Cost Variance Percentage (CV%) normalises cost variance so you can compare projects of wildly different sizes. A CV of -£200,000 means something entirely different on a £5M package than on a £50M programme. CV% strips out the scale and gives you a rate. It's the metric that lets a commercial director scan ten projects in a dashboard and instantly spot which ones are bleeding.
CV% = (CV / EV) x 100
Or equivalently:
CV% = ((EV - AC) / EV) x 100
Where:
- CV = Cost Variance (EV - AC)
- EV = Earned Value (budgeted cost of work completed)
- AC = Actual Cost (what you've actually spent)
CV% is part of the earned value definitions glossary. For the full variance reference, see the cost and schedule variance page.
Why CV% Exists
Raw CV is great for one project. It tells you the pound figure. But the moment you need to compare across projects, it falls apart.
PROJECT COMPARISON – Same CV, Very Different Stories Project A: £5M fit-out Project B: £50M highway ───────────────────────── ───────────────────────── EV = £4,800,000 EV = £28,000,000 AC = £5,000,000 AC = £28,200,000 CV = -£200,000 CV = -£200,000 CV% = -£200K / £4.8M CV% = -£200K / £28M = -4.2% = -0.7% ┌──────────────────┐ ┌──────────────────┐ │ █████████████░░ │ -4.2% │ ██████████████▒ │ -0.7% └──────────────────┘ └──────────────────┘ Significant problem Rounding error Same £200K. Completely different severity.
Project A is haemorrhaging. For every £100 of value delivered, it's spending £104.20. That compounds across the remaining works and could wipe out the margin entirely. Project B has a minor blip, £200K on a £50M scheme is within normal variance. Without CV%, the dashboard shows two red flags. With CV%, it shows one red flag and one green tick.
The Relationship Between CV%, CPI, and CV
These three metrics are mathematically linked. If you know any two, you can derive the third.
| Metric | Formula | Output | Use Case |
|---|---|---|---|
| CV | EV - AC | Pounds (£) | Absolute overrun for board reporting |
| CV% | CV / EV x 100 | Percentage | Cross-project comparison, thresholds |
| CPI | EV / AC | Ratio | Forecasting EAC, trend analysis |
The mathematical link: CV% = (1 - 1/CPI) x 100
So a CPI of 0.91 gives CV% = (1 - 1/0.91) x 100 = -9.9%. They're telling you the same thing in different formats. CPI speaks to commercial managers who use it for forecasting. CV% speaks to directors who want a quick health check across the portfolio.
Worked Example: Portfolio Dashboard With CV%
Worked ExampleScenario: A Tier 1 contractor runs five NEC4 packages simultaneously. The commercial director reviews the monthly portfolio dashboard (December 2025).
| Project | Contract | EV | AC | CV | CV% | CPI | Status |
|---|---|---|---|---|---|---|---|
| M6 Junction Improvement | £52M Option C | £28.4M | £28.6M | -£200K | -0.7% | 0.993 | Green |
| A14 Bridge Replacement | £18M Option A | £9.7M | £10.8M | -£1.1M | -11.3% | 0.898 | Red |
| Rail Electrification Phase 2 | £35M Option C | £19.2M | £18.4M | +£800K | +4.2% | 1.043 | Green |
| Water Treatment Upgrade | £8M Option C | £4.8M | £5.0M | -£200K | -4.2% | 0.960 | Amber |
| Depot Refurbishment | £5M Option A | £3.1M | £2.9M | +£200K | +6.5% | 1.069 | Green |
Reading the dashboard:
Without CV%, the M6 and Water Treatment projects both show -£200K cost variance. Identical headline number. But CV% reveals the truth: the M6 is barely off budget (-0.7%) while the Water Treatment project is losing 4.2% on every pound of work. That's the difference between "no action needed" and "schedule a commercial review this week."
The A14 at -11.3% is the real emergency. That CPI of 0.898 means for every £1 spent, the team delivers 90p of value. On the remaining £8.3M of work, if efficiency doesn't improve, the EAC is £18M / 0.898 = £20.04M, a £2.04M overrun on an £18M fixed-price contract. On Option A, where you can't adjust the target, that overrun comes straight off the margin.
The action: The commercial director calls a recovery meeting for A14, requests a root cause analysis, and asks whether any of the cost overrun qualifies for compensation events. Meanwhile, the M6 and Depot projects carry on with normal monthly monitoring.
Threshold Table
I've seen various threshold schemes across different clients and frameworks. This is the one I use for UK construction projects:
| CV% Range | RAG Status | Interpretation | Action Required |
|---|---|---|---|
| > +5% | Blue | Significantly under budget | Check EV measurement, are you overclaiming progress? |
| +1% to +5% | Green | Under budget | Healthy. Monitor for sustainability. |
| 0% to -3% | Green | Minor variance | Normal range. Investigate if trending downward. |
| -3% to -7% | Amber | Moderate overrun | Root cause analysis. Review at monthly commercial meeting. |
| -7% to -12% | Red | Significant overrun | Recovery plan required. Escalate to project director. |
| < -12% | Red (critical) | Severe overrun | Immediate intervention. Probably too late for recovery without contract relief. |
These thresholds aren't gospel. A -5% CV% on a £100M programme is £5M. That demands a different response than -5% on a £2M package. Use CV% for screening, then switch to raw CV and CPI for the detailed analysis.
Common Mistakes
- Dividing by AC instead of EV. I've seen this more times than I'd like to admit. CV% = CV / EV, not CV / AC. The denominator is Earned Value because you're expressing the variance as a percentage of what the work is worth, not what you spent. Using AC as the denominator gives you a different ratio that isn't CV% and doesn't compare cleanly across projects.
- Using CV% when EV is very small. Early in a project, EV might be £200K. A CV of -£50K gives a CV% of -25%, which looks catastrophic. It's not. It's one material delivery or one subcontractor invoice. CV% is meaningless until you've got enough EV for the ratio to stabilise. I don't trust CV% until a project is at least 15-20% complete.
- Setting the same thresholds regardless of contract type. On NEC4 Option A (fixed price), a -5% CV% eats directly into your margin. On Option C (target cost), the same -5% is shared between Contractor and Client through the pain/gain mechanism. Your threshold table should reflect the contract risk profile. I use tighter thresholds for Option A and looser ones for Option C.
- Ignoring the trend. A CV% of -3% is amber. But a CV% that's gone from -1% to -2% to -3% over three months is a trend heading towards red. Plot CV% over time, not just the current period snapshot. The direction matters more than the number.
How Gather helps. Gather's AI reads your site diaries daily and maps progress against your cost-loaded programme, giving you accurate earned value data without manual spreadsheet updates. Book a demo to see it working on a live NEC4 project.
Frequently Asked Questions
What is the formula for cost variance percentage?
CV% = (CV / EV) x 100, where CV is Cost Variance (EV minus AC) and EV is Earned Value. A positive CV% means you're under budget; negative means over budget. For example, if EV = £10M and AC = £10.8M, then CV = -£800K and CV% = (-£800K / £10M) x 100 = -8.0%.
What is a good cost variance percentage?
On well-managed UK construction projects, CV% typically falls between -5% and +5%. Anything above 0% is favourable (under budget). Between 0% and -5% is normal variance. Below -5% usually requires investigation and a recovery plan. Below -10% is a serious problem. But context matters, these thresholds should be tighter on fixed-price contracts (NEC4 Option A) where overruns hit your margin directly, and can be looser on target cost contracts (Option C) where the risk is shared.
What's the difference between CV% and CPI?
They measure the same thing differently. CV% expresses cost performance as a percentage deviation from plan. CPI expresses it as a ratio of value delivered per pound spent. They're mathematically linked: CV% = (1 - 1/CPI) x 100. A CPI of 0.90 equals a CV% of -11.1%. Use CV% for portfolio-level screening and threshold alerts. Use CPI for EAC forecasting and trend analysis.
.webp)
