EVM = EVM
earned value management results in early visibility and management opportunities
Harry Sparrow, Vice President, Performance Management Associates, Inc.
Earned Value Management (EVM) is an “early warning” Project Management tool that enables managers to identify and control problems before they become irrecoverable. Earned Value is an enhancement over traditional accounting oriented progress measures. Those methods only compare planned expenditures with how much has been spent. Earned Value Management goes one step further and provides an objective indication of actual accomplishment. This assessment of actual accomplishment and its translation into a metric called Earned Value, gives managers greater insight into both progress and potential risk areas. It also provides a foundation for more accurate estimates of projected completion costs.
Earned Value Management is more than a metric, however. It is the combination of robust planning and integrated management system elements (Gary Christle OUSD). For the benefits of Earned Value Management to be fully realized, thorough planning combined with the establishment of and disciplined maintenance of a baseline for performance measurement are needed. The combination of advance planning, baseline maintenance and schedule, cost and technical performance analysis with risk management yields earlier and better visibility into project performance than is provided by non-integrated methods of planning and control.
The key to the early visibility and the resultant management opportunities is the existence of an integrated baseline. Such a baseline results from a four-step process. First the work of the project is comprehensively defined and subdivided into commonly understood tasks. These tasks are phased consistent with the performance sequence logic and the schedule objectives of the project. Resource estimates are then established for the performance of the tasks and are designated as the budgets or targets for the work; these are timephased consistent with management visibility needs of the project. Finally, this is integrated with the risk assessment/management process.
Okay, given that we have a baseline, then what? We perform the work, status the baseline, analyze “significant” variances, forecast the impacts of our performance on our future performance and make adjustments as appropriate.
There is no substitute for an integrated project schedule, but EVM does offer supplemental insight into project performance. There are several reasons for saying this. First, a meaningfully developed schedule that appropriately incorporates task relationships and dependencies is necessary to develop the baseline, monitor performance and evaluate potential future outcomes based on performance to date; Earned Value Management depends on the existence of this integrated project schedule. Second, EVM actually provides what might be referred to as “accomplishment” performance information as opposed to schedule information. Additionally, EVM schedule information is quantified in terms of resource units and/or dollars rather than units of time. With this as a lead-in you might well wonder how EVM can actually offer anything of value relative to schedule visibility.
As described in the earlier presentations, planned schedule performance in the EVM environment is depicted by the timephased budget element of the plan. This element is called Budgeted Cost for Work Scheduled (BCWS). Another way to think of this element is as the planned value of the work scheduled to be performed. The progress or status aspect of the schedule is denoted in terms of the Budgeted Cost for Work Performed (BCWP). This element can be thought of as the planned value of the work accomplished. If schedule performance is other than the plan, the schedule variance is identified via the following formula:
Schedule Variance (SV) = BCWP – BCWS
The Schedule variance is also depicted graphically in Exhibit
1. At first blush this wouldn’t seem to represent any form of a schedule variance. The BCWP and BCWS data elements are planned in resource units (e.g., hours) or in dollars. Telling your customer or your senior management team that you are $50,000 behind schedule is not usually how you would describe a variance in your schedule performance. However, when you think about it, there is some significance to be imparted by this measure. BCWS represents the budget associated with the work scheduled to have been performed as of the data date. BCWP represents the budget associated with the work that actually has been performed. If there is a difference between the two, it means that we must have performed more or less work than we scheduled and therefore we must either be ahead or behind our schedule.
The reality is that this so-called schedule variance is more of an accomplishment variance. We have accomplished more or less work than we planned. For example, we could be “ahead of schedule” on work that is not on the critical path and thus we aren’t really ahead of our project schedule where it matters. We could even be ahead of schedule on a task not on the critical path and behind schedule on an activity that is. These might “wash each other out” and EVM wouldn’t show any schedule variance, when in fact we have a potentially significant one.
All right already, I thought this EVM stuff added value! Well, the fact is that while the above describes some of the pitfalls associated with relying on EVM as an indicator of schedule performance, it more does point you in the right direction. It will confirm what the schedule shows and if it doesn’t, this can be a tip-off that there may be anomalies in the progress/status that is being reported.
Besides acting as a supplemental indication of whether work is being accomplished as scheduled, EVM can provide keen insight into the schedule performance efficiencies being achieved on a project. This schedule efficiency metric is called the Schedule Performance Index. This is calculated as follows.
Schedule Performance Index (SPI) = BCWP/BCWS
This indicates efficiency relative to the plan. If the index is less than 1.00, you are performing less work than you planned. If the SPI is greater than 1.00, you are accomplishing more than planned.
The greatest value of this measure lies with your assessment of the work that remains to be performed. It should help you in thinking about your remaining duration estimates. You can look at what your schedule efficiency has been and make a mathematical projection of how long your project or any given task might take, if the schedule efficiency doesn’t change. This is calculated as follows.
Potential duration = Planned Duration / SPI
As can be seen in Exhibit 2, it would require a total of approximately 14 months to perform this 10-month project, if the schedule performance doesn’t improve. Therefore, since five months have already elapsed, the forecasted remaining duration is another nine months.
The fact that our performance has been at a given level to date of course doesn’t mean that it will remain that in the future. However, it certainly wouldn’t be appropriate to just assume it will be different in the future. If you are the project manager for the example project and you don’t like the 14-month answer, you have to ask yourself whether things have already turned around such that you can expect better schedule performance in the future. If they haven’t, what are you going to do to cause it to improve? Even if you think improved schedule performance will occur, you would certainly have to admit that finishing the project within the original 10-month schedule, is a very steep hill to climb. There are only five months remaining. If the work was originally planned to be performed linearly over the 10 months, this would mean that about two-thirds of the total work remains to be performed. Based on an even distribution, you planned for 10% of the job per month, yet at 0.70 efficiency, you have only achieved a completion status of 35%. Translating this into future performance, we will have to accomplish 65% of the work in a period where you had only planned to do half the work. Your schedule performance in the future would have to be at the 1.30 level to meet the 10-month objective. This doesn’t mean you can’t do it, but it does yield perspective as to the effort required.
Using EVM, you can track your schedule efficiency from the start, if you see trends you don’t like, the earlier you take action to change them, the better your chances of changing the trend are (i.e., EVM = early visibility and management opportunities).
Cost Related Early Visibility and Management Opportunities
The EVM Early Visibility and Management Opportunities associated with the cost element of our performance are particularly potent.
Most projects have a timephased budget and they have information regarding what has been spent on the performance of the work of the project. Thus, at any stage of the project the project team can compare what has been spent to what was planned to have been spent as of that date. The problem is that while they know what our spending variance is, they don’t have a handle on what has actually been accomplished. Without having some way to quantify accomplishments, they have no way of knowing whether they are achieving their estimated budget targets. This is where the Budgeted Cost for Work Performed (BCWP) comes into play and truly adds value that can’t be achieved in any other fashion. The cost variance is identified via the following formula:
Cost Variance (CV) = BCWP – ACWP
The spending and cost variances are depicted graphically in Exhibit 3.
As other presentations have shown, this is a valuable indication of our cost performance, but there are other ways to evaluate this performance that can aid us with our management assessments of likely future performance. The principal one is the Cost Performance Index. (Admit it, you knew that if there was a SPI there must also be a CPI.) The CPI is a measure of the efficiency of cost performance. It is calculated as follows.
Cost Performance Index (CPI) = BCWP/ACWP
If this metric is less than 1.00, the work performed is costing more than was estimated. If the CPI is greater than 1.00, the work performed is costing less than estimated.
As with the SPI, the greatest value of this measure lies with your evaluation of the work that remains to be performed. It should help you in thinking about your cost-related estimates to complete. You can look at what your cost efficiency has been to date and make a mathematical calculation of what your performance in the future will have to be to meet your current estimate of the total costs at completion (EAC) of the project or of any given task. This is called the To Complete Performance Index. It is calculated as follows.
To Complete Performance Index (TCPI) = (BAC–BCWP)/(EAC–ACWP)
The Budget At Completion (BAC) is the target for the project/ task. The numerator of the equation identifies the budget associated with the work that remains to be performed and the denominator is the estimate of the cost to complete the work.
Exhibit 4 indicates the usefulness of this metric. On this project, we have been experiencing a CPI of 0.78 to date which isn’t that good. Our estimate at completion indicates that our performance for the balance of the project will have to be at 0.906 and to satisfy our boss who thinks we can still meet the budget, it has to be 1.18.
The latter is sometimes called the “CPI to go,” but it is really just the TCPI based on an EAC that is equal to the BAC. Let’s start with it. It is unlikely that we will be able to improve our performance efficiency by 50%. To date we haven’t even been able to come close to our budget targets; to meet our boss’ expectations we would have to not just meet, but also significantly beat our budget targets.
The TCPI of 0.906 is based on the current estimate to complete (ETC). It indicates that while our performance on the remaining work won’t meet the budget targets, our efficiency will be better than it has been. This probably won’t make our boss happy, but if this ETC is a true reflection of our best management judgment of what it will take to accomplish the work, it should at least give her some solace that there is some improvement to be expected in our performance.
Okay, let’s say we want to independently develop our own EAC for someone else’s work in order to assess our comfort level with the other person’s EAC. This could simply be a case of evaluating the EAC of the various members of our own in-house project team or it could be the evaluation of the EAC of a contractor or subcontractor. For that matter, we might just want to “test” our own management judgment EAC.
The difficulty with developing our own EAC for the work of others is that we likely don’t have the technical background to knowledgeably evaluate and develop an estimate of the remaining work that. Even if we do have the technical background, we may not be privy to the details of the work that is being performed. However, if we have earned value information about the work for which we want to evaluate the EAC, we can use the following formula to develop our Independent Estimate at Completion (IEAC).
Independent EAC (IEAC) = ACWP + (BAC-BCWP)/Pf
This formula suggests that if we can identify an expected performance factor (Pf) for the work that remains, we could calculate an estimate to complete. Then by adding the ETC to what has already been spent, the ACWP, we would have our IEAC.
The key element of this equation is, of course, the performance factor, Pf. This is our prediction of what the performer’s To Complete Performance Index (TCPI) might be. “Harry, haven’t we just come full circle on this? How are we any better able to estimate what this performance might be than we are able to just develop a management judgment ETC for the remaining work?” My answer is, not really. We can look at the performance to date and use that information. Possible performance factors that are available in an EVM environment are:
• Cumulative performance (i.e., the CPI)
• Recent experience (e.g., the CPI for the most recent three-month period)
• Cost and Schedule Performance (i.e., consideration of both the CPI and the SPI)
The most common and obvious calculation to make is the one that is based on the cumulative to date CPI. The formula above then becomes:
IEAC = ACWP + (BAC–BCWP)/CPI or IEAC = BAC/CPI
The formula reduces to BAC/CPI because the underlying assumption is that the performance of the remaining work will be at the same level of efficiency as has been experienced to date. This literally means we are saying that all of the work will be performed at the same level of efficiency.
It may seem that the use of the cumulative to date CPI reflects a bit of a defeatist attitude, but in fact this is one of the best early warning signs that, if heeded, can lead to early management opportunities. This is based on an extremely powerful empirical observation that has been made relative to over 900 large-scale Department of Defense programs. It applies to so-called “troubled programs,” which are ones that were experiencing a 10% or more cost and/or schedule variance at the 15% complete point. The 900-plus programs that are alluded to are now complete. The observation is that, when you calculate the CPI at the point where only 15% of the contract work was completed and divide that into the contract BAC, in none of the over 900 programs has there been a case where the total costs at completion is less than the number that results from this calculation. In some instances the actual costs have been very close to the result of the calculation and in others they have been much more. In other words, the formula results in more of a substantive warning than of a specific, reliable prediction.
You might conclude from this that, if things aren’t going well at the 15% point, you might as well give up. The actual conclusion that has been drawn is just the opposite. The inference that has been drawn is that instead of treating this as an early warning of potential problems that deserve immediate attention, project teams have tended to assume that, “we have plenty of time to make it up downstream.” At a later stage (e.g., 35 to 50% complete) when things still aren’t as the team had hoped, it is still possible to take effective corrective actions, but it is too late to achieve cost performance better than the result of this formula. Thus, if you don’t like the number that this equation yields at the 15% mark, take action now, not later. Don’t just dismiss the result of the calculation or there is a good chance you will be joining the 900-plus projects teams who also thought they didn’t need to take action this early.
Since the use of the cumulative to date CPI is a good warning, but not always the best predictor of the cost at completion, there is still good reason to look at the use of recent information and at the combined use of the CPI and the SPI.
For calculations of IEACs based on recent information, there are many options available that have proven effective in different stages of completion. These include CPIs for the most recent three-, six- and 12-month periods. Explanations of all of these that have proven effective require more space and time than we have here. So, what we will address is the one CPI related to recent performance that has proven to be effective in all stages of completion in the life of a project (Methods of Estimating Contractor Cost at Completion, Joseph J. Haydon & Richard O. Riether). This is the three-month moving average. Using this, the equation becomes:
IEAC = ACWP + (BAC–BCWP)/CPI3 where
CPI3 = 3-Month moving Average = (BCWPmo + BCWPmo-1 + BCWPmo-2)/(ACWPmo+ACWPmo-1 + ACWPmo-2)
The value in using recent information is that if some technical issue that has been causing cost and/or schedule issues has been resolved, the recent performance period CPI will yield an early indication of this. The performance period needs to be three months or more to be useful, because there can be too much variation in a shorter period.
There are two frequently used equations that involve consideration of both cost and schedule. The logic behind why it would make sense to consider both is captured by the old project management maxim; cost variances tend to follow schedule variances. Maxims aren’t formed unless there is substance behind the precept. The two equations are:
IEAC = ACWP + (BAC–BCWP)/(CPI * SPI) and
IEAC = ACWP + (BAC–BCWP)/(0.2SPI + 0.8CPI)
Both use only the cumulative to date CPI and SPI. The first frequently yields the largest value for the IEAC, because if both the indices are less than 1.0, the product of the two is a smaller fraction than either index. Even though the IEAC is often the highest of the ones that result from the equations described in this paper, it has still been found to be a statistically significant predictor of total costs at completion (ibid).
The second equation assumes a 20% influence associated with the schedule efficiency experienced to date and an 80% influence by the cost efficiency. This formula is the result of a study that was done by the U.S. Air Force on a large body of completed contracts for which they had earned value reports over the entire life of each of the projects (HQ Air Force Systems Command Estimate at Completion Formula Justification, Timothy J. Wallendar). It has been observed by some users of this weighting concept that the 20/80 split may not be appropriate for all projects. The thinking here is that schedule risk varies from project to project and that while this may be a good ratio for many projects, that others with significant amounts of schedule risk might look to a ratio that increases the potential influence of the schedule (e.g., 0.4SPI + 0.6 CPI).
There are two notes of caution that must be made relative to the mathematical derivation of an independent EAC. The first is that no one formula should be selected and looked at exclusively. While the formulas have proven to be good predictors of final costs, they don’t all yield the same result. Therefore, it is better to think of each of them as providing a “ballpark” figure. Exhibit 5 depicts the concept that is recommended here. Pick several formulas and use them to depict a range within which you might reasonably expect the management judgment EAC to fall. If the management judgment EAC falls in this range, you might feel a little more comfortable with the EAC.
This brings us to the second note of caution. Calculated IEACs are not a substitute for those developed based on management judgment relative to the work that remains to be performed. These are simply forms of “sanity checks.” If the management judgment EAC “falls below the range,” it doesn’t mean the formula is right and the other is wrong. However, since there is a significant body of project experience that suggests the formulas are useful predictors of final costs, if there is a substantive difference between the management judgment EAC and the IEACs, it is appropriate to ask what actions are being taken to make the EAC possible.
The key point in this section of the paper is that EVM can assist you from the early stages of the project to quantify your cost performance to date. It can help you understand what your performance in the future has to be in order to meet whatever your current estimate of your at completion costs is. Finally, EVM can help you make “what-if” calculations about final costs by the use of various versions of the IEAC equation. So once again, using EVM, you can track your performance from the start of a project and, if you see trends that you don’t like, the earlier you take action to change them the better your chances of changing the trend.
Scope/Technical Related Early Visibility and Management Opportunities
The primary focus of this paper is on the schedule and cost topics discussed above, but there is a clear and direct connection between those and scope/technical management. In order to truly understand the character of one’s performance to date as depicted by the CPI and SPI the work that has been performed and the technical issues that may have contributed to the e efficiency or lack thereof is needed. Indices that fall below certain performance levels then become triggers for conducting this analysis. In looking to the future and developing an estimate to complete for the work that remains to be performed requires a similar understanding of the work, the technical issues that remain and the corrective action plans that are in place or will be implemented. If the TCPI represents performance that isn’t consistent with these, then the EAC should be reassessed. If the management judgment EAC falls below the range of IEACs and you have not implemented an action plan that can make your EAC probable, the EAC and/or your operating plan needs to be reassessed. Thus, EVM again is providing the early warning that can lead to the early management opportunities.
Many of you may be thinking at this point that this sounds good, but what about all of those project management disasters in the Defense contracting world where EVM is routinely used. There have been disasters, but all too often the postmortem has suggested that the problem wasn’t with the EVM approach, instead it a case of people not believing the information when it first became available. Even given the disasters, there are clear signs of successes in the management of Defense projects. The Defense Contract Management Command released statistics last fall that indicated at that point on programs with a total value of $72.8 billion, which, on average were two-thirds complete, were experiencing cost overruns of just 5.5%. Preliminary, more recent figures show an improvement even over this. Thus, there is substantive reason to say that good project management, including the use of the earned value tool is yielding very good performance results.
The main conclusion of this paper is that Earned Value Management can provide added value to the project team by offering earlier visibility into project issues. This in turn can lead to the opportunity to shift the direction of undesirable trends earlier and therefore increases the likelihood of success. Or EVM = EVM.
Proceedings of the Project Management Institute Annual Seminars & Symposium
September 7–16, 2000 • Houston,Texas,USA