SPI Versus SPI(t) on a Late Rail-Signaling Project
At the end of month 8 on a 24-month rail-signaling project, the control account reports PV = $5.0M, EV = $4.1M, AC = $4.6M, and BAC = $12.0M. The contractor's monthly report shows SPI = 0.82 and CPI = 0.89, and forecasts EAC = BAC / CPI = $13.47M. The project manager says the schedule is 'recoverable' because the critical path has only six months of remaining work and the team will work overtime. The scheduler's independent earned schedule analysis gives SPI(t) = 0.78, an earned schedule of 6.24 months at the data date, and an independent forecast duration of 30.8 months. The client's governance board wants one forecast completion date and one EAC, with confidence levels. The schedule contains 18 open-ended activities and a 30-day lag on the critical path.
1. Interpret the EVM and earned schedule results. Why do SPI and SPI(t) differ, and which is more appropriate for forecasting completion?
Read the worked answer
SPI = EV/PV = 4.1/5.0 = 0.82; CPI = EV/AC = 4.1/4.6 = 0.891. SPI is cost-weighted: it compares earned value to planned value at the data date. SPI(t) is time-based: earned schedule is the planned time at which EV should have been earned; here ES = 6.24 months against an actual data date of 8 months, so the project is about 1.76 months behind in time. SPI(t) = ES/AD = 6.24/8 = 0.78. The two differ because cost-loaded planned value is not linear in time; near the end of a project SPI converges to 1.0 even when the project is late, so SPI is a poor schedule forecast. For completion forecasting, use SPI(t) or the critical path remaining duration. A simple time-based forecast is PD/SPI(t) = 24/0.78 = 30.8 months. The PM's 'six months remaining' claim must be checked against the longest path after removing open ends and the 30-day lag; if the critical path is contaminated, the remaining duration is understated.
2. Assess the PM's claim that the schedule is recoverable through overtime. What schedule and cost evidence would you require?
Read the worked answer
Recoverability through overtime requires evidence: available craft hours, overtime productivity loss, resource histogram, and a corrected critical path. The 18 open-ended activities and the 30-day lag are schedule-quality defects; they can make the critical path look shorter than it is. Require a schedule update with all open ends closed, lags converted to activities, and constraints removed, then re-run CPM. Compare the remaining longest-path duration with the six-month claim. Cost evidence: overtime premium, expected productivity loss, and the EAC impact. If the schedule is already 1.76 months behind and the critical path is actually longer, overtime alone may not recover without adding resources. The PM must show a resource-loaded recovery plan, not an assertion. Also test whether the remaining work is on the critical path; overtime on noncritical work does not shorten completion.
3. Recommend what to report to the governance board, including forecast completion and EAC, and the caveats.
Read the worked answer
Report a range, not a single point. The independent time forecast is about 30.8 months against a 24-month plan, or roughly 6.8 months late; the EAC at the current CPI is $12.0M / 0.891 = $13.47M. State that the EAC is cost-based and assumes the CPI remains stable; if overtime is added, EAC will rise. The schedule forecast assumes the critical path is corrected and no further risk occurs. Caveats: 18 open ends, a 30-day lag, and schedule-quality defects mean the forecast has low confidence. Recommend a schedule risk analysis and a management reserve. Report SPI(t) = 0.78, not just SPI = 0.82, and explain that SPI can converge to 1.0 near completion. Give the board a P50/P80 range and the conditions for recovery, including the amount of overtime, the resource additions, and the schedule-quality corrections required before the forecast can be relied upon.