What Is EVM (Earned Value Management)? How to Read the Six Metrics and Where to Use Them
Published:
Last Updated:
Category:
Published:
Last Updated:
Category:
Authors: Shusaku Yosa
Progress is at 70 percent, budget consumption at 60. Set those two numbers side by side and you still cannot say whether the project is healthy or in trouble. Only by putting progress and cost on the same scale can you connect both into a single judgment. That is what EVM does. This article covers the meaning of the six metrics, how to read them, and where they are worth using.
EVM (Earned Value Management) handles a project's progress and cost in the same unit and measures the gap against plan. Money is the usual unit, but person-days or effort work just as well.
The idea underneath is that finished work carries a value equal to what it was worth. If half of a million-yen piece of work is done, you count it as having earned half a million yen of value. EVM sets plan and actual side by side against that earned value.
A completion percentage tells you only how far the work has advanced. It says nothing about whether you are getting there with more hands than planned, or fewer.
Look only at budget consumption and you cannot tell whether the work finished justifies the money spent either. Having used 60 percent of the budget can be evidence of health or a warning sign.
As long as you view the two separately, the judgment stays a matter of feel. Putting them in the same unit is what makes subtraction and division possible.
First, the three measured values. Everything else rests on these.
The value, in money, of the work that should be finished as of the status date according to plan.
On a 10 million yen project where the plan has 40 percent complete by today, PV is 4 million yen. It is fixed when the plan is made, regardless of what is actually happening.
The value at the centre of EVM. It is the work actually completed, converted at the planned rate.
In the same example, if only 30 percent is actually done, EV is 3 million yen. The important point is that you convert at the planned rate, not at what was actually spent. However many people did the work, the value of what is finished does not change.
What has actually been spent up to the status date. This includes not only external costs but your own team's effort converted into money.
Treat only external payments as AC and your internal labour becomes invisible, producing numbers that look better than reality. This is a point to watch.
Derived from the base three by subtraction. They show the size of the gap in money.
SV = EV − PV
The plan says 4 million yen of work should be done; the actual is 3 million. SV is minus 1 million yen, indicating you are behind. Positive means ahead of plan.
One caution on reading it. SV comes out in money, but what it means is schedule delay. Read it as "a million yen worth of work is behind," not "a million yen has been lost."
CV = EV − AC
If getting 3 million yen of work done has actually cost 3.5 million, CV is minus 500,000 yen. This one does mean cost overrun, directly.
What budget consumption cannot show becomes visible here, because you are comparing what was spent directly against what was earned.
These express the same gaps as ratios rather than amounts. They are used as a pair.
SPI = EV ÷ PV (schedule performance index)
CPI = EV ÷ AC (cost performance index)
For both, 1.0 means on plan. Below 1.0 signals a problem; above 1.0 means better than planned. In the example above, SPI is 0.75 and CPI is roughly 0.86.
The advantage of ratios over amounts is that you can compare projects of different sizes. A million-yen overrun on a 100 million yen engagement and the same overrun on a 5 million yen one mean different things, but CPI puts them on the same scale.
Setting thresholds in advance — review the situation below 0.9, revisit the plan below 0.8, for instance — fixes what you do when you see the number.
The value of EVM lies less in the individual metrics than in the combination. SPI and CPI together give four states.
The second and third are where judgment most often goes wrong in practice. Look at either number alone and both appear to be in good shape. Only side by side can you tell short-handed from over-resourced.
A useful method, but the numbers stop meaning anything once its premises break down.
The largest premise. Without deciding what each piece of work is worth, neither PV nor EV can be calculated. Identifying the work and assigning effort and a rate to each item is the groundwork for adopting EVM.
EV is calculated by counting finished work. If people are reporting by feel that something is "more or less done," EV becomes a number based on feel.
One remedy is to restrict the steps. Use two states, 0 and 100 percent; or 50 percent on starting and 100 on completion. Not allowing intermediate figures removes the subjectivity.
AC has to include your own team's effort. Without time records, neither CV nor CPI can be produced. Adopting EVM makes effort tracking a prerequisite.
You do not have to adopt all of it at once. Starting with SPI alone — just EV ÷ PV — is enough. Convert in person-days rather than money and you can skip setting rates as well.
Even that much lets you respond to "progress is at 70 percent" with "the plan says 80 by the end of this week, so we are behind."
Budget-versus-actual compares two things, budget and spend. EVM adds earned value and compares three. That one addition lets you judge whether the money spent is appropriate in relation to progress.
Yes, where the deliverables are well defined. Website builds, video production, exhibition stands — anything whose tasks can be enumerated. It does not suit continuous work such as running paid media. Work with no ending has no definition of completion to place.
Not necessarily. You may simply be doing the easy work first, with the hard parts still ahead. Also, toward the end of a project SPI converges on 1.0 by definition, so late-stage figures are poor at surfacing delay.
Weekly is the guide. Monthly finds problems too late; daily swings too much to judge. Aligning the calculation with your regular meeting is the practical approach.
The formulas are simple enough to build in a spreadsheet. The problem is that the inputs are scattered: the plan in a schedule, progress in a task tracker, actual costs in timesheets and finance's files. The effort of gathering and consolidating these every week is what causes the practice to lapse.
Xtrategy manages campaign schedules alongside budget and KPIs on a single screen, which removes the re-keying from the work of setting plan against actual.
You do not need everything in place before starting EVM. Begin with EV ÷ PV — setting "the amount that should be done by the end of this week" beside "the amount actually done." Delays that were invisible while you were staring at a completion percentage start showing up as numbers.
What a PMO does and how it differs from a PM, organised around who holds responsibility. Covers the supportive, controll...
How PMP, Japan's national Project Manager Examination, and P2M differ, organised by the nature of each scheme. Covers wh...
A five-step method for mapping a business process, plus the three things that create key-person dependency, how to inter...
