There are so many flaws in the "Earned Hours" approach that I don’t know where to start.
But do you want to hear them, anyway?

I am glad we can rule out long distance extrapolation from small amounts of discrete data points.
I never suggested relying only on EV, although that was the only thing being discussed, so it is fair enough that you might have thought that I might be.
Ruling out long distance extrapolation, it is useful to know that some number was such-and-such last month and the one before and is something else this month.
Then it is a discussion about which numbers are most important, depending on what they are saying about what (hard to know for sure), and what might be done about it, if anything. Especially, you will want to see if anything that you did last month appears to have any effect on the numbers but often you can’t make a direct connection.
I had a look at some of your other posts and you seem to give EV very little credit. If you favour other approaches it is not necessary to slay EV, is it? Unless everyone is so hooked on EV that they won’t consider other approaches?

I think EV deserves a bit more credit, way more powerful than is generally appreciated. But then so is CPM. Both CPM and EV hardly ever get fully unleashed. Neither will work and both are subject to mis-interpretation and way worse if not applied very carefully and the necessary data cannot be obtained and crunched fast enough etc. Other approaches will suffer the same fate if the same kinds of implementation shortcomings prevail.

I think EV is under-estimated and also that more gets read into it than should be.

There is not enough appreciation of either the power or the limitations.

I am familar with PERT (optimistic/pessimistic etc) and tools such as @RISK, Monte Carlo analysis, estimate ranges to express confidence, rather than single numbers etc etc. A pre-requisite would have to be flawless CPM and EV implementation, which are rare and even resisted.

Ok, so I imagine that project cost success probability and project schedule success probability gets worse or improves, ie becomes a lower or higher number than it was. Sounds good. Does it cost anything to get these numbers and is the pay-off worth it?

So what? Is this an argument that knowing this is better than EV or sufficient without EV, and even if it is so then this is a reason to not do EV?

Of course the numbers, whatever they are, and especially whether they go up or down and how fast are all good numbers, and, provided the correct meaning can be extracted from what is happening, do describe different aspects of performance and can be used for informed decision-making.

How would you define:
success?
cost success?
schedule success?
project cost success probability?
project schedule success probability?

The future performance of the project does not depend necessarily on the past performance.
There is no reason why the future performance cannot be perfect, measured against the plan now, even though the past performance has been a shambles measured against the plan made at the beginning. Of course, if the ball is already dropped then it will be hard to pick it up again in the future. That would be expecting the future plan to make up for past mistakes and problems, so it is overloaded with performance expectations.

With measuring trends, well, I’ve got problems with em. At the beginning we only have a very small amount of past performance data to extrapolate from. eg if you are doing a 20 month project and have 1 month of actuals data which results in a single discrete data point and no idea of the shape and slope of the curve that got it there, how reliable or useful is an extrapolation out to 20 months?
Even 2 or 3 months of data doesn’t help much, especially if the numbers aren’t accumulated and plotted and trended continuously in the period between the discrete data points.
Reliable, mathematically rigorous extrapolation can be done when we are only extrapolating a little ahead of a known curve.
2 months of bad indicators might lead us to extrapolate from 2 early data points that next month some indicator will equal zero and the finish date and cost will be infinity or vice versa.
The future is not connected to the past, although 2 months of shambles does not auger well for the future.
Likewise I would not put my feet up and relax just because I had 2 months of good numbers.
What will happen depends on what we intend to do and how feasible that is and how closely we can come to actually doing it.
EAC is not measured from the trend of the BCWP or BCWS or even ACWP. It is the ACWP up to now (ie a fact) + the estimate to complete from now based on what we intend to do, assuming it is feasible.

Of course, EV, Variances, CPI and SPI are not the only way to control a project.
Or rather EV alone won’t do it.
But without EV nothing else will do it alone either.
And why not do EV?
All it takes is data, communication, education and organisation.
Unfortunately, these are exactly the things which pose such a challenge to the construction industry especially.
It also takes hardware and software but the cost of these is insignificant in comparison to the benefits that effective use can bring to the project.

To return to the original question “why bother?”, there seems to still be a question over whether it’s worth it, whether the benefits outweigh the costs. I find it hard to believe that this question is still considered worthy of an answer.

To use another analogy, all those instruments at the front end of the aircraft aren’t just there because the pilots like the pretty lights, and they aren’t there because someone did a cost-benefit analysis and they passed. They are essential to the operation of the aircraft.
I have been on aircraft which didn’t have all of the instruments and it didn’t prevent flight but they were all missed.

Software (and management systems) leverages the initial investment if it used productively. Good tradesmen don’t buy the cheapest tools in the hardware shop.

If EV is coming, if not already routine and established, then choosing not to adopt it will not stop it coming.

Any construction business that doesn’t adopt EV is in great danger from the competition that does. You would have to be very, very sure that there was no advantage in it, or that it is just a fad, or that no one can make it work, if these are the kinds of reasons for not adopting. And how can you tell without investigation and trying it out.

A few short years ago, no one would produce 3D CAD modeling even though they had the 3D CAD software.
All of the old “reasons” were there. “it’s an extra cost for the contractor”, “the clients won’t pay for it” ,“it’s a steep learning curve and a training overhead”. Now it is routine and early adopters did not get left behind in the dust, scrambling to catch up.

The construction business is a jungle. Darwinian natural selection will ultimately answer the “why bother?” question.

That’s “why bother?”.

Trevor,
I strongly disagree with your statement that Earned Value Analysis is the only way to control the project. There are other and more efficient ways. I think that trend analysis is much more useful for proper project management. Analyzing trends of project success probabilities or trends of deviations from the baseline targets you will discover project problems much earlier and will be able to apply necessary corrective actions in time. I already wrote that EVA data shall be thoroughly checked because they may lead to wrong conclusions. It does not take into account network logic, project risks, resources that caused problems. It is not safe to make conclusions based only on EVA results. It can be used only as supplementary tool.
Regards,
Vladimir

I will try to keep them shorter but we advancing on this topic on a wide front and you make many useful points that need to respond to.

I like the South Pole example because it illustrates how all projects are alike wrt scope definition, distribution of risk and responsibility, performance measurement, definitions and consequences of success and failure etc.

I looked back at the start of the thread and sure, ok, we’re off topic a bit. If the starting question is "why bother?" I say because it is the only way to control the project. It is not really a choice based on we like it or not, or it costs too much or is too much trouble or not, but how it is implemented, calculated, used etc makes all the difference in whether it pays off.

Re SPI and CPI "independence":
CPI = BCWP/ACWP
BCWP = CPI x ACWP

SPI = BCWP/BCWS
therefore SPI = CPI x ACWP/BCWS

Not very independent at all and actually at the core of the strength of EV.

I thank you for this excellent example of such a typical scenario and the detail of the discussion.
I have been writing up just such a scenario as an article/teaching exercise, so this helps a lot.
May I post the relevant draft parts as my answer and be excused for the lecture/text book tone?
Any input or correction humbly appreciated.

Our difference of opinion is over whether the 300 m3 is scope, whether it is additional scope, whether it costs more or takes more time or more Hours, and whether and/or when it is included in the EV calculation. We can just consider all possible combinations one at a time. Then we can say, if such-and-such is the case, then this is the answer. If some other such-and-such is the case, then this is the other answer. Then we can eliminate the cases which are not this case. If we make a mistake here, and choose a case which is not this case, then we will have an answer but it will not be applicable to our reality.

I think we can agree on the definition of the mathematical parts of the problem:

CPI = BCWP/ACWP
SPI = BCWP/BCWS

And BCWP, ACWP and BCWS are all measured UP TO the Status Date. Especially, BCWS is not measured up to the completion of the project. CPI and SPI are determined AT the Status Date.
The calculation hinges on the Status Date, relative to the overall Duration, which we have to assume since it is not provided in the example.
If the 500 m3 is switched for the 800 m3 this effectively changes the necessary assumption about the Status Date, relative to the overall Duration (I think).

Perhaps we can’t usually re-cast the baseline to accommodate the 300 m3 hole that we have discovered since the project started, until we have measured performance against the existing baseline. This seems unfair because we could say that the baseline is defective to the extent of what was left out, and better information has become available, but this is what happens when the baseline does not include all of the now emerging extra scope (and emerging extra cost, duration, material and work resources). This is one of the strongest possible arguments for a precise, complete estimate at the start, with the minimum possible changes of any sort, and especially minimum possible emergent scope.

I think the client would insist on measuring performance against the baseline plan that was handed over on day 1, at least until persuaded that it should be replaced (not easy). This is perhaps also a weakness in EV, which assumes, quite rightly since what else is possible, that the estimate is, in all respects, correct, which is impossible if the scope is not correct. The longer you keep on measuring performance against an estimate that is becoming more and more superseded by actual reality and the differences are large enough to mount up to become significant, the less meaningful all of the indicators become. It is not EV itself which prevents re-casting the plan and the mechanism exists for constant, frequent and continuous monitoring, updating, re-scheduling, re-budgeting and eventually when required re-baselining. Throw out the baseline and get a new one and start measuring again. This “constant, frequent and continuous” element aligns nicely with the basic principles of the Critical Path Method (everything moves all the time). If EV forces estimates to be complete and accurate then this is a strength of EV. But many Estimates are rough and incomplete and, amongst all of the pain that that will cause, The EVA won’t work.

Contracts (or the client’s lawyers), however, freeze the early baseline overall and (some say, although very few) even freezes the early baseline detail, and rigidly prevent the baseline from being changed or replaced, regardless of how much improvement is available or needed. Talk about a terrific way to choke your own project.
Legally, this makes sense because the objective is legal certainty.
But it does not assist the achievement of the project scope, time and cost objectives and completely contradicts the concepts, principles and methods of the CPM and EV which are the tools available to achieve the project scope, time and cost objectives.
It is lazy or uninformed contract drafting, too, because it should be possible to achieve both legal certainty and unleash the project control power of CPM and EV.
The contractor’s only defence, other than educating the client and his lawyers about the difference between artificial legal constructs and real world project management, is to ensure that the Planned/Estimate/Baseline plan is so good, so exact, so error free, that it can actually be worked to as-planned and that no significant changes will occur.

The contractor in the example has contracted to fill a hole with concrete, as you say. If this is a fixed price contract, rather than cost-plus reimbursable or schedule of rates, and if the contractor under-estimates the materials or quantity or Cost or Hours or Duration it’s his risk. After all, anyone can win a contract by under-estimating the scope and bidding low. That’s called buying yourself a world of trouble and pain.

I don’t feel comfortable using the words “approved” and “forecast” in the EV discussion.
I prefer sticking to the EV lexicon which has “budget” and “actual”.

This “forecast” 800 m3 looks to me like something you know exists, that you didn’t know about at the start, and has to be dealt with but is not yet part of the plan/schedule, either at the beginning of the project or now at the Status Date (whenever that is) when we have poured 200 m3 of concrete.

Planned Cost = Estimated Cost = Baseline Cost= “what we thought it would take to get the job done, no more and no less, given what we knew about the scope at the time we made the plan, regardless of what we know now”.

Estimates must drive approval, not vice versa.

The Estimate is the minimum amount that must get “approved” and provided for if this project is to start at all. Otherwise a guy (maybe you, not me) might be walking 500 km to the South Pole with a Planned/Estimated/Baseline cost = enough to buy 1 tonne of food but an “approved” cost = only enough to buy 0.5 tonne of food. What do you think will happen? And if you do start, and are doomed already, you had better not find another 300 km on the ice or on the map that you hadn’t noticed (or it makes no difference to the outcome anyway). If someone (not me and not the board of directors and not the client/sponsor) did start off to walk to the South Pole with half the food estimated to be needed (departing from project management to gambling), and didn’t find another 300 km and actually made it (on schedule) that would be a magnificent achievement. Or, more likely, it would tell us that there was something wrong with the estimate in the first place, such as how much energy it would take or how much energy could be squeezed out of 1 tonne of food. Next time we won’t ask for so much. Notice that the sponsor/client and the board of directors who approved only half of what was asked for, but didn’t do the walking, now takes some of the credit! If the walking guy doesn’t make it and starves in the snow he gets fired by his board of directors and they get sued by the sponsor/client or the LD clause kicks in.

So Approved = Planned/Estimated/Baseline or we aren’t walking anywhere snowy (or pouring any concrete).

Then he breaks a leg or runs into bad weather or finds 300 km that was or was not on the map.
Some of these things are bad luck that he’s not responsible for or things that he is responsible for.
Un-imagined bad weather or getting struck by a meteor are bad luck, beyond the reach of risk management and scope definition. The contract will provide relief. The sponsor will fly in supplies and aid to get the job finished according to a new plan or will fly him out, probably depending a lot on where the calamity occurs and how achievable the remainder is, and how it can be achieved.

The 300 km though is hard to talk his way out of. The Estimate was wrong (not just inaccurate or imprecise) because the maps were bad (who made the maps?) or the maps were good but the counting was bad (who did the counting?).
If you spring this surprise on the client he will fly in the necessary supplies if he must to get the job done. Or he may tell the walking guy’s board of directors to fly in the supplies at their expense, as he would be entitled to do if the contract was to walk to the pole for the contracted quantities of supplies and support. They can do it and attempt to argue against the contract later and recover from the client. If they can’t do it then the client can do it under duress and sue them later. In this contest, depending on who made the maps and who did the counting, and which one (or both) was wrong, if it wasn’t the client who did it I would be betting on the client.

It would be better for the Contractor if the walking guy was hit by a meteor (and survives) before the client finds out about the 300 km. Preferably, for the contractor, the client never finds out about the 300 km. Anyway, the arbitrator/claims consultant/analyst/forensic accountant can apportion the effects and the blame and reponsibility later, using rules about “concurrency”.

The discovered 300 m3 of concrete looks like nothing except an estimating blunder by the contractor, with all of the typical nightmare that goes with it. Here it comes. How do you break this news to the project manager, the board of directors, the shareholders and ultimately to the client? Who has to fix it? Heads will roll.

When the 300 m3 hole turns up, if it has cost, and/or materials, and/or work, and/or duration and/or schedule consequences then it is definitely scope If it is not scope because it does not have any of these consequences then we don’t have to consider it.

Actually, as an estimating blunder, it’s not the planner’s problem. It’s the estimating department’s problem. The planner just plans what he knows, what the estimating department tells him. Unfortunately, if the same person is the estimator and the planner, and perhaps also the project manager, board of directors and shareholder, then he definitely has an un-shiftable problem. The client will not be enthusiastic about the Contractor’s blunder becoming his problem, especially if he gets told about it late after a serious of glowing assurances and hyped-up progress reports.

Is the $12000 the best estimate of enough money to pour 500 m3 or 800 m3?
Is the Duration (not given) the best estimate of enough time to achieve 500 m3 or 800 m3?

The example leaves out some essential information, namely everything to do with the schedule/timing.
The example does not incude any consideration of SPI = BCWP/BCWS
CPI is just about meaningless without SPI as well.
EV is inextricably linked with schedule/timing and rates of consumption of dollars compared to rates of achievement of Tasks (not compared to the rate of consumption of hours or materials, neither of which is achievement or progress).
No schedule information essentially renders meaningful EV calculations impossible or else forces the adoption of a high risk assumption that planned and actual production has been achieved proportional to time elapsing at the planned production rate, and similarly that planned Cost is proportional to planned duration (pro-rata distribution of production and Cost over time).

So, if 25% of the Task (measured by concrete) has been achieved up to the Status Date, in the absence of any timing information, and, forced to make an assumption, mainly that the Task started at the planned ES and production rates have been exactly as planned, we can take this to imply that the Status Date is at 25% of planned Duration (whatever that is/was).
This is a really risky assumption because it so unlikely to be exactly correct, and if you are forced to make this on-site in real life because you don’t have the records of actual start/finish compared to planned start/finish, especially this early in the project, then you can expect to be losing control. First rule in number crunching: get the numbers.

Examples follow.

Trevor is right... you shall be considering budgeted quantities till this point. Any variance will be covered in actual quantities. Forecast has nothing to do with CPI at this time.

However if you feel this difference is so significant or it may be due to scope change/creep, you shall follow the change control mechanism to rebaseline your project. This is the only way to cater for forecasted quanitites in this example/situation.

Regards

Sorry,...me too

Rashid,

No, the CPI is still 0.75 because you haven’t accounted for the additional quantities required. Work accomplished is based on a forecast total of 800 m3 to install, not 500 m3.

In my example, I didn’t include a Forecast cost because it isn’t relevant for indicating the value of te work accomplished. Note that the Forecasted costs would include the actual costs, so it isn’t a good measure for calculating percent complete and subsequently, earned value.

Regards,

Ett

Trevor,

One of the key requirements of earned value is coming up with an accurate method for calculating the % complete.

To do this, you must first determine what "drives" the task/activity based on the nature of the work it represents. i.e. what is the best indicator of percent complete for this task? Is it quantities? hours? cost? level of effort? apportionment?

Once you’ve determine this, then you can apply the best method to calculate earned quantities, hours and costs.

E.g. Assume a task has the following data:

Description - Install concrete foundations
Approved Quantities - 500 m3
Forecast Quantities - 800 m3
Approved Hours - 2000
Approved Costs - $120,000

You’ve determined that the best method for % complete for this task should be based on quantities installed.


So, assume 200 m3 have been installed at an acutal cost of $40,000.
Percent complete = Actual Qty/Forecast Qty = 200/800 = 25%

Therefore, earned value calcs will be:
25% x 500 m3 = 125 m3 earned quantities
25% x 2000 hours = 500 earned hours
25% x 120,000 = $30,000 earned.

From there you can determine the CPI: $30,000/40,000 = 0.75
This means that for every $ spent on this task, you are only receiving 0.75 worth of work.

So, earned value can be applied using quantities or hours or any other indicator that best determines the work accomplished. It need not be based only on using costs.

Hi Naveed,

I got your point. Thank you.

Regards,

manu.

Hi Naveed,

Thank you for your input.

How can the BCWS be 100?

Manu.

Alex

In this example you just cited, the Cost performance index will be BCWP/ACWP = 25/70 = 0.35 . Since a value less than 1 is a negative indication, Don’t you think CPI may give a confusing indication even if the project is progressing as scheduled?

Manu.

I also agree with Alex, but Vladamir makes a good point.

Earned Value is a useful tool but it can be abused by project managers to give the perception that a project is running well, when in fact, thinks are not what they seem.

The reporting of Earned value components, such as CPI and SPI, can be influenced by having work done on high cost, non-critical tasks.

I think that when reporting CPI and SPI, there needs to also be some indication of total float.

Comments?

Alex,
I agree with you. Using EV as one of program performance measurement indicators is useful.
Regards,
Vladimir

Vladimir

EVMS is only one of the many tools/methodology should be use in a project.

I am still waiting for one tools fix all...

I guess in the mean time we have to use a series of tools and indicators to help manage a project.

I have to disagree with your statement - about the project manager try to play smart. A full, EVMS should includes a set of rules to goverance the progress of a activity, training to the manager, benchmark, change management, roll out methodology, audit... Of course if you think that all U need is a software tools without invest in Process / Procedure and People then I think any system will fail.

A schedule developed with no rules and standard will give you wrong information.

Hope that ease the concerns that you might have in EVMS.

Alex

Dear All

I been promoting EVMS in many organisation. And observed the followings:

Accountant - Cannot visualise what is Value Earned, they only know what is being spend... If you said your project is 50% ahead of progress, he will only complaint about your 30% overspend on your original cash flow...

Project Manager - They only want to know the end result "not the means", whether the project is going to finish on time on schedule, what is the trend, he will uses his 6 senses rather than a quantify figures??

Portfolio Manager - They want no of projects are currently active and no of them are forecasted a profit and ahead of schedule time, again end result only cannot see in between.

However when you ask them question like these they cannot give you an answer, like
**How is the overall performance of your portfolio? (A)How can I add them together they are all different type of projects.
**What is the physical % complete of the overall projects ? (A) How would I know they are all different activities I know I am ahead in procurement, and installation but late in T&C.....
** What is the progress trend of the project?? Well they spend the money in accordance with the cash flow, so the project is running great...

EVMS in my opinoin is a Tool to measure performance for all stakeholders Cients, PM, Accountant, Engineers, Site supervisor. And it is a critical tools in a project organisation.

Cheers

Alex

Vladimir

Agree!!
As I said EV indicators are only part of the KPIs in a projects. (One tool out of many)

However, "Dont rely on EV" I would resentence it to "Dont rely on EV only" IMO that is more true. Because EV have a lot of functionality and charactistic that other indicator dont have.

What I want to said is EVM/EVMS do have a lots of value dont overlook it’s power. Especially in a portfolio management area and that is why the US Defence develop this system as a tool to manage their portfolio

Cheers

Alex

I know alot of you think that EV is just a US OMB requirement as a result of frivolous expenditures on government contracts resulting in ridiculous cost data statements in the media like " 1 Hammer... $215, 1 Toilet Seat... $415"

Earned Value Management is an integrated approach to Project Management. It requires the definition of a Performance Measurement Baseline, a Change Control Process, reporting of cost and schedule data ( CPI, CV, SPI, SV) and using all of these to ensure that a contractor finishes the project on time and within his budget, for the scope that was documented. In a FFP contract, EV is actually beneficial for the contractor to ensure that changes are identified and additional cost and schedule is given to perform out of scope work, as well as to protect the profitability of the project. For CPFF or CPAF, it is beneficial for the client, to ensure that they are getting their monies worth. Validating progress is an essential aspect of the system, and this is the crutch of the whole thing. All prices should be broken down to the lowest possible quantity, allowing for simple calculations of installed quantity and thereby giving accurate installed cost data. This prevents the contractor from stating that more was installed then really was, and it pushes the owner ( or the prime in a FFP) to go out and take a look, be more hands on and ensure accurate quantities are being reported in invoices and schedules. Being an arm-chair quarterback gets you no-where these days, and if you are one, EV is certainly not for you.

Ett

"In fact, Earned Value is a set of guidelines that when implemented on a project results in an integrated project management approach that, when followed, integrates scope, cost, scheudule, physical progress and actual cost."

I wholeheartedly agree. In a previous incarnation working on a contract for MoD we were asked to implement an EVMS by them and following an independent baseline review actually moved to payment by earned value. The benefits were immense. We improved our project management process significantly. Scheduling improved. There were no problems getting accurate and timely progress reports from line. Estimates for work packages were realistic and supported by risk analysis. Cost control improved. Our understanding of the project status was very good. I could go on.

EV is a misunderstood animal. It is an excellent performance measurement tool and every planner should be thoroughly familiar with it and the subtleties of it.

Regards

Hi Jaco,

Well said, however, there is an out, someone is proposing the use of QS’s as planning engineers, if this happens, and all the estimators are fired as well as planning engineers, what are we going to do without any QS’s around. Who is going to count the beans?

Regards (Groete)
Philip

Earned Value = 50% job costing

-does not take into account production units but just percentage
-unit cost is intuitive, earned value terms are not
-contractors are interested in unit costs, a value meaninfull to them, it is their bread and butter
so they perform earned value + the other 50%, they just don’t use the terminology
-earned value is basic budgeting in a foreign language
-it is insane doing job costing with CPM software although they provide, or attempt to provide some functions
-true job costing must be integrated into accounting
-accounting cannot be determined by a CPM software although vendors will push for the sale
-I have no idea on who besides government (or worst pharmaceuticals) would be interested in earned value over job costing

Hi Jim,

Has this question anything to do with Vadim’s EVA? The reason I ask is that I had a query earlier this week about it, but cannot find any definitive information about it. If you know, more let me know where to find it. The problem I have is that in my thinking, Earned Value is different in planning than in costing, maybe just a sign of age:-)
Earned value in costing is fairly simple, however in planning I have always understood it to be the percentage of scheduled work completed, excluding any activities ahead of schedule. In other words, when you run a s-curve, you end up with actual/early/late and earned curves, and if you want, a scheduled or planned curve ( in the case of resource leveling). The reason being as you rightly say is to determine, how good is the planning. However, this can be done using costing info (performance factors) as well. If you use both methods, it could be a very handy tool for contractors. But as I say I am a bit unclear on the EVA methods you are talking about.
I think this is an interesting topic and deserves some discussion, and might be of value to most planners.

Regards

Philip