Earned value management (EVM) is a way of measuring project performance by comparing three numbers: what you planned to have spent by now, what the work you have actually finished is worth, and what you actually spent to finish it. Those three values, planned value, earned value, and actual cost, are enough to tell you whether a project is over budget, behind schedule, or both, and to forecast where it will land. It is the antidote to the status update that says "we are 60% done and we have spent 80% of the budget" without ever admitting that those two facts, together, are bad news.
Key takeaways
- EVM runs on three inputs: planned value (PV), earned value (EV), and actual cost (AC). Everything else is derived from them.
- The two numbers you will use most are the cost performance index (CPI = EV / AC) and the schedule performance index (SPI = EV / PV). Above 1.0 is good, below 1.0 is bad, and 0.80 means you are getting 80 cents of work for every dollar you spend.
- EVM forecasts, it does not just report. Estimate at completion (EAC = BAC / CPI) projects the final cost from performance so far, which is what makes it useful to a portfolio board.
- The technique has a real limit: SPI always drifts back to 1.0 as a project finishes, so it stops being a useful schedule signal in the final stretch. Use milestone dates there instead.
What is earned value management?
Earned value management is a project performance technique that integrates scope, schedule, and cost into a single set of measures. Instead of tracking spending and progress separately, EVM asks what the completed work was actually worth against its budget, then compares that figure to both the plan and the money spent. It is codified in the ANSI/EIA-748 standard and built into PMI's guidance.
The reason it matters to a PMO is that it is comparable. A percent-complete figure means whatever the project manager wants it to mean. A CPI of 0.82 means the same thing on every project in the portfolio, which is what lets you rank, roll up, and challenge.
The three numbers EVM runs on
Every EVM measure is derived from these three. Get them right and the rest is arithmetic.
| Input | What it answers | How you get it |
|---|---|---|
| Planned value (PV) | What should the work done by now have cost, according to the baseline? | Read it off the time-phased budget at the reporting date. |
| Earned value (EV) | What is the work we actually completed worth, in budget terms? | BAC x percent of the work genuinely complete. |
| Actual cost (AC) | What did we really spend to complete that work? | Timesheets, invoices, and expenses posted to the project code. |
BAC is the budget at completion, the total approved budget for the work. The weak link is almost always AC: earned value only tells the truth if every hour and every supplier invoice has actually landed against the project. If contractor bills sit in someone's inbox for six weeks, your actual cost is fiction and your CPI is flattering. Getting every expense and invoice captured against the right project code is unglamorous plumbing, and it is the thing that decides whether EVM works at all.
Earned value management formulas
This is the full working set. BAC is the budget at completion.
| Measure | Formula | How to read it |
|---|---|---|
| Cost variance (CV) | CV = EV - AC | Negative means over budget. |
| Schedule variance (SV) | SV = EV - PV | Negative means behind schedule. |
| Cost performance index (CPI) | CPI = EV / AC | Value earned per dollar spent. Below 1.0 is over budget. |
| Schedule performance index (SPI) | SPI = EV / PV | Work done against work planned. Below 1.0 is behind. |
| Estimate at completion (typical) | EAC = BAC / CPI | Use when current performance is expected to continue. |
| Estimate at completion (atypical) | EAC = AC + (BAC - EV) | Use when the overrun was a one off and the rest will run to budget. |
| Estimate at completion (cost and schedule) | EAC = AC + [(BAC - EV) / (CPI x SPI)] | The pessimistic view. Use when schedule pressure is driving cost. |
| Estimate to complete (ETC) | ETC = EAC - AC | What finishing the job will still cost from here. |
| Variance at completion (VAC) | VAC = BAC - EAC | The forecast overrun or underrun. Negative means an overrun. |
| To complete performance index (TCPI) | TCPI = (BAC - EV) / (BAC - AC) | The efficiency you now need to hit the original budget. |
A worked earned value example
The numbers below are illustrative, chosen to divide cleanly. A project has a budget at completion of $500,000 over ten months. At the end of month four, the baseline said $200,000 of work should be done. The team has actually completed work worth $160,000 of the budget, and finance shows $200,000 spent.
So PV = $200,000, EV = $160,000, AC = $200,000.
| Measure | Calculation | Result | Verdict |
|---|---|---|---|
| Cost variance | 160,000 - 200,000 | -$40,000 | $40,000 over budget already. |
| Schedule variance | 160,000 - 200,000 | -$40,000 | A month's worth of work short. |
| CPI | 160,000 / 200,000 | 0.80 | 80 cents of work per dollar spent. |
| SPI | 160,000 / 200,000 | 0.80 | Running at 80% of planned pace. |
| EAC (typical) | 500,000 / 0.80 | $625,000 | Forecast final cost if nothing changes. |
| VAC | 500,000 - 625,000 | -$125,000 | A forecast overrun of $125,000. |
| ETC | 625,000 - 200,000 | $425,000 | Still to spend. |
| TCPI | (500,000 - 160,000) / (500,000 - 200,000) | 1.13 | Needs 1.13 efficiency from here to hold budget. |
That last line is the one that ends the argument in a portfolio review. A team running at 0.80 efficiency is being asked to suddenly run at 1.13 for the rest of the project. It will not happen. The honest options are more money, less scope, or a later date, and the board can now pick one on month four instead of month nine.
What is a good CPI and SPI?
A CPI or SPI of 1.0 means performance is exactly on plan. Above 1.0 is favorable, below 1.0 is unfavorable. In practice most PMOs treat 0.95 to 1.05 as a normal working band, flag anything below 0.90, and escalate below 0.85. A CPI of 1.4 is not a triumph, it usually means the baseline was padded.
| Index value | Cost (CPI) | Schedule (SPI) |
|---|---|---|
| Greater than 1.0 | Under budget | Ahead of schedule |
| Equal to 1.0 | On budget | On schedule |
| Less than 1.0 | Over budget | Behind schedule |
What is the difference between CPI and SPI?
CPI compares the value of completed work to the money spent on it (EV / AC), so it measures cost efficiency. SPI compares the value of completed work to the value that should have been completed by now (EV / PV), so it measures progress against plan. One is about money burned, the other about work delivered.
They move independently. A project can be ahead of schedule and over budget at the same time, which usually means it bought speed with overtime or extra contractors. Reading both together is the point.
How do you calculate earned value?
Earned value equals the budget at completion multiplied by the percentage of work actually complete. If a $500,000 project is genuinely 32% done, EV is $160,000. The hard part is not the multiplication, it is establishing the percentage honestly, which is why serious EVM programs avoid subjective estimates.
The usual methods for setting that percentage are worth knowing:
- 0/100: a work package earns nothing until it is finished, then earns everything. Conservative and very hard to game. Best for short packages.
- 50/50: earns half its budget when it starts, the rest when it completes.
- Milestone weighting: value is earned at defined, verifiable milestones. The best fit for portfolio level reporting.
- Percent complete: the project manager estimates. The most common and the easiest to fudge, which is exactly why the PMO should sample and challenge it.
How a PMO uses earned value across a portfolio
At portfolio level EVM stops being a project control technique and becomes a comparison tool. Because CPI and SPI are dimensionless ratios, they can sit side by side on a dashboard for a $50,000 project and a $5 million program without any scaling.
Three things a PMO does with them:
- Roll up. Portfolio CPI is the sum of all EV divided by the sum of all AC. It tells you whether the whole portfolio is converting money into delivered value, and it is a much harder number to spin than a page of green status circles.
- Rank and triage. Sort the portfolio by CPI, and the bottom five projects are your agenda for the next portfolio review meeting. No debate about whose project feels troubled.
- Forecast the funding gap. Add up VAC across the portfolio and you have the forecast overrun for the year, early enough to reprioritize instead of asking for emergency money in Q4.
Keep the roll up honest by only aggregating projects on comparable baselines. Mixing a fixed-price vendor build with an internal agile team produces a portfolio CPI that means very little. Where EVM fits with your other measures is covered in project portfolio management KPIs and metrics, and where the numbers land is the project portfolio dashboard.
Does earned value management work in agile?
Partly. EVM needs a stable scope baseline to compare against, and an agile team deliberately keeps scope flexible, so classic EVM fights the method. Teams that make it work substitute story points or completed features for dollars of scope, and treat the release or increment as the baseline rather than the whole project.
If your portfolio is mostly agile, the honest answer is that a flow measure such as throughput or cycle time will tell you more than an SPI will. Blending both is common in a mixed portfolio, and the trap is pretending one number covers both delivery models.
What are the limits of earned value management?
Three limits are worth stating plainly, because most guides skip them.
SPI becomes useless at the end. As a project finishes, EV climbs toward BAC and PV climbs toward BAC too, so SPI is dragged back to 1.0 no matter how late the project is. A project delivered six months late shows an SPI of exactly 1.0 on its final report. Once a project is past roughly 80% complete, use forecast milestone dates, not SPI, to talk about schedule.
It is only as good as the actuals. Late invoices and unposted timesheets make CPI look great right up until the accruals land.
It says nothing about value. EVM measures whether you built the thing efficiently, never whether the thing was worth building. A project can finish at a CPI of 1.05 and deliver no benefit at all. That question belongs to benefits realization management and to the project business case, and no cost index will ever answer it for you.
Is earned value management still used?
Yes, and it is mandatory in places. US federal contracting, defense, aerospace, construction, and large capital programs generally require an EVM system compliant with ANSI/EIA-748. Outside those sectors adoption is lighter, and many commercial PMOs use a simplified version: CPI, SPI, and EAC only, reported monthly, with no formal control-account structure.
That simplified version is the sensible default for most portfolios. You get the forecasting power without the overhead of a full compliant system.
Where earned value fits in the reporting stack
EVM is one input to portfolio reporting, not the whole of it. The rhythm of reporting, who gets what, and how often, is covered in PMO reporting, while the color-coded health rating that sits next to your CPI on most status reports is explained in RAG status. Cost and schedule indices are the objective half of a status report. The subjective half is a judgment call, and the two disagree more often than anyone likes to admit. When they do, the number is usually right.
For the wider discipline these measures serve, start at project portfolio management, and for how the money gets committed in the first place, see project budgets and portfolio spend.