Which project KPIs should a dashboard show? A table with formulas

What makes a good project KPI

A key performance indicator is a measure someone acts on. If a number can turn red and nobody would do anything differently, it is a statistic, not a KPI. Good project KPIs share four traits:

• Defined formula: everyone calculates it the same way.
• Reliable source: it comes from the schedule, cost data or registers, not from someone's impression.
• Threshold: green, amber and red bands agreed in advance.
• Owner: a person who explains it and acts when it moves.

Mix lagging indicators (what already happened, such as cost variance) with leading indicators (what is likely to happen, such as falling float or rising risk exposure). Leading indicators give you time to act.

Match the KPIs to the audience. A sponsor needs the four answers and the forecast: will it finish on time and on budget, and if not, by how much? A project manager needs the detail behind them: which activities are slipping, which cost accounts are overrunning, which risks are growing. A PMO needs the same small set for every project, calculated the same way, so a CPI of 0.92 means the same thing on every line of the portfolio report.

The project KPI table

KPI	Formula	How to read it
Cost performance index (CPI)	CPI = EV / AC	Below 1.0: each dollar spent earns less than a dollar of planned work
Schedule performance index (SPI)	SPI = EV / PV	Below 1.0: less work done than planned by now
Time-based SPI, SPI(t)	SPI(t) = ES / AT	Like SPI but stays meaningful late in the project
Estimate at completion (EAC)	EAC = BAC / CPI (one common form)	Forecast final cost if current efficiency continues
Variance at completion (VAC)	VAC = BAC − EAC	Negative: forecast overrun
To-complete performance index (TCPI)	TCPI = (BAC − EV) / (BAC − AC)	Efficiency needed on remaining work to finish on budget
Forecast finish variance	Forecast finish − baseline finish (working days)	Positive: late
Critical path float	Total float on the critical path	Negative: the plan cannot meet its deadline
Milestone hit rate	Milestones met on time / milestones due	Trend matters more than one period
Risk exposure	Sum of EMV of open threats (EMV = probability × impact)	Compare with remaining contingency
Open change requests	Count and total value awaiting decision	A growing backlog means decisions are stalling
Overdue issues and tasks	Count past due date	Shows whether the team keeps up

For the full set of cost formulas, see the earned value management guide; for SPI(t), the earned schedule guide.

Worked example: one month's numbers

A project has a budget at completion (BAC) of $500,000. At the end of month four: planned value (PV) is $200,000, earned value (EV) is $180,000 and actual cost (AC) is $200,000.

• CPI = 180,000 / 200,000 = 0.90. The project earns 90 cents of planned work for every dollar spent.
• SPI = 180,000 / 200,000 = 0.90. It has done 90% of the work planned by now.
• EAC = 500,000 / 0.90 = $555,556 (rounded).
• VAC = 500,000 − 555,556 = −$55,556, a forecast overrun.
• TCPI = (500,000 − 180,000) / (500,000 − 200,000) = 320,000 / 300,000 = 1.07 (rounded). To finish on the original budget, the remaining work must be done about 7% more efficiently than planned, while the team is currently running 10% less efficiently.

Read together, these say more than any one number: the project is behind and over cost, and recovering to budget would require a jump in efficiency that is rarely realistic. The useful dashboard conversation is therefore about re-forecasting or reducing scope, not about "trying harder".

How to lay out the dashboard

1. Top row: the four answers. Schedule (SPI or forecast finish variance), cost (CPI and EAC), risk (exposure against contingency) and scope (open changes).
2. Second row: trends. CPI and SPI over the last six periods, and an S-curve of planned value, earned value and actual cost. A single 0.95 means little; a slide from 1.02 to 0.95 over three months means a lot.
3. Third row: exceptions. Late milestones, top five risks, overdue issues. Lists of what needs a decision.
4. Portfolio view. For a PMO, one row per project with the same KPIs and the same thresholds, so projects can be compared fairly.

Agree thresholds once and publish them. A common pattern is green at 0.95 or above, amber from 0.90 to 0.95, red below 0.90 for CPI and SPI, but choose bands that suit your projects and keep them stable.

Common KPI mistakes

• Percent complete as the only KPI. "60% done" without a link to planned value says nothing about whether 60% is good.
• Spend against budget as a cost KPI. Spending less than planned can simply mean work is late. CPI compares cost with work done, which is what matters.
• Too many indicators. Twenty tiles hide the two that matter. Start with the table above and remove what nobody uses.
• Changing formulas mid-project. Trends break and trust goes with them.
• Green dashboards over a broken schedule. If the schedule has missing logic, SPI and forecast dates are unreliable. Check schedule quality, for example with the DCMA 14-point checks, before trusting the KPIs.

How to do this in Critova

Critova calculates CPI, SPI, SPI(t), the EAC family and TCPI from your schedule and cost entries, and shows each figure with its formula so nobody has to guess how it was derived. Eleven dashboards, including executive, portfolio, cost, schedule and risk views, place the KPIs above side by side, and the risk register adds up expected monetary value for exposure. Each project also has a printable status report. See the portfolio dashboards, or try the numbers from the example in the free earned value calculator.