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
| Metric | Formula | What It Tells You | Good/Bad |
|---|---|---|---|
| PV (Planned Value) | From baseline | How much work should be done by now | Reference only |
| EV (Earned Value) | % complete × BAC | How much work IS done (in budget terms) | Higher is better |
| AC (Actual Cost) | From accounting | How much we have spent | Lower is better |
| BAC (Budget at Completion) | Total baseline budget | How much the whole job should cost | Reference only |
| CV (Cost Variance) | EV – AC | Over or under budget | Positive = under budget |
| SV (Schedule Variance) | EV – PV | Ahead or behind schedule | Positive = ahead |
| CPI (Cost Performance Index) | EV ÷ AC | Dollars of work per dollar spent | >1.0 = under budget |
| SPI (Schedule Performance Index) | EV ÷ PV | Work accomplished vs. planned pace | >1.0 = ahead |
| EAC (Estimate at Completion) | Multiple formulas | What the total job will actually cost | Lower is better |
| ETC (Estimate to Complete) | EAC – AC | How much more we need to spend | Lower is better |
| VAC (Variance at Completion) | BAC – EAC | Expected overrun/underrun at finish | Positive = 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: 36-month duration. BAC = $120M.
At month 18: PV = $65M, EV = $52M, AC = $68M.
| Metric | Calculation | Result | Interpretation |
|---|---|---|---|
| CV | $52M – $68M | –$16M | $16M over budget |
| SV | $52M – $65M | –$13M | $13M behind schedule |
| CPI | $52M ÷ $68M | 0.765 | Getting 76.5 cents of work per dollar spent |
| SPI | $52M ÷ $65M | 0.800 | Accomplishing 80% of planned work rate |
| EAC (CPI) | $120M ÷ 0.765 | $156.9M | Program will cost $156.9M if current efficiency continues |
| ETC | $156.9M – $68M | $88.9M | Still need $88.9M to finish |
| VAC | $120M – $156.9M | –$36.9M | Expected $36.9M overrun at completion |
| TCPI | ($120M – $52M) ÷ ($120M – $68M) | 1.308 | Must 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:
| Technique | How It Works | Best For |
|---|---|---|
| 0/100 | 0% until complete, then 100% | Short-duration tasks (≤2 months) |
| 50/50 | 50% when started, 100% when complete | Short tasks where start is a meaningful milestone |
| Milestone Weighted | Assign % to intermediate milestones (e.g., 25% at PDR, 50% at CDR) | Engineering tasks with defined review gates |
| Percent Complete | Subjective assessment of completion | Use sparingly — prone to “90% done syndrome” |
| Units Complete | Units delivered ÷ total units planned | Recurring 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.
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.