Earned Value from the Project

So far, we have discussed monitoring segments of a project—tasks, subtasks, and such. Of primary importance, however, is deriving some measure of overall project progress in terms of performance, budget, and schedule. Such a measure is “earned value.”

The common problem with comparing actual expenditures against plan (a.k.a. base­line or budget) for any given time period is that the comparison fails to consider the actual progress made on the project. Thus, if expenditures are lower than expected for a given period, it may be either good or bad, depending on whether progress is in line with that amount of expenditure. Similarly, if expenditures are higher than expected, this may be acceptable if progress is sufficiently greater than planned for that period. The performance of a task or project cannot be evaluated by considering cost factors alone.

The earned value (EV) of a task (or a project) is the budgeted cost of the work actu­ally performed. It is calculated by multiplying the budgeted cost of the task by the percentage of completion of the task and summing over all tasks for the project. This process is more difficult than it might sound. The budgeted cost of a task is clear, but the percentage of completion is not. In particular, the percent of a task’s budget actually spent is not, in general, a particularly good indicator of the percent of that activity’s
completion. For example, the major cost for a task might be for obtaining the machinery to do the task, a cost that will be incurred before any progress is made on that task. Or perhaps the major cost will not be charged until completion of the task. To the best of our knowledge, there is no satisfactory way to measure accurately the percent of completion of most tasks, let alone to measure accurately the percent of completion of an entire project.

As a result, three conventions have been adopted for estimating progress on tasks, but they must not be confused with reality. The most popular is 50-50: the task is listed as 50 percent complete when work on it is initiated, and the other 50 percent is added when the task is completed. This approach avoids the difficult problem of trying to estimate progress while the task is being executed. Clearly, this overstates the EV of tasks that have recently begun, but understates the EV of tasks nearing completion. In a large project, of course, with multiple on-going tasks starting and stopping at different times, the overstat­ing and understating may tend to even out resulting in a relatively accurate portrayal of project progress. Another convention is 100 percent when the task is complete and zero percent before that. This is a very conservative approach that will only show project pro­gress that is definitely achieved. Projects, however, will always appear to be “behind schedule,” and upper management will be in a constant state of worry about the project’s progress (not to mention the project manager’s competence). The last approach is the common, intuitive one of trying to estimate percentage completion by using the ratio of cost expended to the total cost budgeted for a task, the ratio of the actual time elapsed relative to the total scheduled task time, or some combination of both. There is no strong evidence that either the time or cost ratio is an accurate estimator of percent completion.

These conventions are meant for application only to individual tasks on a project, not to the project as a whole. (At the end of this section, we show how aggregating the task EVs can be used to estimate the progress of the project as a whole.) In fact, applying these conventions to the project as a whole may result in a seriously misleading number because none of the conventions noted has much semblance of reality. If one must esti­mate the percent complete for an activity—and as we shall see, one must—the best guess of a knowledgeable person is probably the most that can be expected for most tasks.

As the earned value of a project is calculated, a graph such as that shown in Figure 7-2 can be constructed. Given the limitations of the input data, such a figure provides a basis for evaluating cost, schedule, and performance to date. The earned value completed to date tells the manager whether progress is up to expectation, the “baseline” planned for this point in time. Any difference is called the “schedule variance,” which shows how much the project is ahead of or behind schedule. As seen in the figure, the value com­pleted to date is less than the baseline estimate for this point in the project’s life and represents about a 10-day delay, resulting in a negative schedule variance (delay is nega­tive). The actual cost, however, is considerably above the amount of value completed, resulting again in a negative (bad) “spending variance,” and even exceeds the baseline cost for this point. Thus, this figure represents a difficult situation—a project signifi­cantly behind schedule and considerably over budget at this time in the project’s life.

The variances on the earned value chart are calculated based on two simple rules:* (1) A negative variance is “bad,” and a positive variance is “good”; and (2) the spending and schedule variances are calculated as the earned value minus some other measure. Specifically, the cost or spending variance is the earned value (EV) less the actual cost (AC) of the work performed. The schedule variance is the EV less the planned cost (PV) of the work budgeted and scheduled to have been performed to date as determined from the baseline plan. All measures are made as of the same date. For a variety of examples of different possible scenarios, see Figures 7-3(a), (b), and (c).

Although the above method of calculating variances is relatively standard, another way of handling the data is more useful for making comparisons at different points in time, or across different projects, or among different project managers. This procedure is simply to take the ratios of the measures rather than their differences. Thus, the spending or cost variance becomes the Cost Performance Index (CPI) = EV/AC, and the schedule variance becomes the Schedule Performance Index (SPI) = EV/PV. Values less than 1.0 are undesirable.

We illustrate the above with a simple example. Suppose that work on a project task was expected to cost $1,500 to complete the task and the workers were originally sched­uled to have finished today. As of today, however, the workers have actually expended $1,350, and our best estimate is that they are about 2/3 finished. Our calculations follow:

Cost / spending variance = EV-AC = 1,500(2/3) -1,350 = -350

Schedule variance           =EV — PV = 1,500(2/3] 1,500 = —500

CPI                                   =EV/AC = 1,500(2/3)/1,350 = .74

SPI                                   = EV/PV = 1,500(2/3)/1,500 = .67

Thus, we are spending more than the baseline plan indicates and, given what we have spent, we have not made as much progress as we should have.

We can use these calculations to determine some additional items of interest as well, such as the estimated (remaining cost) to completion (ETC) and the projected (total cost) estimated at completion (EAC) if nothing is done to correct the problem. Given that the budget at completion (BAC) is 1,500 and the EV is 1,500(2/3)                                                                                          1,000,

ETC = (BAC- EV) /CPI = (1,500 -1,000)/.74 = 676

Note that this assumes that the work will be completed at the same level of efficiency or inefficiency as conducted thus far. The total cost to complete the task is

EAC = ETC + AC = 676 + 1350 = 2,026

That is, the projected additional cost to complete this task is $676 which, when added to the cost accrued to date of $1,350 gives a total task cost to completion of $2,026 rather than the original estimate of$1,500. (See Barr, 1996, and Flemming and Koppleman, 1996.)

Tables 7-1 and 7-2 illustrate the application of some of these variance calculations for a project to produce a documentary film. Tables 7-1 and 7-2 were generated by MSP. Table 7-1 depicts a status report for the documentary film project outlined in Chapter 6.

Table 7-1 is an overview of the actual information on the project steps compared to the baselines for duration, start dates, finish dates, cost, and hours worked. The project began 3/29 and is scheduled to be completed by 5/26. The project baseline or budgeted task and cost information is compared to what actually happened.

At any point during the project the project plan can easily be updated. MSP uses earned value analysis as a tool to compare the actual to the baseline plan. Table 7-1 shows the project updated as of 5/2. Task 2.0 has a baseline (budgeted) duration of 7 days. It actually took 12 days to complete. Task 3.3 took only 3 days instead of the baseline of 5 days. Since we hired the secretary earlier, we could start task 3.4 for scheduling the shoots earlier than expected. Because task 2.0 took longer, however, task 4.0 was delayed in starting. (MSP makes these adjustments automatically once actual information is entered.) All other tasks were on schedule. Step 6 is underway and not yet completed. Steps 7 to 9 have not yet begun.

Table 7-2 is the earned value table prepared by MSP once the project information is updated. Note that only completed stand-alone tasks or subtasks have final cost variances calculated for them. Summary tasks show the roll-up variances for their subtasks. The col­umns titled EAC, BAC, and VAC show “estimate at completion,” “budget at completion,” and “variance at completion (VAC = BAC – EAC),” respectively. MSP does not automati­cally adjust the estimate to complete by the CPI of the project to date as we did above.

As noted above, some older versions of MSP do not calculate earned value variances recommended by the PMI. The MSP earned value numbers in Table 7-2 have been cor­rected to reflect the PMI Project Management Body of Knowledge (PMBOK, Chapter 11) definitions. (In some older versions, MSP calculation methods even differ from the MSP “help” file definitions.) The calculations used in this book are those recommended by the PMI. Our detailed calculations above were for a particular project task, but the approach can sensibly be extended to an entire project to give a reasonable estimate of the percent completion of the project. The conventional 0/100, 50/50, or percent of budget (time) expended are too ham-handed to apply to the project as a whole. An alternative is to aggregate the calculations made for individual tasks, as shown in Table 7-2. Using the overall task data for the project to date, one can estimate the project CPI and SPI, and thus the cost and schedule at completion for the project.

If the earned value calculations indicate a cost or schedule deficiency, the PM must still figure out what to do (control) to get the project back on budget or schedule. Options include borrowing resources from other tasks, holding a brainstorming meeting with the project team, requesting extra resources from senior management, or informing the cli­ent and/or senior management of the projected deficiency. Unfortunately, research on 700 military projects indicated that the chances of correcting a poorly performing project more than 15-20 percent complete were effectively nil (Christensen, 1994).

An interesting exception to this situation is that of ViewStar Corporation, under contract to design an Accounts Receivable imaging system for Texas Instruments. As the project progressed, although costs were under control, earned value progress fell behind. To catch up, ViewStar directed the project team to meet only key project requirements at the least cost and time (Ingram, 1995). Earned value thereafter rapidly increased back to the planned schedule.

Earned value represents a way to capture both in-process performance and cost on a certain date as measured against budget or schedule. Including the planned costs and actual costs allows the calculation of spending and schedule variances, where nega­tive values are undesirable. Using these figures, a projection can be made of costs to completion and total cost for the task or project under consideration. Although percentage completion makes limited sense for individual tasks and work elements, it has little meaning for the project as a whole. Nonetheless, one can aggregate the individual task earned values and variances to make reasonable estimates of project completion costs.

Source: Meredith Jack R., Mantel Jr. Samuel J., Shafer Scott M., Sutton Margaret M. (2017), Project Management in Practice, John Wiley & Sons, Inc. 3th Edition.