Earned value with an unfunded change and a commitment

Calculation

• CPI = EV / AC = 980,000 / 1,150,000 = 0.85
• SPI = EV / PV = 980,000 / 1,200,000 = 0.82

What they tell you

• For every £1 spent the project is earning 85p of value; roughly a 15% cost inefficiency on work performed to date (CV = EV − AC = −£170,000).
• The project is delivering 82% of the value it planned to have delivered by now (SV = EV − PV = −£220,000).

What they do not tell you

1. Nothing about the critical path. SPI is value-weighted across all work; a project can have SPI of 0.82 while the critical path is on time, if the shortfall sits in non-critical packages. Always check milestone slippage directly.
2. Nothing about the future. CPI is a historical average. Here the variance is concentrated in one work package caused by a one-off interface error. If that rework cannot recur, CPI understates likely future performance - which is exactly why a bottom-up estimate to complete matters.
3. Nothing about commitments. CPI uses cost incurred, not cost committed. The £180,000 purchase order is a commitment, not a cost, and must not be added to AC.
4. Nothing about scope. The £90,000 approved change is not yet in the baseline, so EV is being measured against an out-of-date plan.

If you did add the commitment to AC

CPI would fall to 980,000 / 1,330,000 = 0.74 and the forecast would inflate by hundreds of thousands. That is a measurement error, not a performance signal: no work has been done for that money.

Step 1 - restate the baseline

Revised budget at completion: RBAC = 2,400,000 + 90,000 = £2,490,000. The approved change is real scope and must be in the baseline it is measured against.

Step 2 - formula forecasts

Typical variance (variance continues at the current rate):

• EAC = RBAC / CPI = 2,490,000 / 0.852 = ≈ £2,922,000
• Equivalently AC + (RBAC − EV) / CPI = 1,150,000 + 1,510,000 / 0.852 = 1,150,000 + 1,772,000 ≈ £2,922,000

Atypical variance (the rework was one-off, future work proceeds at budget):

• EAC = AC + (RBAC − EV) = 1,150,000 + (2,490,000 − 980,000) = £2,660,000

Step 3 - bottom-up cross-check

• AC + ETC = 1,150,000 + 1,520,000 = £2,670,000

Step 4 - reconcile and report

The bottom-up figure (£2,670,000) is within £10,000 of the atypical formula and £252,000 below the typical formula. That is strong independent evidence that the variance is largely non-recurring: the team, who know the remaining work, are effectively saying 'the rest will run to plan'. Report EAC ≈ £2,660,000–£2,670,000, use £2,670,000 as the single point, and quote £2,922,000 as the 'if it goes wrong again' upside. Headline overrun: £170,000–£180,000 on RBAC of £2,490,000, about 7%.

Treatment of the two items

• £180,000 purchase order: exclude from AC (it is a commitment, not a cost, and no value has been earned against it). Include it in the cash-flow and funding forecast - the funding requirement this year is higher than the EAC progression implies. It is already inside the RBAC; do not add it again.
• £90,000 change: add to BAC to form RBAC, never to AC. Note it is approved but unfunded in the baseline - that is a governance defect to fix now, not at year end.

Caveat to state explicitly: the £1,520,000 ETC excludes hardware. Confirm the hardware is in neither AC nor ETC before signing the forecast, or EAC is understated by £180,000.

Recommend against it. Distinguish two different things, because they get conflated:

• Updating the baseline for an approved change - legitimate and overdue. The £90,000 change is approved scope; the cost baseline should be revised to £2,490,000 through change control, with the original retained in configuration records.
• Re-baselining - replacing the approved performance measurement baseline with the current forecast so that variance disappears.

Why re-baselining is the wrong move here

1. It destroys the performance record. The £170,000 cost variance and the schedule variance are the evidence that the interface error happened and that the estimate was optimistic. Once reset, the organisation learns nothing and repeats it on the next project.
2. It hides an emerging trend. Re-baselining is usually requested when the number is uncomfortable; the discomfort is the signal.
3. It is a governance decision, not a management one. The baseline was authorised by the sponsor/board as the measure of success. Changing it retrospectively without the same level of authorisation is a control failure.
4. It converts a reported overrun into a silent one, which will resurface as a funding shortfall at the next stage gate.

When re-baselining is legitimate

• Scope has been fundamentally removed or the project re-authorised with a new mandate.
• The original baseline was never valid (e.g. built on estimates formally recorded as provisional).
• A stage boundary with a formal re-plan and re-approval.

What to do instead

• Revise the baseline for the approved change only, and record it as a controlled revision.
• Report against RBAC with a clear variance narrative and a named root cause (interface specification control).
• Present the EAC range to the sponsor as a decision: fund the £170,000–£180,000 forecast overrun from contingency, or reduce scope by an equivalent amount.
• Add the forecast overrun to the risk register and cost report so it is visible at the next gate, rather than absorbed silently into a new baseline.