Cost Performance Index (CPI)

How CPI Works

The Cost Performance Index is an earned value management metric that measures cost efficiency by comparing the budgeted cost of work completed to the actual cost of that work. The formula is CPI = EV / AC, where EV (Earned Value) is the budgeted cost of work actually completed and AC (Actual Cost) is the money actually spent to complete that work.

A CPI of 1.0 means the project is exactly on budget: every dollar spent produced exactly one dollar of planned value. A CPI greater than 1.0 means the project is under budget: the team is producing more value per dollar than planned. A CPI less than 1.0 means the project is over budget: each dollar spent is producing less value than planned.

CPI is expressed as a ratio, not a percentage. A CPI of 0.90 means the project gets 90 cents of value for every dollar spent. A CPI of 1.15 means the project gets $1.15 of value per dollar spent. The further CPI deviates from 1.0, the more significant the cost performance issue.

Why CPI Is the Most Reliable Forecasting Metric

Research by the US Department of Defense and independent project management researchers has consistently shown that cumulative CPI stabilizes by the time a project is 15% to 20% complete. After that point, cumulative CPI almost never improves by more than 10% for the remainder of the project. This means that if a project has a CPI of 0.85 at the 20% mark, the final CPI will likely fall between 0.77 and 0.94. The project will finish over budget.

This stability makes CPI the single best predictor of final project cost. It outperforms subjective assessments, optimistic re estimates, and management promises to “get back on track.” When a project’s CPI is consistently below 1.0, the appropriate response is to re estimate the budget, reduce scope, or both.

How to Calculate CPI

CPI = Earned Value / Actual Cost. To calculate it, you need two data points. Earned Value: take the total budget for the work completed to date and sum it. If the project has 10 work packages and 4 are complete, the EV is the total budgeted cost of those 4 work packages. Actual Cost: sum all costs incurred for the work completed, including labor, materials, overhead, and any other direct costs.

Example: A project has a total budget (BAC) of $500,000. At the Month 4 review, the team has completed work that was budgeted at $200,000 (EV = $200,000) but has spent $220,000 to do it (AC = $220,000). CPI = $200,000 / $220,000 = 0.91. The project is over budget, getting 91 cents of value per dollar spent.

CPI feeds directly into forecasting formulas. The simplest Estimate at Completion (EAC) formula is BAC / CPI. In the example above: $500,000 / 0.91 = $549,450. If current cost efficiency continues, the project will cost approximately $549,450 instead of the planned $500,000.

When to Track CPI

Track CPI on any project using earned value management. Report it at every performance review period (monthly for most projects, biweekly for fast moving or high risk projects). Plot cumulative CPI over time to identify trends. A declining CPI requires immediate investigation and corrective action. A stable CPI below 1.0 requires budget re estimation or scope reduction.

CPI is especially critical on fixed price contracts where the provider absorbs cost overruns. A CPI below 1.0 on a fixed price contract means the provider is losing money on every unit of work. Early detection allows for scope negotiations or efficiency improvements before losses become unrecoverable.

When CPI Is Less Useful

CPI requires accurate cost data, which is not always available in real time. Organizations with delayed cost reporting (costs posted weeks after they are incurred) will have lagging CPI values that do not reflect current performance. In these environments, use CPI for trend analysis rather than real time decision making.

Very early in a project (before 15% completion), CPI can be volatile because small cost variations produce large swings in the ratio. Do not overreact to CPI fluctuations in the first few reporting periods. Wait for the metric to stabilize before using it for forecasting.