EVM in MSP

James,



I am not confused. Didn’t say I was. However, I think that I would understand Steve’s situation a bit better, and be able to offer some relevant contribution to the discussion, if he could add some further explanation.



Projects are about the achievement of results and outcomes.

Merely burning Duration, consuming Hours or spending money do not constitute achieving results, and if I was the client I wouldn’t be paying for those unless that was what the contract was for. In a cost-plus or T&M contract the agreement is something like that if the contractor spends $50 the client will pay him $100, say, but fixed price contracts are more common.



A small correction to your post: you mention "cost of $10K". In my example, this is the Price, ie what the client pays. Cost is what the contractor spends to produce something that the client will pay $10K for. Obviously, Cost < Price unless the contractor is in deep trouble.



I am not advocating any particular method (not yet).



If I say that x% can mean a % of the planned or baseline

Duration

Hours

Dollars

Words, bricks, m3 of concrete, or whatever..



I just want to know which one Steve is referring to.

Which one do you think he is referring to?

Which one do you think his client thinks he is referring to?



Reporting progress to management or the client as "x% complete" is ambiguous and uninformative and only a partial picture at best, and at worst is mis-representation, deliberate vagueness or a downright lie. It implies a calculation derived from a numerator and a denominator, and I would want to know what they are, that’s all.



This isn’t method, just facts and arithmetic.



Suppose x% applies to one of the 4 of Duration, Hours, Dollars or Words (or bricks etc). What about y%, z% and w% for the other three? They exist. They are facts. Do they get measured, calculated and reported?

Trevor,



I understand your confusion. However, the confusion originates from the fact that you haven’t defined the method of calculating progress. Strictly speaking, you should use only one method, otherwise the whole thing becomes total chaos. Your last sentence describes precisely how the confusion arises. You should use either DURATION, HOURS, or WORDS - but you CANNOT and MUST NOT try using all three. Moreover, if you take each method, you will see that you can get wildy differing results for the same point in time. However, assuming that, at the end of the day, you are still going to deliver the book within the 50 days, at a total cost of $10k, then the interim results are misleading.



I’ll try to provide a brief series of calcilations in order to show you how the results differ, and what interpretations and assumptions need to be made. Gorra get this Bid ready for today.



Cheers.



James

I am still not sure I understand what the project is intended to achieve, or what the basis of the bid and the contract are.

Suppose the contract/project is to write 10 chapters of a book. You estimate 10 direct hours of work per chapter = 100 hours. Perhaps the estimated duration is 5 days per chapter and they are all done sequentially, to a total 50 days. Suppose the hours cost $50/hour, ie $500 per milestone, total estimated cost $5000.

You have to make a profit and cover some overheads and indirect costs that haven’t yet been included in the plan, so you bid say $10000.



It seems to me that your bid, ie your offer, is to deliver the book for $10000, ie you bid a result and $, not a number of hours.



You could also tell the client that you have estimated 100 hours total or 10 hours per milestone. You could also tell them that you estimate you can deliver one milestone every 5 days. You could also tell the client the hourly cost rate. But you don’t have to reveal any of these.



Once you start the project, duration elapses, hours get logged, $ get spent and you achieve milestones. Some might take more or less than the estimated duration, hours and/or cost.



If you report that you are "x% complete against a particular milestone", are you saying:



x% of the 5 days duration has elapsed?

or

x% of the 10 hours of work have been clocked up?

or

x% of the estimated $500 budget has been spent?

or

x% of the words have been written?



While it would be unusual for them all to be the same at, say, 20%, it is not impossible to be right on plan/schedule/budget. They are all only loosely connected.



What if after 1 day (20%), 5 hours of work (50%) have been clocked up, $300 (60%) has been spent and 80% of the words are done? .... or any other combination.

Addendum:

Sorry folks, I have been reading this thread and think I’ve hit on a fundamental flaw in my problem statement. When we bid the project we bid a set of hours to complete with a bottom line cost ($$ not broken out by WBS).

Hours WERE bid against specific milestones and this has formed the basis of the basline establishment. Therefore, when I report to the customer that we are x% complete against a particular milestone for which we bid abc hours, they can get a picture of where we are at and how long we’ve spent getting there.

Its a small point but I think an important one.

Thanks Trevor,

I guess my point is that:

a. I want to manage by EV, so I need to be able to report to my management the state of play in true EV terms,

b. I want to provide EV performance to the customer.

c. I don’t want to run two sets of books. This is a labour resource project only so the only costs ARE labour costs.



I don’t have a problem with the EV aspects of the Project (I’ve been doing this stuff for a while) and I have several work-arounds that I’m exploring - just wanted to know if anyone knew of a way that MSP could generate S, P, and A in hours, like most other serious PM tools. As it apparently can’t I revert to more mandraulic processes.



Thanks for your help on this folks.

Steve,

I hope you don’t mind if I make a few observations..



When you start off intending to run transparent progress reporting for the client, the intention is immediately diluted by a need to use Hours instead of $ to obscure the cost. Why? Because the client is so narrow minded that they only see the difference between (only a part of) your cost and their price and don’t like it. Apparently. a client isn’t happy unless he sees evidence that a contractor is making a loss.



In a fixed price contract there is no automatic client entitlement to know anything about your costs. You are not obliged to show the costs, and you shouldn’t. If this is a fixed price contract, the $500K price was agreed at the start, probably after a competitive tender process. After that, the only party that cares about the cost is your outfit, not the client and he should know that he doesn’t have any reason to know your costs.



If you must show EV numbers to the client, just show SPI and CPI which are ratios. But will the client know whether they are good or bad or what they or you should do about them?



I am not in favour of isolating out just the hours component of cost and doing the EV calculation on those only. Obviously, it only produces a partial, distorted and very ambiguous picture because there are also Material Resource Costs, Fixed Task Costs and other indirect costs, all of which can and should also be accounted for in the MSP project plan. Also, different kinds of hours have different costs and they also have different "worth" to the project depending on which tasks they are associated with. Hours consumed to achieve important Tasks on the Critical Path are worth a lot more to the project wrt "progress" than hours on trivial, non-critical Tasks.



However, it is easy to substitute hours for cost by just setting the rate for everyone to $1/hour so $1 = 1 Hour on all EV reports and graphs such as from the Analyse Time-phased Data To EXCEL button, and easy to amend labels.

Hi Steve,

I’ve just had an “oh Doh” moment. If you’re supplying the customer with EV data – why not just give him a “picture” of the graph with the Y axis marker values “hidden”. Doing this means that he can see the ratios of the ACWP, BCWP and BCWS without you having to divulge any form of $ values. The fact that it is a “picture”, means that he cannot retrieve the Y axis marker values.



It’s intriguing that you say that, even when using the Physical % Complete option in MSP, it still calculates EV using the % Complete value. If that is genuinely the case, it then negates the whole idea of introducing the Physical % Complete option. I have a copy on my PC at home, so I’ll perform a few experiments and keep you informed. At work, we are still using MSP 2000, so we don’t get the option to even try anything.



Just to help clarify (and you probably already know this, so forgive me if this is a “granny-sucking-eggs” statement: BCWP (Earned Value) is calculated by the “Progress [same as % complete] x BCWS”. “Progress” [or % complete – whatever term you wish to use] is calculated using any data set that you care to imagine. Yes, that’s right – literally any unit of measurement can be used. It is entirely up to the project to define what it’ll be. The important thing is that people are aware of the units being used. When you use software to calculate the EV curves, you only have the options of the pre-defined methods that reside in the software. In most cases, it will probably be Schedule % Complete, Physical % Complete or Hours % Complete. All of them are perfectly valid. You will note, however, where you have the Physical % Complete, it is literally the input of a progress value that can be calculated in any way that you wish [usually calculated external to the software, and then the result is input back in]. This is where it gets interesting!!!!



As a very basic example: Produce a Report – and assume [therefore programme it] that it’ll take 40 hours over 5 days and working at 100% continuously from day 1, on the profile of a conventional working day. Therefore, you’d do 8 hours per-day for 5 days. In terms of the BCWS profile; it would be linear. Progress [or % complete] could be calculated using

1. Schedule % Complete,

2. Hours % Complete

3. Physical % Complete. If you use Physical, then the only real physical element of a report is the quantity of pages. However, to complicate matters further – many companies often “claim” a pre-agreed progress value based on the physical, discrete steps of the report production activity; such as. Start – Issue 1st Draft – Update – Issue Final. These are just the quantifiable methods. We have yet to include the “gut-feel” methodology of progress calculation.



Considering the above, it is quite straight-forward to work-out that you can get significantly differing BCWP values for any given status-date, depending upon which progress method you use. However, the equation itself is exactly the same ie. (Actual/Actual+Remaining). The problem is that, in order to calculate a physical % complete, you’ll probably need to maintain an external data-set – which is just then adding to the burden of project stewardship. Things can become hellishly complex as you start re-scheduling and re-profiling the activity. This is even before incurring any scope changes. Now if you multiply this over several hundred activities and a relatively long duration, you can imagine that it becomes virtually impossible to independently verify the true EV. Finally, if you add-in the fact that very few people really understand how EV works, will not have time to listen to anyone and probably wouldn’t understand it if they tried, you then have a recipe for continuous debate and mucho slagging-offo. Ultimately, however, you should still be able to answer the fundamental questions of:

1. When’s it going to finish?

2. How much is it going to cost?

3. Are we going to make a profit?



Cheers.



James.

HiJames,

Love a debate/discussion, especially on EVM, and its uses, abuses, and excuses.

Being a systematic simpleton I’d like to take your points in order and clarify my position or explain better where I’m coming from.

1. Other products vs MSP. I don’t doubt that the other high end tools have their own idiosynchrasies. Indeed OPP can be a "rheet bugger" at times and I sometimes feel like you need a post doctorate degree in scheduling and s/w tools to keep it fed and watered. Nonetheless, once its up and running, and being maintained, its a very robust and capable Project Management tool. To date I cannot say the same for MSP. I’m prepared to take a hit on user ignorance, but just recently I’ve had MSP do some pretty squirrely things to my schedule without telling me about it.

2. Hours vs Dollars. It is my intention to provide EV performance to my customer (v.laudable I hear you say) but this is a fixed price contract. I manage my resources internally by cost and let the bean counters make all the O/Head, G&A, Margin and Contingency adjustments outside of my PM sphere. So, and this is by way of example only, a job of work that may cost me $250,000 may get charged to the customer at $500,000. "Outrageous!", the customer says, "You are gouging me and making 100% profit." From this point on I have lost their trust in my professionalism and integrity. After all the other legitimate costs have been added to my direct labour, this effort may represent a cost to the company of $490,000, much closer to a 2% profit (better off just putting the cash in the bank). By only reporting labour resurces EV (and there is no material on this contract) I get to plan 2000hrs of labour effort and report accomplishment against that. If there was only one category of labour on the project the customer "could" work out the average labour charge out price on his contract, but I have subcontracted labour which attracts different overheads, and other effort which is charged out higher based on degree of R&D effort involved, face time with customer, etc etc. No longer is the straight line comparison of price vs resource quite so straight. All I want MSP to do for me is allow for BCWS, BCWP, ACWP in hrs. It has to be able to do this in order to provide the $$ figures. Oops, no it doesn’t because it doesn’t calculate EV based on EV inputs but on schedule %complete.

I too have noticed the Physical %Complete option in MSP, which in EV world would be equivalent to BCWP (Earned Value), but the few experiments I’ve tried with it have not been encouraging - the EV is still derived from %complete.



I’m not really "having a go" at MSP - it wouldn’t occupy the market position it does if it didn’t satisfy its user community needs - I just approach the whole planning and scheduling project controls activity with a great deal more trepidation knowing it is the tool I’m staking my project and customer relationship on.

Hi Steve,



I appreciate your sentiments on MSP and EV. However, MSP really isn’t as bad as some people make it out. Judging by many of the postings in this forum, and some of my own personal experience, Primavera has equally as many flaws and idiosyncratic behavioural patterns - some of which I consider to be appalling.



With regard to your specific statement about preserving commercial sensitivities, and displaying EVM in labour hours, it would take about ten seconds to estimate the $ value, merely by assuming an average hourly-rate. Moreover, in MSP, as maybe also in Primavera, you can select the activities on which you perform the EV – thus giving you the opportunity to select activities that contain only labour hours. In conclusion, you can actually perform EV on resources, as opposed to activities. This can be quite useful but, like any method of calculation, you have to be aware of precisely how it is doing it in order to feel totally comfortable with the results. In order to feel comfortable, you have to perform many experiments and understand how the individual variables of the EV equations affect the outcome. Even so: when applying the methodology across many hundreds of activities, you are unlikely to determine where errors of significance reside.



I’ve always maintained that, regardless of the method of calculation, EV results should never be perceived as being an absolute truth and numerical value. This is because there are so many airs-and-variations on its calculation methodology that it becomes impossible to ascribe it an inviolate value. It is, therefore, the TREND that is important. Consistency of values, within the expected norms of statistical variation, are of far greater benefit than attempting to analyse precisely why last month’s SPI value was 0.84 and this month’s is 0.81.



As a footnote, I know that MSP 2002/3 has the option of calculating EV using Physical % Complete. However, as I’ve not done any real experiments with this, I cannot compare the results.



Feel free to debate.



Cheers.



James.

Thanks James,

I’ve always had a beef with MSP and EVM. It stumps me that other products like Open Plan and Primavera can get it right but MSP continues to perpetuate bad EVM algorithms. The thing I like about OPP Professional is that I could choose whether I would represent EVM data in $ and thus capture all effort (labour, material, and ODC)under the project, or, if I wanted to preserve commercial sensitivities wrt $$ but still represent labour EVM I had that option as well. I guess thats why there’s such a price differential between the two products - one is a serious PM tool and the other is great for pretty graphs and some prelim planning effort.

Cheers,

Steve

Hi Steve,



EV is always represented in cost, as cost is the lowest common factor that will apply to both labour and Bought-out activities (purchasing). However, if you have only labour hours in your project, then you could just convert the cost into an average labour-cost-per-hour....once you have dumped the data into Excel.



I would proffer the warning that performing EV in MSP 2000 may give you some mildly unusual results, owing to the fact that it calculates it on the basis of DURATION % COMPLETE, as opposed to specifically the hours. Over a long period, and many activities, however, there may be little error or noticeable difference.



HTH



James.

Another question for MSP and EVM. Possibly a numpty question but How do you represent EVM data as per the EVM report but in hours vs dollars?

Hi, if you give your mail account I can send you some