11.2 EVM Variances & Performance Indices (CV, SV, CPI, SPI)

11.2 EVM Variances & Performance Indices (CV, SV, CPI, SPI)

This section covers 4.D track cost performance and 4.F perform cost/schedule control analysis. The blueprint's verbs matter again: 4.D is the running record of cost performance, while 4.F is the analytical step that turns variances and indices into a control decision.

Once the three fundamental Earned Value Management metrics—Planned Value ($PV$), Earned Value ($EV$), and Actual Cost ($AC$)—are established at the Data Date, cost engineers calculate derived variances and performance indices. These metrics quantify past performance, detect emerging cost and schedule trends, and serve as mathematical inputs for project forecasting.

For Certified Cost Professional (CCP) candidates, mastering the formulas, interpretation of dimensionless indices, variance percentage derivations, diagnostic performance quadrants, and the theory of Earned Schedule ($ES$) is essential.

---

1. Absolute Variances: Cost Variance & Schedule Variance

Variances in EVM measure the absolute divergence between actual accomplishment and baseline expectations. By standard convention across ANSI/EIA-748 and AACE International, favorable variances are positive (+) and unfavorable variances are negative (-).

+-----------------------------------------------------------------------------+
|                        EVM ABSOLUTE VARIANCE FORMULAS                       |
|                                                                             |
|   1. COST VARIANCE (CV):                                                    |
|      CV = EV - AC  (or BCWP - ACWP)                                         |
|      • CV > 0 (+): Under Budget (Favorable - earned more than spent)        |
|      • CV = 0: Exactly on Budget                                            |
|      • CV < 0 (-): Over Budget (Unfavorable - spent more than earned)       |
|                                                                             |
|   2. SCHEDULE VARIANCE (SV):                                                |
|      SV = EV - PV  (or BCWP - BCWS)                                         |
|      • SV > 0 (+): Ahead of Schedule (Favorable - earned more than planned) |
|      • SV = 0: Exactly on Schedule                                          |
|      • SV < 0 (-): Behind Schedule (Unfavorable - earned less than planned) |
+-----------------------------------------------------------------------------+

[!IMPORTANT] Critical Exam Distinction: Schedule Variance is Expressed in Dollars or Hours! Although termed Schedule Variance, $SV = EV - PV$ is measured in monetary units ($) or labor hours, not calendar time (days, weeks, months). $SV$ represents the monetary value of work accomplished compared to the monetary value of work planned. It does not directly indicate the number of days a project is delayed.

2. Performance Indices: CPI and SPI

Performance indices express project efficiency as dimensionless ratios of output to input, standardizing performance comparisons across projects of different dollar magnitudes.

+-----------------------------------------------------------------------------+
|                     EVM EFFICIENCY PERFORMANCE INDICES                      |
|                                                                             |
|   1. COST PERFORMANCE INDEX (CPI):                                          |
|      CPI = EV / AC                                                          |
|      • CPI > 1.00: Cost efficient (e.g., CPI = 1.20 -> $1.20 earned per $1) |
|      • CPI = 1.00: On budget ($1.00 earned per $1.00 spent)                 |
|      • CPI < 1.00: Cost inefficient (e.g., CPI = 0.80 -> $0.80 earned per $1)|
|                                                                             |
|   2. SCHEDULE PERFORMANCE INDEX (SPI):                                      |
|      SPI = EV / PV                                                          |
|      • SPI > 1.00: Schedule efficient (progressing faster than plan)        |
|      • SPI = 1.00: Progressing exactly at planned baseline rate             |
|      • SPI < 1.00: Schedule inefficient (progressing slower than plan)      |
+-----------------------------------------------------------------------------+

Percentage Variances ($CV%$ and $SV%$)

In executive management and CPR Format 1 reports, variances are normalized into percentage deviations:

1. Percentage Cost Variance ($CV%$): $CV\% = \frac{CV}{EV} = \frac{EV - AC}{EV} = 1 - \frac{1}{CPI}$ (Note: An alternative convention in some cost engineering literature normalizes against baseline budget: $CV%_{\text{alt}} = CV / BAC$. In standard ANSI/EIA-748 variance reporting, $CV / EV$ is the official ratio).
2. Percentage Schedule Variance ($SV%$): $SV\% = \frac{SV}{PV} = \frac{EV - PV}{PV} = SPI - 1$

---

3. Diagnostic Analysis of the Four Performance Quadrants

Plotting $CPI$ against $SPI$ creates four diagnostic quadrants that reveal underlying root causes and execution strategies:

---

4. The Critical Flaw of SPI & The Theory of Earned Schedule ($ES$)

The "End-of-Project" SPI Anomaly

Traditional Schedule Performance Index ($SPI = EV / PV$) suffers from a fatal mathematical defect as a project nears completion:

1. At project completion, all required scope is eventually performed, meaning $EV$ converges to $BAC$ ($EV \to BAC$).
2. The baseline schedule also reaches its end, so $PV$ equals $BAC$ ($PV = BAC$).
3. Consequently, at the end of the project: $SV = EV - PV = BAC - BAC = 0$ $SPI = \frac{EV}{PV} = \frac{BAC}{BAC} = 1.00$

[!WARNING] The SPI Failure Mode: Even if a project is one year late, once all work is finished, $SV$ returns to $0 and $SPI$ returns to 1.00. Beyond 65% to 70% project completion, $SPI$ loses predictive validity because the denominator ($PV$) stops growing, forcing the ratio to improve artificially toward 1.00.

Earned Schedule ($ES$) Formulation

Developed by Walt Lipke in 2003, Earned Schedule ($ES$) bridges this gap by mapping the dollar value of Earned Value ($EV$) back onto the planned value curve to find the exact point in calendar time when that amount of work was supposed to have been achieved.

+-----------------------------------------------------------------------------+
|                        EARNED SCHEDULE (ES) MECHANICS                       |
|                                                                             |
|   Let:                                                                      |
|   • C = Number of whole time increments where EV >= PV_C                    |
|   • PV_C = Planned Value at time increment C                                |
|   • PV_(C+1) = Planned Value at time increment C + 1                        |
|   • AT = Actual Time elapsed (calendar time from start to Data Date)        |
|                                                                             |
|   Linear Interpolation Formula:                                             |
|   ES = C + [ (EV - PV_C) / (PV_(C+1) - PV_C) ]                              |
|                                                                             |
|   Time-Based Variances and Indices:                                         |
|   • Schedule Variance (Time):  SV(t) = ES - AT  (in months/weeks/days)      |
|   • Schedule Index (Time):     SPI(t) = ES / AT                             |
+-----------------------------------------------------------------------------+

Why $SPI(t)$ Outperforms Traditional $SPI$:

• When a project finishes late, $ES = \text{Planned Duration } (PD)$ and $AT > PD$.
• Therefore, at late completion, $SV(t) = PD - AT < 0$ and $SPI(t) = \frac{PD}{AT} < 1.00$.
• $SPI(t)$ correctly reflects late schedule performance throughout project closeout!

---

5. Comprehensive Step-by-Step Worked Numerical Problem

Consider an industrial EPC substation modernization project at Month 6 ($AT = 6.0\text{ months}$). The contract baseline budget is $BAC = $2,000,000$.

Time-Phased Cumulative Planned Value Baseline ($PV$):

• Month 1: $PV_1 = $150,000$
• Month 2: $PV_2 = $400,000$
• Month 3: $PV_3 = $750,000$
• Month 4: $PV_4 = $1,150,000$
• Month 5: $PV_5 = $1,550,000$
• Month 6: $PV_6 = $1,800,000$ (Current Planned Baseline at Data Date)
• Month 8: $PV_8 = $2,000,000 = BAC$ (Planned Duration $PD = 8.0\text{ months}$)

Performance Data at Month 6 (Data Date):

• Earned Value ($EV$): $$1,350,000$
• Actual Cost ($AC$): $$1,500,000$
• Actual Time Elapsed ($AT$): $6.0\text{ months}$

Step 1: Calculate Traditional Variances and Indices

1. Cost Variance ($CV$): $CV = EV - AC = \$1,350,000 - \$1,500,000 = -\$150,000 \quad (\text{Unfavorable / Over Budget})$
2. Schedule Variance ($SV$): $SV = EV - PV_6 = \$1,350,000 - \$1,800,000 = -\$450,000 \quad (\text{Unfavorable / Behind Schedule})$
3. Cost Performance Index ($CPI$): $CPI = \frac{EV}{AC} = \frac{\$1,350,000}{\$1,500,000} = \mathbf{0.900}$
4. Schedule Performance Index ($SPI$): $SPI = \frac{EV}{PV} = \frac{\$1,350,000}{\$1,800,000} = \mathbf{0.750}$
5. Percentage Variances: $CV\% = \frac{CV}{EV} = \frac{-\$150,000}{\$1,350,000} = -11.11\%$ $SV\% = \frac{SV}{PV} = \frac{-\$450,000}{\$1,800,000} = -25.00\%$

Step 2: Calculate Earned Schedule ($ES$) and Time-Based Performance

1. Identify time increments bounding $EV = $1,350,000$:
   • At Month 4: $PV_4 = $1,150,000$ (Since $$1,350,000 \ge $1,150,000$, $C = 4$).
   • At Month 5: $PV_5 = $1,550,000$ ($PV_{C+1} = $1,550,000$).
2. Interpolate Earned Schedule ($ES$): $ES = 4 + \frac{\$1,350,000 - \$1,150,000}{\$1,550,000 - \$1,150,000} = 4 + \frac{\$200,000}{\$400,000} = 4 + 0.50 = \mathbf{4.50\text{ months}}$
3. Calculate Time-Based Schedule Variance ($SV(t)$): $SV(t) = ES - AT = 4.50\text{ months} - 6.00\text{ months} = \mathbf{-1.50\text{ months}}$ (The project has accomplished 4.5 months of work in 6.0 months, representing an actual schedule delay of 1.5 months).
4. Calculate Time-Based Schedule Performance Index ($SPI(t)$): $SPI(t) = \frac{ES}{AT} = \frac{4.50}{6.00} = \mathbf{0.750}$

Performance Summary Table