EVM in 60 Seconds
Planned Value (PV), Earned Value (EV), Actual Cost (AC). From those three numbers you get everything else:
- SPI = EV / PV — are you ahead or behind schedule?
- CPI = EV / AC — are you over or under budget?
- EAC (Estimate at Completion) — what’s this project actually going to cost?
- ETC (Estimate to Complete) — how much more money do we need?
- VAC (Variance at Completion) — how far off from the original budget?
- TCPI (To-Complete Performance Index) — what efficiency do we need going forward to hit budget?
I’m not going to belabor the theory. If you’re reading this, you probably know the acronyms. The problem is never understanding EVM concepts — it’s getting P6 to calculate them correctly. That’s where most implementations fall apart.
The P6 Setup That Makes or Breaks Your EVM
Navigate to Project Settings → Calculations (or Project Details → Calculations in older versions). This single screen determines whether your earned value numbers reflect reality or fantasy.
% Complete Type
This is where most implementations go wrong. P6 offers three options per activity:
Duration % Complete (the default) — measures time elapsed relative to remaining duration. A 10-day activity on day 5 shows 50% complete. Sounds reasonable until you realize it has nothing to do with actual work performed. A crew could sit idle for 5 days and P6 would still report 50%. Almost never what you want for EVM.
Physical % Complete — manually entered by the scheduler or field team. This reflects real progress against deliverables. An activity to install 100 valves with 60 installed is 60% physically complete. This is the gold standard for EVM, but it requires discipline. Someone has to update it accurately every reporting period.
Units % Complete — calculated from resource units consumed vs. budgeted. If you budgeted 200 labor hours and 120 have been charged, Units % shows 60%. Good for labor-heavy activities where hours correlate with progress. Dangerous when they don’t — burning hours doesn’t always mean producing work.
The hybrid approach is what I recommend on most projects: assign different % complete types at different WBS levels or for different activity types. Construction and fabrication activities get Physical %. Engineering and design activities that are truly effort-driven get Units %. Level of Effort activities (project management, site supervision) get Duration %. P6 supports this — you set the % complete type at the activity level, not just the project level.
Earned Value Technique Selection
P6 lets you choose how earned value is calculated for each WBS element:
- Activity % Complete — rolls up activity-level % complete. The most common and usually the right choice.
- WBS Milestones — earned value is based on weighted milestones within the WBS. Good for summary-level reporting when you don’t trust activity-level data.
- 0/100 — zero credit until the activity is complete, then full credit. Works for short activities (a few days). Disaster for anything longer — a 3-month activity at 95% gets zero earned value.
- 50/50 — 50% credit when started, 50% when complete. A reasonable middle ground for activities where measuring progress is impractical.
- Level of Effort — earned value equals planned value. No variance, ever. Use only for true support activities (PM overhead, site security).
I once audited a project where someone had set every activity to 0/100 because “it’s simpler.” The project was a 2-year pipeline construction. Thousands of activities, many running 2-3 months each. SPI showed 0.4 when the project was actually on track — almost every in-progress activity was showing zero earned value. The project director was about to cancel a subcontractor over numbers that meant nothing.
Resource Loading for EVM
EVM without resource loading is meaningless. Here’s what you need:
- Every activity needs at least one resource — labor, material, or expense. Activities without resources contribute zero to earned value calculations.
- Resources need rates — either price/unit for labor and materials, or lump sum costs for expenses. No rates means no cost, which means no PV or EV.
- The budget must be baseline-locked before EVM makes sense. Your baseline is PV. Without it, there’s nothing to measure against.
A common mistake I see on new implementations: the team loads resources after setting the baseline, then wonders why Planned Value is zero across the board. The baseline captured the schedule before resources existed. You need to re-baseline (or set an initial baseline) after resource loading is complete.
Another mistake: loading resources at the WBS level using WBS summary resources instead of at the activity level. P6 supports this, but your earned value granularity drops to the WBS level. Fine for a summary program, not acceptable for a project that needs real cost control.
The % Complete Problem
This is the single most common EVM issue I encounter in the field, and I’ve seen it on everything from $20M plant shutdowns to billion-dollar infrastructure programs.
The scenario: P6 is configured with Duration % Complete as the default (because that’s the out-of-box setting). The scheduler updates the schedule, and P6 auto-calculates Duration %. The project controls team pulls that number and plugs it into earned value reports as if it were Physical % Complete.
The result is garbage.
Duration % says “how much time has passed relative to the remaining duration.” Physical % says “how much work has actually been done.” A 6-month activity at month 3 is 50% Duration % complete even if zero deliverables have been produced. I once inherited a schedule where the EVM setup was so wrong that CPI showed 1.4 on a project that was 6 months behind and $12M over budget. The client had been getting green status reports for a year. Nobody questioned it because the numbers “came from P6” — as if the tool validates its own inputs.
How to audit this: In P6, create a filter for activities where Duration % Complete > 0 but Physical % Complete is blank or zero. If that filter returns hundreds of in-progress activities, your EVM is based on duration, not progress. Every one of those activities needs a real Physical % assessment.
Then check the project-level setting. If Earned Value % Complete is set to “Duration” at the project level and nobody has overridden it at the activity level, the entire project’s EVM is measuring time, not work.
Building EVM Reports That Matter
Raw P6 columns are a starting point, but real EVM reporting requires context. Here are the five reports I build for every project that needs earned value:
1. Project-Level SPI/CPI Trends
Monthly time series showing SPI and CPI over the project lifecycle. Not just the current values — the trend. A CPI of 0.95 means one thing if it’s been stable for 6 months, and something completely different if it dropped from 1.05 last quarter. Plot both on the same chart with a 1.0 reference line.
2. CPI by WBS
Where are the cost overruns hiding? A project-level CPI of 0.92 might mask a WBS element at 0.65 that’s dragging everything down while other areas run at 1.05. Break it out. Sort by worst CPI first.
3. SPI by Phase
Which phases are behind schedule? If engineering is at SPI 1.1 but procurement is at 0.78, you know where to focus. This also helps predict downstream impacts — late procurement means late construction, regardless of what the construction SPI currently shows.
4. EAC Comparison
P6 calculates an EAC using the formula you configure. But that’s just one estimate. Present three side by side:
- P6 calculated EAC — based on current CPI
- Management EAC — the project team’s bottom-up re-estimate
- Independent EAC — using CPI × SPI factor (EAC = BAC / (CPI × SPI))
When all three converge, you have confidence. When they diverge wildly, someone is wrong — and that conversation needs to happen.
5. TCPI Reality Check
TCPI tells you the cost efficiency you need on remaining work to hit your budget (or your EAC). If TCPI > 1.0, you need to be more efficient than you have been. If TCPI > 1.2, be honest: the project isn’t recovering without a scope change, a schedule extension, or additional funding. I’ve never seen a project sustain a CPI improvement of more than 0.15 through “trying harder.”
What Executives Actually Want to See
Stop showing them 20 metrics. I’ve sat through hundreds of executive reviews and the questions are always the same:
- Are we on budget?
- Are we on schedule?
- What’s the forecast at completion?
- What are the top risks?
That’s it. Four questions. Your EVM report for leadership should answer those four questions on a single page.
Traffic light summary: Green / Amber / Red for cost, schedule, and overall status. One line of narrative for each. “Cost performance has declined 3% this period due to unplanned rework on Unit 2 foundations. Corrective action underway — revised estimate reflected in next month’s EAC.”
Trend lines over tables. A chart showing CPI trending from 0.98 to 0.95 to 0.93 to 0.91 over four months is more alarming — and more actionable — than a table of numbers. Executives process visuals faster than spreadsheets.
Reserve the detail for the appendix. The full WBS breakdown, activity-level variances, resource loading curves — put them in the back. They’re there if someone asks. But the first page should tell the story.
When EVM Adds Value vs. When It’s Overhead
Not every project needs earned value. Here’s my honest assessment after 15 years:
EVM adds real value when:
- The project exceeds $50M or has complex, interdependent scopes
- You’re on a cost-plus or cost-reimbursable contract where the client has a right to see efficiency metrics
- Government or defense contracts where ANSI/EIA-748 compliance is required
- Multi-year programs where early trend detection prevents late-stage disasters
EVM is usually overhead when:
- Projects under $5M with straightforward scope — track milestones and budget burn rate instead
- Time-and-materials engagements where the scope is fluid
- The organization lacks the discipline to update Physical % accurately every period
- Nobody actually reads the reports
That last point is critical. The key to useful EVM is that the earned value must reflect real work. If the % complete is fiction — estimated by a scheduler who hasn’t been to the field in weeks, or auto-calculated by duration — then your SPI, CPI, EAC, and every other metric built on top of it is fiction. Well-formatted fiction, but fiction.
Get the inputs right. The math takes care of itself.