Your project shows a CPI of 0.97 and an SPI of 0.95. Green on every dashboard. Management is comfortable. Six months later the project is 40% over budget and a year behind schedule.
The earned value numbers were technically correct. The interpretation was wrong, and the reporting made the wrong interpretation easy. On mega-projects this pattern repeats so reliably that it is worth taking apart properly.
Three ways the numbers mislead
Claimed progress is not physical progress. If earned value is based on contractor self-reported progress without independent verification, the CPI is fiction. We have seen projects where claimed progress ran 25% ahead of physical progress for months before anyone noticed. The index was calculated flawlessly from a false input.
Front-loaded baselines flatter Year 1. When the performance measurement baseline allocates disproportionate value to early, easy activities, the project looks healthy in its first year and collapses in its second. The CPI was never real. It was an artefact of the baseline structure.
SPI is blind to the critical path. SPI tells you whether value is being earned on time. It does not tell you whether the critical path is slipping. A project can hold an SPI of 1.0 while its critical path runs three months late, provided enough non-critical work is being completed. Near completion, SPI drifts back toward 1.0 regardless of performance, which is precisely when schedule risk peaks.
None of these are failures of earned value management. They are failures of inputs, baseline construction and interpretation. Which is to say, failures of reporting.
The four questions an executive is actually asking
A steering committee does not need the definition of CPI. It needs four answers, in order, on one page.
Where are we? Cost and schedule position, stated plainly, with the variance in currency and weeks rather than indices alone. R80 million adverse lands differently than 0.94.
Where are we going? The estimate at completion, presented as a range rather than a point. A single EAC is a fiction of precision. Calculate it several ways: CPI-based, combined CPI times SPI, and a bottom-up build from the remaining work. When the methods converge, say so; that is confidence. When they diverge, say so louder; the spread is itself an early warning.
What changed since last period? Trends carry more information than positions. A stable 0.92 is a known problem being managed. A 1.05 that has slid to 0.97 over four periods is a project changing character, and it deserves more attention than the healthier-looking number suggests.
What decision is needed? Every report should end with either "no action required" or a specific ask with a date. A report that ends in neither is a document, not an instrument.
The early-warning signals worth escalating
The indices themselves are lagging indicators. The leading signals live one layer down, and a good reporting pack surfaces them before they reach the headline numbers:
- Three or more consecutive periods of CPI or SPI deterioration, however small each step.
- A widening gap between claimed progress and independently verified physical progress.
- A falling baseline execution index while SPI holds, meaning planned work is being skipped and backfilled with easier scope.
- EAC methods diverging from one another period on period.
- Contingency draw-down running faster than progress.
Any one of these, sustained, is a project asking for intervention months before the dashboard turns amber.
Cadence: matching the rhythm to the decision
On mega-projects the reporting rhythm matters as much as the content. Weekly, the controls team works the data: update quality, progress verification, logic health. This is internal and unglamorous, and it is where report integrity is won. Monthly, the full earned value cycle runs: variance analysis, EAC range, trends, and the one-page executive summary with its decision line. Quarterly, step back: baseline integrity review, independent verification of physical progress, and an honest test of whether the performance measurement baseline still represents the work.
The discipline this cadence protects is simple. Earned value is only as current as its inputs, and every week of verification lag is a week added to management's reaction time.
The standard worth holding
Earned value management remains the most powerful tool in project controls. But only when the inputs are honest, the baseline is sound, and the analysis goes beyond the headline indices. The gap between an EVM system that satisfies a contract clause and one that changes decisions is not the mathematics. It is whether the person reading page one can see the trend, the range and the required decision without asking for page two.
Executives do not ignore earned value because they misunderstand it. They ignore reports that bury the signal. Fix the report, and the numbers start earning their keep.
Faolan sets up and runs earned value systems on capital projects, from baseline construction to executive reporting. If your EVM produces numbers but not decisions, contact us.