Earned Value Management Without the Jargon: Five Mistakes That Wreck the Numbers
Earned value management (EVM) has a reputation for being complicated, but the idea behind it is plain: instead of asking "how much have we spent?" you ask "how much of the work did we actually get for that spend?" That single shift catches trouble months before a burn-rate chart would. The trouble is that EVM only tells the truth if you feed it honest inputs, and in early 2021 — with split teams, half the staff working from kitchen tables, and materials arriving late or not at all — a lot of those inputs quietly went bad.
Strip it to three numbers. Planned value (PV) is the budgeted cost of the work you said you'd finish by today. Earned value (EV) is the budgeted cost of the work you've genuinely completed. Actual cost (AC) is what you actually paid. From those, schedule variance is EV minus PV, and cost variance is EV minus AC. Negative means behind or over. That's the whole engine. The mistakes below are about feeding it badly.
Where teams go wrong
Claiming credit for work that isn't done. The classic failure is the 90-percent-complete task that stays at 90 percent for a month. If you let people self-report progress with no rule, optimism inflates EV and your numbers look healthy right up until they don't. Use a defined rule — 0/50/100, or weighted milestones — and apply it the same way every period.
Confusing money spent with progress made. Paying an invoice is not the same as earning value. In 2020 and 2021 many teams pre-paid suppliers to lock in scarce materials. That cash left the account, but the work it represented hadn't been performed. Booking it as earned value made projects look on track while the real schedule slipped.
A baseline nobody maintains. EVM compares against the plan you committed to. If scope grew through three change requests but the baseline was never re-set, every variance is measuring against a fiction. Re-baseline through change control — deliberately, with a record — not by quietly editing the spreadsheet.
Reading cost and schedule as the same thing. A project can be under budget and badly behind, or over budget and ahead. Cost variance and schedule variance answer different questions. Treat them separately, and remember schedule variance fades to zero at the end even on a late project, because all the planned value eventually gets earned.
Reporting numbers without a story. A dashboard that says CPI 0.92 helps no one who can't act on it. Pair every metric with the reason — a delayed shipment, a vacant role, rework on a flawed drawing — and what you're doing about it. EVM is an early-warning system, not a scoreboard.
Keeping it honest
Two ratios do most of the work. The cost performance index (CPI = EV / AC) tells you how much value you get per dollar; the schedule performance index (SPI = EV / PV) tells you how much of the planned pace you're holding. Below 1.0 on either is a flag. They also let you forecast: if your CPI is running at 0.9, finishing the rest of the work at today's efficiency means the estimate at completion is roughly your budget divided by 0.9 — a sobering, defensible number to put in front of a sponsor.
Above all, keep the cadence steady and the rules fixed. EVM is only useful when this period is measured the same way as the last one. Change the progress rule mid-project and you've broken the comparison that gives the method its value.
If your project numbers feel disconnected from what's actually happening on the ground, XNM's program & project delivery advisory can help you build controls that tell the truth.