Earned value management in plain numbers: CPI, SPI and EAC with one worked example

"We've spent 40% of the budget and we're about 40% done." It sounds like a project on track. It's also the sentence that precedes most budget overruns, because it compares money spent with a feeling about progress.

Earned value management (EVM) replaces the feeling with a number. It asks one simple question: how much of the planned work has actually been finished, valued at what we budgeted for it? With that, you can tell whether a project is over budget, behind schedule, or both, and forecast where it will land, months before the end.

Three numbers

One more you'll need: BAC, the budget at completion, which is simply the total budget.

The example: a €200k, ten-month program

A company is replacing its warehouse management system. The budget (BAC) is €200,000 over 10 months. At the end of month 4, the project manager collects the numbers.

The plan said 40% of the work should be done by now, so PV = €80,000. Finance says AC = €75,000 has been spent. So far, so reassuring: less spent than planned.

Now the earned value. It's calculated per work package, from the budget and how much is really finished:

Work package	Budget	Complete	Earned
Requirements and design	€20,000	100%	€20,000
Data migration design	€30,000	100%	€30,000
Build integrations	€50,000	20%	€10,000
Testing	€40,000	0%	€0
Training	€30,000	0%	€0
Go-live and hypercare	€30,000	0%	€0
Total	€200,000		EV = €60,000

Only €60,000 of the planned €80,000 of work is done, and it cost €75,000. The picture has changed completely.

Variances and indices

Read them like this. A CPI of 0.80 means every euro spent is buying 80 cents of planned work. An SPI of 0.75 means the team has completed three quarters of what it planned to by now. Both below 1.0: over budget and behind schedule, despite having spent less than planned.

Forecasting the finish

The real value of EVM is the forecast. The estimate at completion (EAC) depends on what you believe about the rest of the project, so there are a few standard versions:

If you believe…	Formula	Result
Current cost efficiency will continue	EAC = BAC ÷ CPI	€250,000
The overrun was a one-off; the rest will go to plan	EAC = AC + (BAC − EV)	€215,000
Cost and schedule pressure will both keep hurting	EAC = AC + (BAC − EV) ÷ (CPI × SPI)	€308,333

From the EAC you get two more useful numbers: the estimate to complete (ETC = EAC − AC = €175,000 still to spend, in the first case) and the variance at completion (VAC = BAC − EAC = −€50,000).

Which formula should you use? Be careful with the optimistic one. A study of 155 US defence contracts found that by the time they were 20% complete, the cumulative CPI had already stabilised on 134 of them, and earlier research found it rarely moved by more than 10% after that point. Small commercial projects are less rigid than defence programs, but the lesson holds: cost problems tend to persist. "The rest will go to plan" is a claim that needs evidence, such as a clearly one-off cause that has been fixed.

Is the budget still reachable?

The to-complete performance index (TCPI) asks: to finish on the original budget, how efficient does the remaining work need to be?

The team has been working at 0.80. To hit the budget, it would need to work at 1.12 for the rest of the project, 40% better than so far. Unless something big changes, that's not a plan, it's a hope. The U.S. Government Accountability Office uses a simple test: if the TCPI is more than 0.05 above the current CPI, the forecast is too optimistic. A TCPI well above the current CPI is the moment to re-baseline the budget or cut scope, and to say so early.

The SPI trap, and earned schedule

SPI has a known flaw. At the end of every project, all the work is done, so EV = PV = BAC and SPI returns to 1.0, even if the project finished six months late. In the last third of a project, SPI drifts towards 1.0 and stops telling you much.

Earned schedule fixes this by measuring schedule in time instead of money. Ask: at what point in the plan should we have earned €60,000? The plan reached €44,000 at month 3 and €80,000 at month 4, so €60,000 falls at about month 3.4.

So the program is about 0.6 months behind today and heading for roughly 11.6 months instead of 10. Unlike SPI, SPI(t) stays meaningful right up to the end.

Setting up EVM on a real project

You don't need a defence contract to use it. A lightweight version needs five things:

1. A work breakdown with budgets. Every work package has an owner and a budget, and together they add up to the BAC.
2. A time-phased baseline. When each package is planned to be done. This gives you PV, month by month.
3. Objective progress rules. Agreed up front, like the 0/100 and 50/50 rules above.
4. Actual costs by work package. Timesheet codes that match the breakdown, so AC lines up with EV.
5. A monthly rhythm with thresholds. For example, any CPI or SPI below 0.9 triggers a short review of cause and recovery options.

Formal EVM systems for government contracts follow the EIA-748 standard, whose 2026 revision condenses the original 32 guidelines into 27. For most commercial projects, the five points above get you most of the value.

Build the schedule first: the critical path method →EVM tells you whether you're behind. The critical path tells you which delays actually matter.

Sources

• U.S. Government Accountability Office (2020). Cost Estimating and Assessment Guide (GAO-20-195G).
• NASA (2021). Earned Value Management Implementation Handbook.
• SAE International (2026). EIA-748-E, Earned Value Management Systems.
• Christensen, D. S. & Heise, S. R. (1993). Cost performance index stability. National Contract Management Journal, 25.
• Lipke, W., Zwikael, O., Henderson, K. & Anbari, F. (2009). Prediction of project outcome. International Journal of Project Management, 27(4).

The program and its numbers are illustrative.