New to this topic?
We recommend reading these guides first to get the most out of this one:
Earned Value Metrics
1.0
On-Plan Target
±0.10
CPI Stability Range
20%
Prediction Point
0.90
Red Threshold

Cumulative vs. Current-Period Indices

Index TypeFormulaBehaviorUse For
Cumulative CPIΣEV ÷ ΣACStable, smooths out monthly variation, converges over timeEAC prediction, program health assessment, trend direction
Current-Period CPIMonthly EV ÷ Monthly ACVolatile, sensitive to monthly events (large procurements, milestone completions)Detecting recent changes, identifying emerging problems, validating recovery claims
Cumulative SPIΣEV ÷ ΣPVStable but converges to 1.0 at program end (all planned work eventually done)Mid-program schedule assessment. Less useful near program end.
Current-Period SPIMonthly EV ÷ Monthly PVVolatile but reveals current execution rateMonthly performance assessment, recovery tracking

💡 SPI Converges to 1.0 — Use Schedule Variance Instead

At program completion, all planned work is done, so cumulative EV = BAC = cumulative PV, and SPI = 1.0 regardless of whether the program was 2 years late. This makes SPI less useful for schedule assessment in the program’s final third. Use time-based schedule metrics (slip in months, critical path analysis) alongside SPI for schedule health assessment.

The CPI Stability Rule

The most important empirical finding in EVMS research: cumulative CPI stabilizes after the 20% completion point and rarely improves by more than 0.10 thereafter.

📊 CPI Trend Scenario Program Alpha
Month% CompleteCum CPICum SPIAssessment
610%1.050.92Early — CPI looks good but SPI showing schedule pressure. Indices still volatile.
1222%0.940.88At the 20% point. CPI 0.94 = expect final CPI between 0.84–1.04. SPI declining.
1838%0.910.85CPI continuing to decline. Now expect final CPI 0.81–1.01. Trend is negative — not recovering.
2455%0.880.82Red flag. CPI 0.88 at 55% complete = EAC will be ~14% over budget. Recovery is statistically improbable.

Action at month 24: With CPI at 0.88 and declining, the program team should: (1) develop a realistic EAC using multiple methods, (2) identify the root causes of cost overrun (are we inefficient, or was the baseline wrong?), (3) present management with options: rebaseline, descope, or accept the overrun.

Reading the Trend

PatternWhat It MeansAction
Stable CPI > 0.95Program is executing on or near budgetMaintain. Monitor for emerging issues.
CPI declining month-over-monthGetting worse. New cost drivers are emerging or existing ones are growing.Root cause analysis. Variance Analysis Report on largest CA contributors.
CPI stable at 0.85–0.95Program has a structural efficiency problem. Will not self-correct.Management intervention required. Realistic EAC. Corrective action plan with measurable milestones.
CPI < 0.85Severe overrun. Recovery extremely unlikely without scope or baseline change.Formal rebaseline assessment. Scope trade study. Customer notification.
Current CPI improving but cumulative flatRecent performance is better but not enough to move the cumulative needle.Validate improvement is real (not timing). Check if recovery rate is sufficient to change EAC trajectory.
SPI < 1.0 with CPI > 1.0Behind schedule but under budget. May be “buying schedule” with less work than planned.Examine whether under-budget means underperformance. Check if work is being deferred.

Reporting Thresholds

LevelThresholdAction Required
GreenCPI/SPI ≥ 0.95Normal reporting. Monitor trends.
YellowCPI/SPI 0.90–0.95Variance analysis required. Identify root causes. Corrective action plan within 30 days.
RedCPI/SPI < 0.90Formal CAR/VAR. Management review. Customer notification. Recovery plan or rebaseline assessment.

⚠️ Don’t Manage Indices — Manage the Work

A common dysfunction: program teams focus on “improving CPI” by manipulating earned value claims (claiming more % complete than warranted) or deferring cost recognition. This makes the indices look better temporarily but does not change the reality. Manage the work: reduce rework, improve productivity, control scope. The indices will follow.

🎯 The Bottom Line

CPI and SPI are the vital signs of your program. Cumulative CPI stabilizes early and predicts the future: if it is 0.88 at the 20% point, plan for a final CPI of 0.78–0.98. Current-period indices reveal whether you are getting better or worse. Use both together: cumulative for prediction, current-period for trend detection. And remember — the indices are symptoms, not causes. When they signal a problem, the response is root cause analysis and corrective action on the work, not manipulation of the numbers. Next: Variance Analysis & Reporting — turning index signals into actionable root cause narratives.

Keep Learning
📖
Guide
Variance Analysis & Reporting
📖
Guide
Estimate at Completion Methods
📖
Guide
Earned Value Metrics
🎓 Part of the EVMS Reporting track — view all 8 guides →
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 programme leadership reading monthly indices. They should leave able to distinguish cumulative from current-period and to know when SPI stops being useful.

7 beats · 12 min
  1. 1

    Two indices, two jobs

    Cumulative predicts. Current-period detects. Using one for both is the common mistake.

    • Cumulative CPI is stable and forecasts the finish.
    • Current-period CPI tells you whether this month was better or worse.
    • A stable cumulative with a falling current period is a programme about to turn.

    Ask the room Which one does our report show?

  2. 2

    CPI stabilises at twenty per cent

    This is the most useful empirical finding in EVMS, and it is not widely known.

    • Once about 20 per cent of the work is complete, cumulative CPI barely moves.
    • Final CPI is almost always within plus or minus 0.10 of the value at that point.
    • A CPI of 0.88 at the 20 per cent point means plan for 0.78 to 0.98 at the end.
  3. 3

    So the early number is the forecast

    That changes what an early bad month means. It is not noise you will recover from.

    • Programmes routinely promise to recover a first-year CPI. Very few do.
    • Recovery requires a structural change, not effort.
    • If you are not making that change, the honest EAC is the statistical one.

    Ask the room What did we promise to recover, and what specifically changed to make it possible?

  4. 4

    SPI lies at the end

    At completion all planned work is done, so cumulative EV equals BAC equals PV. SPI reads 1.0 whatever happened.

    • A programme two years late finishes with an SPI of 1.0.
    • In the final third, SPI is close to useless for schedule health.
    • Use time-based measures alongside it: slip in months, critical path analysis.
  5. 5

    Thresholds and escalation

    Decide in advance what number triggers what, or every review re-litigates it.

    • Around 0.90 is the common red threshold.
    • Plus or minus 0.10 is the normal stability band.
    • Set the trigger before you need it, so the conversation is about the work.
  6. 6

    Do not manage the indices

    The dysfunction to name explicitly, because it is common and it is quiet.

    • Claiming more percent complete than warranted improves CPI without improving anything.
    • Deferring cost recognition does the same.
    • It makes the number look better temporarily and changes nothing real.

    Ask the room Would we notice if this were happening on our programme?

  7. 7

    The indices are symptoms

    When they signal a problem, the response is root cause analysis on the work.

    • Reduce rework, improve productivity, control scope.
    • The indices follow the work. The work never follows the indices.
    • Cumulative for prediction, current period for trend, always together.

    Ask the room What corrective action are we taking on the work itself?

Continue the EVMS Track

📋 Variance Analysis
Root cause narratives for cost/schedule variances
📊 EAC Methods
Estimating final cost from performance data
📊 EV Metrics
Complete metrics reference