New to this topic?
We recommend reading these guides first to get the most out of this one:
3
Base Measurements
10+
Derived Metrics
CPI×SPI
Best EAC Predictor
1.0
Target Index
Work earned against money spent Plot work earned against money spent at one scale and the whole metric set collapses into one geometry: this program has flown a ray of slope 0.765, and the only route left to the $120M corner is a ray of slope 1.308 — steeper than the 45° line it has never once touched.

Every figure is the worked example's own: BAC $120M, and at month 18 PV $65M, EV $52M, AC $68M. Both axes run from zero at the same 3 px per $M, so a dollar is as wide as it is tall, the 45° line is CPI = 1.0, and every gap drawn here can be measured off the picture in dollars. The program's ray has slope EV ÷ AC = 52 ÷ 68 = 0.765 = CPI. Extend it to EV = BAC and it arrives at AC = 120 ÷ 0.765 = $156.9M = EAC; the run from today to there is ETC = 156.9 − 68 = $88.9M, and the overshoot past BAC is VAC = 120 − 156.9 = −$36.9M. The vertical drop from the 45° line at today is CV = 52 − 68 = −$16M; the drop from the PV mark is SV = 52 − 65 = −$13M. Where the ray crosses AC = $120M the budget is spent and only 0.765 × 120 = $91.8M of work has been earned — 76.5%, which is what CPI means; the tinted wedge between the 45° line and the ray is that shortfall, widening from zero to $16M at today and $28.2M by the time the money runs out. The green ray is the only route from today to the corner: slope (120 − 52) ÷ (120 − 68) = 68 ÷ 52 = 1.308 = TCPI. The track up to today is drawn straight because the article reports cumulative figures only; its slope is the cumulative CPI and it asserts nothing more.

The Complete Metrics Reference

MetricFormulaWhat It Tells YouGood/Bad
PV (Planned Value)From baselineHow much work should be done by nowReference only
EV (Earned Value)% complete × BACHow much work IS done (in budget terms)Higher is better
AC (Actual Cost)From accountingHow much we have spentLower is better
BAC (Budget at Completion)Total baseline budgetHow much the whole job should costReference only
CV (Cost Variance)EV – ACOver or under budgetPositive = under budget
SV (Schedule Variance)EV – PVAhead or behind schedulePositive = ahead
CPI (Cost Performance Index)EV ÷ ACDollars of work per dollar spent>1.0 = under budget
SPI (Schedule Performance Index)EV ÷ PVWork accomplished vs. planned pace>1.0 = ahead
EAC (Estimate at Completion)Multiple formulasWhat the total job will actually costLower is better
ETC (Estimate to Complete)EAC – ACHow much more we need to spendLower is better
VAC (Variance at Completion)BAC – EACExpected overrun/underrun at finishPositive = underrun
TCPI (To-Complete Performance Index)(BAC – EV) ÷ (BAC – AC)Required future CPI to finish on budget>1.0 = must improve

Worked Example: All Metrics

📊 Program Status at Month 18 Complete Metric Set

Program: 36-month duration. BAC = $120M.

At month 18: PV = $65M, EV = $52M, AC = $68M.

MetricCalculationResultInterpretation
CV$52M – $68M–$16M$16M over budget
SV$52M – $65M–$13M$13M behind schedule
CPI$52M ÷ $68M0.765Getting 76.5 cents of work per dollar spent
SPI$52M ÷ $65M0.800Accomplishing 80% of planned work rate
EAC (CPI)$120M ÷ 0.765$156.9MProgram will cost $156.9M if current efficiency continues
ETC$156.9M – $68M$88.9MStill need $88.9M to finish
VAC$120M – $156.9M–$36.9MExpected $36.9M overrun at completion
TCPI($120M – $52M) ÷ ($120M – $68M)1.308Must achieve CPI of 1.31 on remaining work to finish on budget — extremely unlikely

Bottom line: This program is in serious trouble. CPI of 0.765 at 50% completion has never recovered in DoD history. The TCPI of 1.308 means the program would need to become 31% more efficient than budget on all remaining work — effectively impossible without a fundamental scope or approach change.

Earned Value Techniques

The hardest part of EVMS is objectively measuring “percent complete.” Several techniques exist, each appropriate for different work types:

TechniqueHow It WorksBest For
0/1000% until complete, then 100%Short-duration tasks (≤2 months)
50/5050% when started, 100% when completeShort tasks where start is a meaningful milestone
Milestone WeightedAssign % to intermediate milestones (e.g., 25% at PDR, 50% at CDR)Engineering tasks with defined review gates
Percent CompleteSubjective assessment of completionUse sparingly — prone to “90% done syndrome”
Units CompleteUnits delivered ÷ total units plannedRecurring production (aircraft, engines, assemblies)
Level of Effort (LOE)EV = PV (always on plan by definition)Support activities (PM, admin) that cannot be objectively measured. Minimize LOE — it masks performance.

⚠️ The LOE Trap

LOE work always has CPI = 1.0 and SPI = 1.0 by definition — because EV is set equal to PV. If a large portion of your program is LOE, your indices look better than reality. Programs with >30% LOE are effectively hiding their true performance. Push to convert LOE to discrete work packages with objective earned value measurement wherever possible.

🎯 The Bottom Line

Three base measurements (PV, EV, AC) drive the entire EVMS analysis framework. CPI tells you cost efficiency. SPI tells you schedule efficiency. EAC tells you where you are headed. TCPI tells you what miracle you need to get back on budget. The worked example shows how these metrics turn raw numbers into actionable insight — at month 18, this program’s TCPI of 1.31 tells management they cannot finish on budget without a fundamental change. That is the power of EVMS: early warning, not post-mortem. Next: CPI & SPI Analysis — trend analysis and what the indices really mean over time.

Interactive Demo

Set the budget, the plan, the progress and the spend, and watch every metric on this page fall out of those four numbers at once.

⚡
Try It Yourself
Earned Value Management Calculator
▼
Adjust the budget, planned progress, actual progress, and actual cost to see how EVM metrics change. CPI < 1 means over budget; SPI < 1 means behind schedule.
$500K
1000001000000
60%
10%100%
45%
0%100%
$280K
0500000
Cost
Over Budget
Schedule
Behind Schedule
$0$200K$400K$300KPVPlanned Value$225KEVEarned Value$280KACActual Cost
CV = EV - AC = $-55000
SV = EV - PV = $-75000
EAC = BAC / CPI = $622K
ETC = EAC - AC = $342K
0.80
CPI
0.75
SPI
$622K
EAC
$-122222
VAC
1.25
TCPI
Ready for the full knowledge check? Test your understanding with guided scenarios and data export.
PROTake the Pro Knowledge Check →
Free forever · Every feature included

Stop reading, start modeling

Model your process flow, run simulations, optimize staffing with TOC math, and test your knowledge with 107 interactive checks — all in one platform.

Open Workbench →

Take this to a room

The running order

For CAMs and programme managers. They should leave able to read CPI, SPI, EAC and TCPI on their own programme and to spot the LOE trap.

7 beats · 13 min
  1. 1

    Three bases, ten derivations

    PV, EV and AC. Everything else on the report is arithmetic on those three.

    • CV = EV minus AC. SV = EV minus PV.
    • CPI = EV over AC. SPI = EV over PV.
    • One is the target for both indices.
  2. 2

    What each index actually tells you

    Say it in plain words before anyone opens a spreadsheet.

    • CPI 0.90: we are getting ninety cents of work for every dollar spent.
    • SPI 0.85: we have completed 85 per cent of the work we planned by now.
    • Neither says whose fault it is. They say where to look.
  3. 3

    EAC is where it is heading

    Estimate at completion converts today's performance into tomorrow's number.

    • BAC is what we said it would cost. EAC is what it is going to.
    • VAC = BAC minus EAC - the variance we are currently heading for.
    • ETC is what remains from here.
  4. 4

    TCPI is the reality check

    To-complete performance index: the efficiency you would need on every remaining dollar to hit the target.

    • A worked programme at month 18 with a TCPI of 1.31.
    • That says finishing on budget requires 31 per cent better performance than the track record.
    • Above 1.10 the target is a fantasy. Say that out loud in the review.

    Ask the room What is our TCPI, and does anyone believe it?

  5. 5

    The LOE trap

    Level of effort work always has CPI and SPI of exactly 1.0, because EV is set equal to PV.

    • It is not performing well. It is definitionally incapable of showing a variance.
    • A programme with more than 30 per cent LOE is hiding its real performance.
    • Push to convert LOE into discrete work packages with objective measurement.

    Ask the room What proportion of our budget is LOE?

  6. 6

    Earned value techniques matter

    How you claim EV determines whether any of the indices mean anything.

    • 0/100, 50/50, percent complete, milestone-based, apportioned, LOE.
    • Discrete, objective techniques give honest indices.
    • Percent complete by opinion gives you the number someone wanted.
  7. 7

    What good looks like

    The whole point is turning raw numbers into a decision this month.

    • CPI for cost efficiency, SPI for schedule efficiency.
    • EAC for where you are heading, TCPI for whether the target is real.
    • Early warning, not a post-mortem. If the report only confirms what everyone already knew, it was late.

    Ask the room When did our last EVMS report tell us something we did not already know?