Integrating scheduling and earned value management (EVM) metrics
Earned Value Management (EVM) is a 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. The earned value metric can be used to assess both cost and schedule performance.
Earned Value Management is more than a metric, however. It is the combination of robust planning and integrated management system elements [Gary Christle]. 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. This baseline is the result of a true integration of the work scope, schedule, and cost elements of the project. When this is successfully accomplished we can readily correlate our performance as reported from the perspective of earned value to that reported via our integrated project schedule. Let's start from the point of view of the earned value schedule variance.
The earned value metric is actually the planned value of the work that has been accomplished, but it is often referred to as the budgeted cost for work performed (BCWP). The baseline plan that performance is measured against is an aggregation of the timephased value of the work planned to be performed. This element is often referred to as the budgeted cost for work scheduled (BCWS). It is the difference between these two elements that quantifies the schedule variance in an earned value management system. Both the cost and schedule variances can be seen in Exhibit 1, but for this discussion our focus is the schedule variance.
Translating a Schedule Variance to Time
Even though Earned Value provides us with a Schedule Variance, experienced project management people always say, “Look at your physical schedule or you won't know if you have a meaningful schedule variance issue.” One reason for this is that when you read that you have a schedule of $200,000, what does that mean? We all relate schedule to time not dollars or hours. The other issue is that the EVM schedule variance is really more of an “accomplishment variance. However, despite these concerns, the earned value schedule variance does provide a valid alert to the fact that work is not progressing consistent with the planned schedule.
To provide us with an example to use through out this paper, I've selected an actual project that currently has a schedule variance of $203,200 for one element of its work. So what might that a schedule variance of better than $200,000 mean? Well, it depends. It definitely means that we haven't completed as much work as we had planned to. However, the question is, “If we are behind schedule, what does this translate to in terms of the amount of time behind. One way to approximate the answer to this question is to apply the formula below.
Where the average BCWS may be calculated using the total contract months to date or a selected number of recent months. The idea is that the timephased BCWS indicates the pace at which the work was planned to be performed. As such, if we take the BCWS for a given period of time and divide it into the schedule variance, the result is the amount of time we are ahead or behind schedule. For example, Exhibit 2 shows that, based on the average rate at which the work was planned to be performed from the beginning of the project to the present, we are almost 2.25 months behind schedule. However the results for the second two calculations are less pessimistic. This is because the project startup wasn't planned to be linear. This project has been ongoing for a while, but the plan is only just now beginning to show a ramp up in effort. In this case, it would appear to make more sense to use either the second or third formula. The apparent long slow ramp up causes the first formula to yield an exaggerated behind schedule condition. Whereas the other two use a factor that is a reflection of the pace at which our plan says we should be performing at right now. However, if the plan had called for the work to be linear (i.e., level loaded), the inception to date approach would be just as valid a calculation as the other two.
Correlating the Schedule Variance and the Project Schedule
While this is an adequate method to approximate what the schedule variance translates two in terms of time, it really only tells us about the pace of our accomplishments. Depending on which tasks are behind schedule per the plan, you may or may not be behind schedule relative to completing the project on time. However, whatever the situation is, if $200K represents a significant variance as a percentage, you need to look at your physical schedules NOW.
What we want to know is whether any of the activities that are contributing to this behind schedule earned value variance are on the project's critical path. Exhibit 3 helps us answer this question. This exhibit is courtesy of CS-Solutions and is a partial screen capture from one of their tools, CS Glue.
For the work breakdown structure (WBS) element that has the $203,200 unfavorable schedule variance, it shows all of the activities that are late as well as those that are forecasted to be late. Tasks 12, 20, 10, 28, and 32 are late and are all contributing to the schedule variance, but only tasks 20 and 28 are on the critical path. Regardless of the dollar value of our schedule variance, these are significant, because they put the project end date in jeopardy.
This is an example of schedule and cost integration working effectively. Our earned value information alerted us to an issue relative to our accomplishment of work, which then led us to look directly at the specific activities in the schedule that were contributing to that condition. We might have found that none of the associated activities were on or near the critical path. Had that been the case it would have been a valuable analysis, because it helped us know how to prioritize our concerns. In this case our analysis indicated that we should be concerned about this WBS element. That caused us to do further analysis. I haven't included another exhibit, but what we found was that the implications of the poor performance to date on a number of the as yet unstarted tasks was quite significant. There were ten of those tasks that had a forecasted negative float and/or had three slips of their expected completion date in the last four months.
The ability to readily develop the analysis describes was wholly dependent on how successfully we had developed an integrated baseline for the project. There would have been no way to readily link the schedule variance depicted by the earned value information with the specific schedule activities that were associated with it unless we had successfully developed an integrated baseline plan. Exhibit 4 illustrates this point. The earned value information is generated at what is called the work package level. Those work packages have to be correlated directly with one or more activities in the integrated baseline schedule. In most instances the work packages and the activities are maintained in two different tools (i.e., a cost processor and a scheduling tool). When this done effectively, the correlation described above is a simple one to make.
However, there is more that has to be in place for a project team to have an effective capability to this kind of analysis. The budgets that are planned (i.e., the BCWS) have to be based upon/consistent with the resource planning that the schedule is predicated upon. If they are not, there is a fundamental performance capability difference between the two elements of what is supposed to be an integrated baseline.
Correlating the SPI with Schedule Data
Let's review another earned value metric, the schedule performance index (SPI). The SPI is a measure of how efficient we have been in accomplishing our work relative to our plan for its accomplishment. It is calculated per the formula:
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 SPI can be used in many ways, but is it consistent with what our schedule information is saying? It is possible to confirm that the schedule performance data are consistent with the SPI by use of one of these two formulas.
The first of these assumes that the critical path method of scheduling is being used. The critical path is not necessarily comprised of the most “critical” elements of the project in the context of risk, but instead is the longest path of the schedule. Thus it determines the length of the project based on current information. Float (AKA slack) is a factor that reflects the amount of time that any one activity can be delayed before it delays the project. The critical path always has the least amount of float associated with it. If the project is behind schedule, the critical path will usually have negative float. When that is the case, the formula will yield a number less than unity.
If you aren't using the critical path method of scheduling, but you are maintaining a baseline and some form of an estimate of time ahead or behind schedule you can still make an analogous calculation to the SPI.
If there are significant differences between the SPI and either of these calculations it makes sense to look more closely at the performance information. While the comparison will seldom yield identical numbers, major disparities would seem to imply that the cost and schedule elements of the project baseline aren't truly integrated.
The SPI and Schedule Management Decisions
Once you have satisfied yourself that the SPI is consistent with other scheduling information, there are ways to use it to facilitate decision-making. For example, how long will this project take if our schedule performance efficiencies don't improve and are the remaining durations that are being reported consistent with performance?
To calculate how long any given activity or the project, as a whole will take to complete, if there is no improvement in the schedule performance, you can use this formula.
If the planned duration for the remaining work is 12 months, but our schedule performance has been such that our SPI is 0.80, it will take 15 months to finish the work, unless our schedule performance efficiency improves.
Similarly, project teams can assess the remaining duration estimates in the context of schedule efficiency as reflected by the SPI. This is just a slightly different form of the equation above. Here however, you are focused on the forecasts of remaining durations for individual schedule activities. The equation here is
The object here is to evaluate the remaining duration estimates to assess whether it appears performance experience to date is being considered.
Correlating the CPI With Resource Information
In addition to the SPI, earned value management also provides a metric called the cost performance index (CPI). While the SPI is a measure of our performance efficiency relative to the plan reflected by our schedule, the CPI is a measure of our cost efficiency. It is arrived at by comparing what the work performed has cost us (actual cost of work performed – ACWP) compared to what the planned value (budgeted cost for work performed – BCWP) of that work is. It is calculated per the formula:
This indicates cost efficiency relative to the plan. If the index is less than 1.00, the work you are performing has cost more than its planned value. If the CPI is greater than 1.00, you are accomplishing the work for less than the planned value. The CPI can be used in many ways, but is there a way to apply it to our scheduling efforts? I think there is. We can apply it to the analysis of our planned resources.
What I am about to describe is something that hopefully seems intuitively obvious. However, it is sometimes overlooked. The CPI is often used to conduct a “sanity check” of a project's estimate to complete (ETC), in terms of dollars. When we do this and convince ourselves that the ETC is a meaningful one, we should, of course, align the resource planning with this, but sometimes this gets lost in the mix of providing cost analysis and we forget to look at the schedule. If we develop our management judgment ETC and conduct some confirming analysis of it using the CPI, it should stand to reason that the resource requirements associated with the new ETC should be consistent with the following calculation
In this instance the appropriate CPI to use would be calculated using hours for ACWP and BCWP. You would calculate this for each resource category. The focus of the analysis would then be on whether the currently authorized resources match this calculation. If they don't, then it's logical to infer that they don't support the new ETC.
The title of this paper, Integrating Scheduling and Earned Value Management Metrics, seems to imply that there is something novel about integrating these two. However, that isn't the case. This is just a good example of integrated project management, which is what Earned Value Management is. As I say at the outset, EVM is more than metrics. What I have tried to show here is that when you do have an integrated management environment these EVM metrics can be used to verify and validate our thinking with regard to schedule management.
Proceedings of the Project Management Institute Annual Seminars & Symposium
October 3–10, 2002 • San Antonio, Texas, USA