Earned Value Status and Forecasts
If the project keeps spending the way it has, here's where it ends up.
Before this Lesson
The monthly report on a $400,000 site-work contract lists three activities. Clearing is done, excavation is three-quarters done, and the storm drain is a fifth done. Average the three, and the job looks 65% complete. It's really 50% complete, and it has spent $250,000 getting there. If the rest of the job goes the way the first half has, it will finish at $500,000.
Both of those numbers come straight out of the report. The 50% comes from pricing each activity's progress at its own budget, and the $500,000 from assuming the job keeps spending the way it has.
The Handbook's Earned-Value Analysis list, in its Construction section, names three dollar amounts at a status date:
- BCWS, planned value: the budgeted cost of the work scheduled by now.
- BCWP, earned value: the budgeted cost of the work actually performed.
- ACWP, actual cost: what the work performed really cost.
Earned value is the yardstick. Compare it with actual cost to judge cost, and with planned value to judge schedule:
A positive variance or an index above 1 is good news: under budget, or ahead of schedule. For more, see Earned Value.
Earned value from a status table
A real job has several activities moving at different speeds, so its status report lists each one's budget, percent complete, actual cost, and planned value. Here's the site-work job, with one column added:
| Activity | Budget | Complete | ACWP | BCWS | BCWP |
|---|---|---|---|---|---|
| Clearing | $40,000 | 100% | $45,000 | $40,000 | $40,000 |
| Excavation | $160,000 | 75% | $145,000 | $160,000 | $120,000 |
| Storm drain | $200,000 | 20% | $60,000 | $100,000 | $40,000 |
| Total | $400,000 | $250,000 | $300,000 | $200,000 |
The report hands you actual cost and planned value; those columns just add up. Earned value is the one you build. Each activity has earned its percent complete times its own budget. Excavation is 75% done on a $160,000 budget, so it has earned 0.75 × $160,000 = $120,000. Add the dollars, and the job has earned $200,000 of its $400,000. That's where 50% complete comes from: earned dollars over budget dollars.
The 65% came from averaging the percents, which counts a percent of clearing as if it were worth as much as a percent of storm drain. It isn't. The drain's budget is five times clearing's, so clearing's 100% and the drain's 20% are the same $40,000 of work. A percent means something only against its own budget, so turn each one into dollars before you add anything.
The Handbook defines earned value only in words, as the budgeted cost of work performed, and percent complete times budget is how you count it. Two other shortcuts fail. Percent complete times actual cost means nothing: actual cost is already the cost of the work done, so the percent counts the progress twice, and earned value prices work at the budget anyway. And the fraction of the schedule gone by measures the calendar, not the work.
If a report gives each activity's percent scheduled instead of its planned dollars, build planned value the same way: percent scheduled times budget. With the three totals in hand, the project's status follows from the usual comparisons:
The job is over budget and behind schedule. Each dollar it has spent has bought 80 cents of budgeted work.
A water-main job reports this status at the end of month 5:
| Activity | Budget | Complete | Actual cost | Planned value |
|---|---|---|---|---|
| Mobilization | $20,000 | 100% | $22,000 | $20,000 |
| Pipe laying | $400,000 | 50% | $236,000 | $250,000 |
| Valves and hydrants | $80,000 | 75% | $62,000 | $80,000 |
| Paving restoration | $100,000 | 0% | $0 | $0 |
Running the equations backward
Each status equation ties three numbers together, so a question can give you any two and ask for the third. Often it hands you an index or a variance and one dollar amount, and wants a different dollar amount back.
The cost index is easiest to run backward if you read it as what each dollar buys. CPI = BCWP/ACWP says each dollar spent has bought CPI dollars of budgeted work. So the money spent is the work done divided by CPI, and the work done is the money spent times CPI:
Had the site-work report given only CPI = 0.80 and an earned value of $200,000, the spending would follow: 200,000/0.80 = $250,000. Check the direction every time. A CPI below 1 means the job has spent more than it earned, so its actual cost has to come out larger than its earned value. That check catches a multiplication where a division belongs.
SPI runs backward the same way, with planned value in place of actual cost. For the variances, keep the signs: SV = BCWP − BCWS rearranges to BCWS = BCWP − SV, and a job that's behind has a negative SV, so subtracting it adds. Some questions chain two steps, using one index to reach earned value and the other to leave it. Earned value sits in both indices, so it's always the bridge.
Forecasting the finish
Now the other question a status report has to answer: what will the whole job cost? The Handbook's Forecasting list, in the same Construction section, starts from the budget at completion, BAC, which it defines as the original project estimate. It's the sum of every activity's budget: $400,000 for the site-work job. The forecast takes two steps.
First, how much work is left? In budget dollars, it's the budget less what's been earned: BAC − BCWP = 400,000 − 200,000 = $200,000 of budgeted work still to do.
Second, what will that work cost? So far the job has paid $250,000 for $200,000 of budgeted work: $1.25 for every budget dollar, which is 1/CPI. If the rest of the work goes at the same price, the remaining $200,000 will cost $200,000 × 1.25 = $250,000. Multiplying by 1/CPI is dividing by CPI, and that's the Handbook's estimate to complete:
The estimate at completion adds the money already spent to the money still to spend:
Notice that the budget left over, BAC − ACWP = $150,000, isn't the estimate to complete. It's what remains in the budget, not what the remaining work will cost. At the job's current price, $150,000 buys only $120,000 of the $200,000 of work still to do.
There's a shortcut, which the Handbook doesn't print but its two lines imply. The money already spent bought earned value at that same price, ACWP = BCWP/CPI. Put that into the estimate at completion:
The work done and the work left are priced at the same ratio, so you can price the whole budget at once: 400,000/0.80 = $500,000. That's the whole forecast in one line. Take the original estimate, and price every dollar of it at what budgeted work is really costing.
What the forecast assumes
Because EAC = BAC/CPI, the cost index alone tells you which way the job is headed. A CPI below 1 divides the budget by less than 1, so the forecast lands above the budget and the job finishes over. A CPI above 1 puts the forecast below the budget, and a CPI of exactly 1 puts it on budget. The Handbook's forecast uses only the cost index; SPI doesn't appear in it.
Behind the division sits one assumption, which the Handbook doesn't state: the current cost efficiency continues. Every remaining dollar of budgeted work is assumed to cost what the earned dollars have cost so far. It's the reasoning you'd use on a road trip. If the first 200 miles took 10 gallons, you'd guess 20 more for the next 400, and you'd be right on a flat highway and wrong if the first 200 miles climbed a mountain pass. The forecast is only as good as that assumption, so before you trust it, ask whether whatever set the CPI so far will keep acting on the rest of the work.
Some questions go one step further and ask for a variance at completion. The Handbook defines none, so use the definition the question gives. The usual one is BAC − EAC, which keeps the familiar sign: negative means a projected overrun.