Earned Value Management: Have We Got What We Paid For?

“We have spent half the budget” tells a steering committee almost nothing. Half the budget and half the work is a project on plan. Half the budget and a third of the work is a project in trouble that will not admit it for another two months. Earned Value Management (EVM) exists to force that distinction into the open, and to put a number on it.

EVM has a reputation problem outside the defence and government world where it grew up: too bureaucratic, too much overhead for a normal portfolio. Parts of that reputation are deserved, and we will deal with them honestly. But the core of the method is three numbers and some arithmetic. A PMO can run it lightly, the evidence says it pays for itself when the culture around it is right, and the rise of AI tooling is quietly dismantling the overhead objection. This article covers what EVM is, where it came from, how to size it to your portfolio, and where we think it goes next.

Understand how it works, test your numbers with our calculator, and download our free EVM Dashboard Template.

Origins and history

EVM has no single inventor, and any article that names one should be read with suspicion. The method grew out of industrial work-measurement practice and was codified by the US Department of Defense in the late 1960s as the Cost/Schedule Control Systems Criteria. An industry rewrite in the 1990s streamlined the criteria, and the result became the EIA-748 standard in 1998.

Custody of that standard now sits with SAE International. The current edition, SAE EIA-748E, was published in February 2026 and trims the guidelines from 32 to 27; its companion Intent Guide from the NDIA maps the old set to the new. If you see “ANSI-748” in older material, it is the same lineage under an obsolete label. For practitioners, the more readable document is PMI’s The Standard for Earned Value Management (ANSI/PMI 19-006-2019), which brought earned schedule and agile delivery into scope.

Government remains the method’s heartland. A 2012 study in Project Management Journal traced EVM’s history and practice at NASA, showing how deeply the method became embedded in US public-programme oversight, and why: it is the only common measure that integrates scope, schedule and cost in a single view. That study is now well over a decade old, which is itself evidence. EVM is one of the longest-serving performance methods in project management, and it has survived every “this will replace it” cycle so far.

The three numbers

Everything in EVM is arithmetic on top of three values, each measured at the same point in time:

• Planned Value (PV): what the plan says should have been completed by now, expressed in money.
• Earned Value (EV): what has actually been completed by now, valued at its budgeted cost, whatever it really cost.
• Actual Cost (AC): what has actually been spent getting there.

The comparisons between them answer the two questions every sponsor asks:

• Schedule Variance (SV = EV – PV): negative means behind schedule, measured in money.
• Cost Variance (CV = EV – AC): negative means the work delivered has cost more than it should.
• Schedule Performance Index (SPI = EV / PV) and Cost Performance Index (CPI = EV / AC): ratios that travel well across projects. A CPI of 0.83 means you are getting 83p of work for every pound spent, which is a sentence a sponsor understands immediately.
• Estimate at Completion (EAC), in three main flavours depending on what you believe about the future: EAC = BAC / CPI if current cost performance will continue; EAC = AC + (BAC – EV) if the overrun so far was a one-off; EAC = AC + (BAC – EV) / (CPI x SPI) if schedule pressure will drive cost too. BAC is the Budget at Completion, the total baseline budget.

One refinement worth knowing: classic schedule variance drifts back towards zero as any project nears its end, however late it is, because eventually all the planned work gets done. Earned schedule, folded into the 2019 PMI standard, converts the schedule measures into time units, so you can say “three weeks behind” rather than “£40k of schedule variance”. That is how normal humans think about lateness, and it keeps the measure honest in the endgame.

How to use it

1. Baseline properly. A decomposed scope (a work breakdown structure), time-phased into a budget. If the plan is a bar chart drawn last night, EVM has nothing to measure against.
2. Credit progress by rule, never by feel. Short tasks score 0/100 (nothing until done) or 50/50 (half on start, half on completion). Rule-based crediting is what kills the notorious “90% complete” report, because nobody gets to self-assess momentum.
3. Run the numbers on a monthly drumbeat. CPI and SPI per project, trended over time. The trend matters more than the snapshot; a CPI sliding from 1.02 to 0.91 over three months is the early warning working.
4. Use the EAC variants as a conversation, not a verdict. “If nothing changes we land at £X; if the overrun was a one-off, £Y” gives a sponsor a defensible range and makes the assumptions explicit.
5. Protect the inputs. EVM is only as good as the baseline and the actuals feed. Weak change control or cost data that arrives six weeks late will produce precise-looking nonsense, and precise-looking nonsense is more dangerous than no data at all.

We have built the arithmetic into a Free Excel Calculator. Enter BAC, PV, EV and AC and it returns the variances, both indices and all three EAC variants, with a worked example to check yourself against. Or try it right here, without leaving the page:

When the PMO should use this

Proportionality is the whole game, and it is where most EVM adoptions go wrong. Full SAE EIA-748 compliance, with all 27 guidelines, belongs on large, complex or high-risk programmes, usually where a government or defence contract demands it. Nobody should be running a compliance-grade EVMS on a £200k website build. The PMO’s job is to size the method: which projects justify full EVM, which get EVM-lite, and which get none. EVM-lite, meaning a baselined plan, rule-based percent complete and monthly CPI/SPI, is within reach of any portfolio that plans at all, and it gives the PMO something a spend-versus-budget dashboard can never give: it exposes the project that is “on budget” only because it is quietly doing less work. Within a PRINCE2 environment, this slots naturally into the Progress theme.

The evidence also says the spreadsheet is the smaller half of the job. A 2024 study in Project Management Journal, aptly titled “An Effective Earned Value Management System (EVMS) is a Team Sport”, assessed 27 environment factors across 35 completed projects and programmes worth over US$21.8 billion, and found that a project with a positive EVMS environment, meaning culture, team competence and practices around the numbers, could save up to 25% in cost against baseline. Teams that own their numbers produce numbers worth owning. A PMO that mandates EVM without building that environment gets the reporting burden and none of the return.

Could AI make EVM mainstream?

Here is the question we would put to every PMO wrestling with the overhead objection: if the hard part of EVM is compiling it, what happens when compiling stops being hard? The data EVM needs already exists in most organisations. The baseline sits in the scheduler, actuals sit in finance, progress evidence sits in timesheets, ticket systems and delivery tools. Assembling an earned-value view from those sources is exactly the kind of multi-source, rule-following work that AI tooling is becoming good at. The project manager’s role shifts from compiling the numbers to interpreting them, and the sponsor gets the plain-English version without waiting for month-end.

The same shift could rescue good ideas that never escaped the research literature. A 2021 paper in the Journal of Intelligent & Fuzzy Systems proposed directed EVM using ordered fuzzy numbers, which lets progress estimates carry their uncertainty and their trend instead of pretending to false precision. It is clever, and almost no working project manager will ever compute it by hand. That is precisely the class of technique AI could make mainstream: uncertainty-aware forecasts on tap, with the mathematics hidden and the meaning surfaced. EVM’s future audience is wider than its past, if the tooling meets practitioners where they are: helping project managers compile, and helping sponsors understand.

Is EVM dead? Criticism and limitations

The question gets asked every few years, usually with agile in the next sentence, so let us answer it properly. The genuine weaknesses first. EVM measures cost and schedule performance against a plan; it says nothing about quality or benefit, so a project can hold CPI 1.0 while building the wrong thing, which is why we pair it with an outcome lens such as OKRs. Classic schedule metrics mislead near the finish line unless you use earned schedule. The method inherits every weakness of its baseline, so undisciplined change control poisons it silently. And gold-plated implementations on projects that never needed them have burned goodwill that lighter implementations now have to win back. If your concern is schedule risk specifically, the Critical Chain Method attacks that problem from a different and complementary angle.

Our position: EVM is not dead, it is routinely oversized. Sized to the project, fed honest data and surrounded by a team that owns its numbers, it remains the only widely standardised way to answer “have we got what we paid for?” while there is still time to act on the answer.

How HotPMO can help

We help organisations build PMOs that measure what matters and act on it early, embedding alongside your team rather than advising from a distance. If your reporting says “on budget” more often than it says anything useful, or an EVM adoption is producing numbers nobody trusts, the honest first step is a capability question rather than a tooling question. Our PMO Assessment (the SPACE Audit) looks at exactly that, baselines and behaviours included, and as an ISO 9001 certified consultancy we hold our own measurement discipline to the same standard.

This article is part of our management models library. For more practical models like this one, straight to your inbox roughly once a month, join the HotPMO newsletter below.

Further reading

• SAE International (2026). EIA-748E: Earned Value Management Systems. SAE International. (Current edition; supersedes EIA-748D-2019.)
• NDIA IPMD (2026). Earned Value Management Systems EIA-748-E Intent Guide, revision of 14 May 2026. National Defense Industrial Association.
• Project Management Institute (2019). The Standard for Earned Value Management (ANSI/PMI 19-006-2019). PMI. ISBN 9781628256383.
• Fleming, Q. W. and Koppelman, J. M. (2010). Earned Value Project Management, 4th edition. PMI. ISBN 9781935589082.
• Aramali, V., Gibson, G. E. Jr., El Asmar, M. and Sanboskani, H. (2024). “An Effective Earned Value Management System (EVMS) is a Team Sport”. Project Management Journal. https://doi.org/10.1177/87569728231226226
• Kwak, Y. H. and Anbari, F. T. (2012). “History, Practices, and Future of Earned Value Management in Government: Perspectives From NASA”. Project Management Journal. https://doi.org/10.1002/pmj.20272
• Mortaji, S. T. H., Noori, S. and Bagherpour, M. (2021). “Directed earned value management based on ordered fuzzy numbers”. Journal of Intelligent & Fuzzy Systems, 40, 10183-10196. https://doi.org/10.3233/JIFS-201248