Earned value analysis, worked through end to end

CV  = EV − AC        SV  = EV − PV
CPI = EV ÷ AC        SPI = EV ÷ PV
EAC = BAC ÷ CPI      ETC = EAC − AC
VAC = BAC − EAC
TCPI = (BAC − EV) ÷ (BAC − AC)

What earned value actually answers

Most progress reporting can be gamed by an optimist. "We're 60% done" is an opinion. "We've spent £96,000" is a fact about money, not about progress. Put them together and you still cannot say whether the project is in trouble, because neither number knows what the other one means.

Earned value fixes this by insisting that progress is measured in the same unit as the budget. Three numbers, all in currency:

• Planned value (PV) — what the schedule said you should have completed by today, priced at budget.
• Earned value (EV) — what you have actually completed, priced at that same budget.
• Actual cost (AC) — what completing it really cost.

The trick is that EV is priced at budget, not at what it cost. That single decision is what makes the comparison honest: EV against AC is a pure cost question, and EV against PV is a pure schedule question. Neither can hide inside the other.

Every earned value formula, worked through end to end

Take a project with a budget at completion (BAC) of 200,000. By today the schedule expected 120,000 of work to be finished. In fact 108,000 of budgeted work is finished, and it cost 96,000 to do.

Those are the four numbers in the calculator above. Here is every metric they produce, and what each one is telling you.

Step 1 — the two variances

Variances are subtractions, and they answer in currency.

CV = EV − AC = 108,000 − 96,000 = +12,000
SV = EV − PV = 108,000 − 120,000 = −12,000

Positive is good, negative is bad, in both cases. So this project is 12,000 under budget and 12,000 behind schedule. Note that the schedule variance is expressed in money — earned value has no time axis, which is a real limitation we will come back to.

Step 2 — the two indices

Indices are divisions, and they answer in efficiency. They matter more than the variances because they scale: a variance of 12,000 means something different on a 200,000 project than on a 20,000,000 one.

CPI = EV ÷ AC = 108,000 ÷ 96,000 = 1.125
SPI = EV ÷ PV = 108,000 ÷ 120,000 = 0.90

Read them as exchange rates. CPI 1.125 means every 1.00 spent has bought 1.125 of budgeted work — you are getting good value. SPI 0.90 means you are completing work at 90% of the planned rate.

Step 3 — the forecast

Now the useful part. If today's efficiency continues, what will the whole project cost?

EAC = BAC ÷ CPI = 200,000 ÷ 1.125 = 177,778
ETC = EAC − AC = 177,778 − 96,000 = 81,778
VAC = BAC − EAC = 200,000 − 177,778 = +22,222

So: expect to finish around 177,778, needing roughly 81,778 more, and to come in about 22,222 under the approved budget. That is a forecast you can put in front of a sponsor, and it is derived rather than guessed.

Cost performance index (CPI) and schedule performance index (SPI)

Neither index means much alone. Together they place the project in one of four positions, and each position calls for a different conversation.

The second row is the one worth dwelling on, because it is the example above and it is the one most often misread. A project can look under budget purely because it has not yet spent money on work it has not yet done. Cost efficiency and slow progress frequently arrive together, and the CPI is not the good news it appears to be. If you take one habit from this page, take this one: never report CPI without SPI beside it.

Estimate at completion (EAC): which formula to use

This is where most people go wrong, because there are four accepted formulas and they are not interchangeable. The formula encodes an assumption about the future, so choosing one is a forecasting judgement, not arithmetic.

1. Variances are typical — EAC = BAC ÷ CPI

Assumes the rest of the project performs like the part you have done. The right default on a project running steadily with no known change. In the example: 177,778.

2. Variances are atypical — EAC = AC + (BAC − EV)

Assumes the overrun so far was a one-off and everything remaining runs at plan. Use it when you can name the cause and show it is finished — a resolved supplier dispute, a rework cycle now closed. In the example: 188,000. Note it is the more pessimistic figure here, because it declines to assume today's favourable efficiency continues.

3. Cost and schedule both bite — EAC = AC + (BAC − EV) ÷ (CPI × SPI)

Assumes remaining work is dragged by both indices. Appropriate when a hard end date means lateness will be bought back with money. In the example: 186,864.

4. A fresh bottom-up estimate — EAC = AC + new ETC

When the original baseline no longer describes the project, no index will rescue it. Re-estimate the remaining work and add it to what you have spent. This is the honest answer after a major scope change, and no calculator can produce it for you.

The practical discipline: pick the formula before you look at which number you would prefer, and write down the assumption you are making. An EAC without its stated assumption is not a forecast, it is a preference.

TCPI — the reality check

Every forecast above tells you where you are heading. TCPI answers a harder question: what efficiency would the remaining work have to sustain to still land on budget?

TCPI = (BAC − EV) ÷ (BAC − AC) = 92,000 ÷ 104,000 = 0.88

The rule of thumb is unforgiving and worth memorising:

• TCPI ≤ 1.00 — achievable, no change in efficiency needed.
• TCPI up to about 1.10 — a stretch; name the specific actions that will deliver it.
• TCPI above 1.10 — treat as unrealistic. Teams almost never improve efficiency by more than 10% mid-flight without descoping.

Here TCPI is 0.88, comfortably below 1.00 — this project could get less efficient and still land on budget. Compare that to the optimism you often hear on a project running at CPI 0.85 with a TCPI of 1.3, where the plan silently requires the team to become 50% more effective than they have managed all year.

One trap: once AC reaches BAC, the denominator turns negative and TCPI stops meaning anything. A negative TCPI is not an easy target — it is an unreachable one. At that point the budget is spent with work outstanding, and the answer is to re-baseline and measure TCPI against an EAC instead.

Schedule performance index (SPI)

SPI has a structural flaw that every practitioner should know: it always returns to 1.00 at the end of a project, however late you finish. At completion EV equals BAC, and PV equals BAC too, because all the planned work has come due. SPI = BAC ÷ BAC = 1.00 — on a project delivered a year late.

The distortion starts well before the end. In the last third of a project, PV is running out faster than EV, so SPI drifts upward regardless of real performance. Treat SPI as reliable in the early and middle phases, and stop trusting it as a schedule signal near the finish.

The fix is earned schedule, which puts EV back on the time axis and reports in periods rather than currency. SPI(t) does not collapse to 1.00, so it keeps working when SPI has stopped.

Five mistakes that survive into real reporting

1. Using % spent as % complete. AC ÷ BAC is not progress — it is consumption. In the example 48% of the budget is spent but 54% of the work is done, and the gap is the entire point of the method.
2. Mixing currencies and units. PV, EV, AC and BAC must all be the same currency at the same baseline. A mixed-rate roll-up produces indices that look precise and mean nothing.
3. Reporting CPI alone. See the four-quadrant table above. Cost efficiency without schedule context is the most common way a report misleads honestly.
4. Letting EV be an opinion. Earned value is only as sound as the rule for claiming it. Fix the rule in advance — 0/100, 50/50, or physical measurement — and never let it be a judgement made at month end by the person being measured.
5. Re-baselining to make the numbers look better. A re-baseline resets the variances to zero. Sometimes that is right, because the plan genuinely no longer describes the project. It is never right as a reporting tactic, and reviewers can always see it in the history.

On the exam and on a real project

If you are preparing for the PMP, expect to be given three of the four numbers and asked for a fourth, and expect at least one question that hinges on choosing the right EAC formula from its stated assumption. The arithmetic is easy; the assumption is what is being tested.

In PRINCE2 terms, earned value is a tolerance instrument. CV and SV tell you whether you are still inside the cost and time tolerances delegated to you, and the forecast tells you whether you are about to breach them — which is what turns a routine Highlight Report into an Exception Report. Used properly it is an early-warning system, not a scorecard.

On a live project the numbers matter less than the cadence. Earned value calculated monthly on a twelve-month project gives you about three readings before it is too late to act. Calculated fortnightly, it gives you time.