and rarely do projects shift their CPI significantly.
This simple EAC calculation method has one risk, however. It depends on the accuracy of percent complete. If project tasks are being measured based on quantitative EV methods, overall percent complete should be reasonably accurate. If the project uses simple percent complete or other "management judgement" EV measurement methods at the task level, percent complete accuracy and the resulting EAC will suffer.
Mark, this sounds so easy that I almost (almost!) want to believe it. The problem is the sloppiness of the term "percent complete."
This, by some, is taken as percent of the schedule used up, cost encumbered, or even effort applied (we're half done!) and none are often accurate as to the actual progress made.
If you use earned value as the percent complete, you may be on the right track.
Easier said than done. The problem is since there is no standard methodology to calculate the percent complete, project managers in the same organization might calculate different EACs for the same project. Wouldn't this be an issue?
Michael, thanks for the thoughtful, clear and concise piece on EAC. Definitely it is going to provoke comments!
The simplicity of it is seductive. And in many cases I'm sure it would work very well. But to be most useful, its limitations should be understood. It assumes that the remaining work will require an proportionally equivalent amount of resources.
In my field of expertise, building construction, the work at the front end of the construction process, namely excavation and foundations, generally presents the highest level of risk and variability. You never really know what you're going to encounter when you start digging a hole. So, excavation and foundations are the most likely part of a project to see cost overruns.
Using your system, rather than re-estimating remaining work, would overestimate cost at completion, because the remaining work is far less risky and variable than the earlier parts of the work.
Of course, there are risk management practices that can address the early problems. But that's another subject.
I just wanted to point out that your system is very useful SO LONG as its limiting assumptions are recognized. In the end, this applies to EVM altogether. The calculations are all useful, but only so long as their inherent assumptions are understood. It's always a mistake to apply a mathematical formula, BLINDLY, to an area of human endeavor so complex as a large project. Proceed with eyes open and all will be well.
Antonio Bettencourt, PMP
Robert R. Morel
Oh wow! - "I could of had a V8" - the simpler more common sense methods are the most frequently overlooked.
Robert R. Morel
Morel Consulting and Sales LLC
It was quite historic to read about a piece of American history and EAC together.
Although there are some formulaes that experts provide, I would think that we would still need some form of human intervention here.
Let's say that the Project Manager could provide the estimate at completion (EAC). After having a first hand understanding of the project status, the re-baseline time and costing and forecasts forward.
Although I do not really understand your detailed explanation here at the fifth paragraph.
Good day on this Sunday.
Your method assumes that costs will be incurred in the future and until the completion of the task at the identical rate that was incurred job to date. That is only true in a very small portion of job tasks, such as rent or leases. Many tasks are either front loaded (incurring more costs up front) or back loaded (incurring a disproportionate amount of costs near the completion of the task). The task may include a finite number of purchases. You must know if these purchases have or have not been made or if your buy out was over or under budget. Your method is what I would use if I just didn't know anything about the task and was guessing. It is the easy way out because it does not require in depth knowledge of the task.
I've been on many programs where your simpler approach should work - especially extended execution (vice development) actions where there is a defined, usually limited, level of effort allowed and an equally well constrained period of performance. And heaven knows I pray for simpler approaches to everything so that we can put the money into program goals rather than the management.
But applicability really depends on the situation and like a good chef a PM best know what he/she is doing with the ingredients of the meal. My current program with a CPI of 1.25 and 47 % complete looks great until I do a bottoms up EAC. Then the unplanned rework, fielding efforts and previously uncommunicated customer expectations result in a very troublesome end-game challenge.
I'd like to use your simplified EAC approach and I expect that it is both a good alternative measure and useful corroborative data when EACs are calculated the normal way. But I submit that the key to its effectiveness is the same key to EV as a whole - accurate forecasting of the complete tasks and efforts / costs associated with them. That artful portion of the process continues to present the real challenge.
How do you handle people manipulating %complete to show a better calculated EAC?
As an active practioner of EV I've repeatedly informed others in my organization that the most difficult part of establishing and tracking earned value is collecting actual costs. The second most difficult part is correctly evaluating task completion, especially for partially completed tasks.
I agree, your formula for "simplified" EAC makes sense - but it depends on collecting these two critical pieces of information. If you do that it would appear you've already done 90% of the work required to implement a full blown EVM system, lacking only a baseline.
Hmmm. I think that Michael's shortcut depends heavily on how work is executed in a project. Of the various types of work (e.g., design, civil, mechanical, rigging, structural, architectural, etc.)on one of my projects, I think that every type of work would need to be represented in the actuals, for the resulting [actuals/% complete] factor to have any overall meaning. In construction projects, since work types are consecutive, this would mean that only near the end of the job would a meaningful EAC be calculable. That's way too late for me to take corrective action based on EAC updates.
Carlos Eduardo Martins Serra
This explanation is very easy to understand and is very usefull to explain how the concept of EVM is built.
But is importante to highlight that it can be truth only if we have a linear relationship between cost and % complete.
Although, it isn't truth in a little bit more complex environment when we have to handle with normal situations like diferent hourly rates and an indirect relationship between actual cost and % complete. In the article's kind of situation normally the project manager could know the phisical % complete. But the scenario that he's considering when says that status is not necessairlly directly related to the financial aspects.
So, before start to apply the formula we need to be very specific when we ask how much complete is the project to understand exactly about what kind of variable - % complete - each one is talking. Otherwise we could calculate an EAC very different of the one that we'll have at the conclusion of the project.
Is there specific cost elements that should be included/excluded? When looking at the bottom line, is there a possiblity you will be including some fixed costs that could inflate your EAC.
I find it very interesting. The principal issue I can see on everything we've been told to do in general, is that you have to follow many steps to get accurate info and sometimes, mostly on short projects, you just need to simplify and get the data you need fast and easy. That's why I like resources like this one.
I tend to agree with you. I work in an environment where we are project management challenged, but working on it to introduce a discipline. Our folks are putting status reporting in their WBS. I showed someone how to use MS Project to calculate EVM and then I also created a template in excel to do just the cummulative measurement you discuss. What a difference!
I understand your position, and really I think it is possible in a theoric view point. BUT
In PMBOK 4th Ed. the AEC is:
EAC = AC + (BAC - EV)/ (CPIxSPI)
So, now would be impossible to do that you say and in my experience (construction projects) the EAC is diferent if it is calculated from item line level or using summary numbers, or calculating % complete from Price or Quantity.
Henry LeÃ³n, PMP
a Peruvian working in Chile
I have been working with a multitude of projects in various organisations.
Earned value is simply like the game of cricket, it is all about run rates, how many over you have remaining to the end of the game will determine how many runs you need to make per over to beat your competitors score.
Be it as it may, at least runs are easier to count, whereas % complete is not so easy.
A few simple rules for % complete, firstly you need to be clear how you want to measure it, and there are 3 options;
1. Physical % complete- based on objective review or inspection of actual work, even better to break an activity or task into weighted steps that all contribute to the physical % complete.
2. Duration % complete - simply based on planned duration and how much of that has elapsed.
3. Units % complete (volumetric) - for example a task consists of laying 100 bricks, if 10 bricks have been laid then it is 10% complete.
You will have many projects that could have a combination of any of these activities, and need to be measured as such.
Earned value is really a simple thing, but proper EVA requires a baseline, in other words a performance benchmark, a key part of planning a project is always baselining the schedule when you have the first plan finalized, once it is in execution you must be able to compare your existing performance against your original plan.
The simple mathematical calculation would be OK to use as suggested but must be seen in the context of it's potential limitations and does not tell me about my required run rate to make the score, in other words it does not tell my where I should be on my project vs. where I am now in reality.
There are many other elements that make % complete a whole separate discussion but hope this provides a good idea of some of the possinilities.