Earned Value
Spending exactly the budget so far can still mean you're behind schedule and over budget.
A crew is building 1,000 ft of fence on a $20,000 budget, planned at 100 ft a day for 10 days. At the end of day 4, the owner asks how it's going. The contractor has two numbers ready: the plan called for $8,000 of work by now, and the job has spent $7,000. That sounds like $1,000 under budget.
Then you walk the line and count 300 ft of fence. That one count turns the story around. The job is over budget and behind schedule, and the two numbers the contractor compared can't show either problem.
Three numbers at the status date
Controlling a project comes down to plan versus actual. Earned value analysis makes that comparison in dollars at a status date, a moment when you stop and take stock, and it takes three numbers, not the contractor's two:
- Planned value is the budgeted cost of the work the schedule said would be done by now. The plan called for 400 ft of fence by day 4, at the budget's $20 a foot: $8,000.
- Earned value is the budgeted cost of the work actually done. The 300 ft that are up were budgeted at $20 a foot, so the job has earned $6,000. It's credit for work in place, priced the way the budget priced it, whatever the work really cost.
- Actual cost is what the work done so far really cost: the $7,000 in payroll and materials.
The Handbook's Earned-Value Analysis list, in its Construction section, names them with acronyms that read like sentences:
| Plain name | Handbook | Stands for | Fence |
|---|---|---|---|
| Planned value | BCWS | Budgeted cost of work scheduled | $8,000 |
| Earned value | BCWP | Budgeted cost of work performed | $6,000 |
| Actual cost | ACWP | Actual cost of work performed | $7,000 |
Adapted from the FE Reference Handbook, Construction: Earned-Value Analysis
Read each label in two halves. The front says how the work is priced, at the budget or at what it actually cost. The back says which work: the work scheduled or the work performed. Most project-management texts shorten the plain names to PV, EV, and AC, and some problems do too.
Now look at what each pair of numbers has in common. Planned value and earned value both price fence at the budget's $20 a foot; they differ only in how much fence they count, 400 ft against 300 ft. Earned value and actual cost count the same 300 ft; they differ only in the price on it, the budget's against the real one. Actual cost and planned value differ in both ways at once.
Earned value is the yardstick
Earned value shows up in both good comparisons. It's the yardstick: hold it against actual cost to judge cost, and against planned value to judge schedule. The Handbook's Variances list writes the two comparisons as differences, the cost variance CV and the schedule variance SV, with earned value first in both. It spells them out in words as Earned − Actual and Earned − Planned:
For the fence:
The 300 ft of fence cost $1,000 more than the budget said they would, and $2,000 worth of fence that should be up by now isn't.
The Handbook doesn't say which sign is good news, but the order settles it. Earned value is what the job has to show for itself. Subtract what was spent, or what was promised, and a positive result means the job is ahead of that number. So a positive cost variance means under budget, a positive schedule variance means ahead of schedule, and a negative one of either is bad news.
The chart shows one more gap, running sideways. The 300 ft that are up were planned for day 3, so you could say the fence is a day behind, and people do. But that horizontal gap isn't the schedule variance. SV is the vertical gap, in dollars: −$2,000.
The comparison that answers neither
Back to the contractor's two numbers: $7,000 spent against $8,000 planned. It's the comparison you'd make with a household budget, and on a project it misleads. Imagine the crew had stayed home for all four days. Nothing built, nothing spent, so actual cost is $0 against $8,000 planned: $8,000 under budget, by that comparison. Nobody saved a cent. The job is four days behind, and the fence it didn't build still has to be paid for. Comparing spending with the plan makes falling behind look like saving money, because work that isn't done hasn't cost anything yet.
The real fence did the same thing on a smaller scale. Its actual cost, $7,000, sits between its earned value, $6,000, and its planned value, $8,000. Whenever actual cost lands between the other two, comparing it with the plan gets the budget status backward, not just the amount. A crew running ahead of schedule shows the mirror image: it spends more than the plan, looks over budget, and can be under. When actual cost lies above or below both, the comparison gets the direction right and the size wrong. Either way the cure is the same: judge cost against earned value, every time.
Performance indices
Variances are in dollars, so their size depends on the size of the job. A $1,000 overrun matters on a fence and is a rounding error on a highway interchange. The Handbook's Indices list makes the same two comparisons as ratios, with earned value on top in both:
Each ratio reads as a plain sentence. The cost performance index is how many dollars of budgeted work each dollar spent has bought. The schedule performance index is the fraction of the planned work that's actually done. For the fence:
Each dollar spent on the fence has bought about 86 cents of budgeted fence, and three-quarters of the fence planned by now is up.
Keeping earned value on top keeps the indices in step with the variances. An index above 1 goes with a positive variance and good news, an index below 1 with a negative variance and bad news, and exactly 1 with a variance of zero: on budget, or on schedule. Turn a ratio upside down, as in ACWP/BCWP, and the verdict flips with it.
Reading a pair
The two indices answer separate questions, so read each one on its own, then put the verdicts together. Nothing ties them to each other. A job can be under budget and ahead, over budget and behind like the fence, or good on one count and bad on the other. A pair of variances reads the same way.