The Cost Growth Index (CGI) is a single ratio that tells you how much a project's forecast final cost has grown compared to its original budget. The formula is simple: CGI = EAC / BAC. A CGI of 1.0 means the project is forecast to land exactly on budget. A CGI of 1.15 means you're looking at a 15% overrun. It's the metric that board members and programme directors care about most, because it strips away the detail and gives them one number: how much is this going to cost us compared to what we approved?
This term is part of the earned value definitions glossary. For the underlying formulas that feed into CGI, see the EAC, ETC, and TCPI formula page.
The Formula
CGI = EAC / BACWhere:
- EAC (Estimate at Completion) = forecast total cost at project completion
- BAC (Budget at Completion) = original approved budget (adjusted for approved scope changes)
That's it. One division. But the number it produces changes boardroom conversations.
What CGI Tells You at a Glance
CGI Scale – Portfolio Health Dashboard 0.90 0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 |-------|-------|-------|-------|-------|-------|-------|-------| | UNDER BUDGET | ON | MILD | MODERATE | SEVERE | CRISIS| | (savings) | TRACK | CREEP | OVERRUN | OVERRUN | | | | | | | | | BAC = £10M examples: CGI 0.95 → EAC £9.5M (£500K saving) CGI 1.00 → EAC £10.0M (on budget) CGI 1.08 → EAC £10.8M (£800K overrun) CGI 1.15 → EAC £11.5M (£1.5M overrun) CGI 1.25 → EAC £12.5M (£2.5M overrun – red flag)
| CGI Range | Interpretation | Board Response |
|---|---|---|
| Below 0.95 | Significant underspend | Check if scope is being delivered. Savings are good; descoping is not. |
| 0.95 to 1.00 | On or under budget | Healthy. Monitor monthly. |
| 1.01 to 1.05 | Mild cost growth | Normal on complex projects. Investigate drivers. |
| 1.06 to 1.10 | Moderate overrun | Cost recovery plan needed. Quarterly board update. |
| 1.11 to 1.20 | Severe overrun | Formal review. Consider programme replanning. |
| Above 1.20 | Crisis territory | Escalate to sponsors. Assess viability. |
Why CGI Beats Absolute Numbers for Portfolio Reporting
Here's the problem with reporting absolute cost variances to a board. If you tell a programme director that Project A is £800K over budget and Project B is £1.2M over budget, they'll focus on Project B. But what if Project A has a BAC of £5M (CGI = 1.16) and Project B has a BAC of £60M (CGI = 1.02)?
Project A is in far worse shape. CGI makes that obvious. Absolute numbers don't.
I've sat in programme reviews where the board spent 45 minutes debating a £2M overrun on a £120M package (CGI 1.017) while completely ignoring a £400K overrun on a £2.5M enabling works package (CGI 1.16). The small package was the one haemorrhaging margin. CGI would have flagged it immediately.
Worked Example: Portfolio of Six Projects
Worked ExampleScenario: A Tier 1 contractor's regional programme includes six live NEC4 Option C projects. The commercial director wants a single dashboard for the February 2026 board meeting.
| Project | BAC | EAC | CGI | Status |
|---|---|---|---|---|
| A14 Road Improvement | £28,000,000 | £27,160,000 | 0.97 | Under budget |
| Midlands Rail Platform | £14,500,000 | £15,225,000 | 1.05 | Mild creep |
| Birmingham Schools Phase 2 | £18,000,000 | £19,260,000 | 1.07 | Moderate |
| Leeds Flood Defence | £42,000,000 | £44,100,000 | 1.05 | Mild creep |
| M62 Bridge Replacement | £8,200,000 | £10,086,000 | 1.23 | Crisis |
| Liverpool Hospital Wing | £35,000,000 | £35,700,000 | 1.02 | On track |
| Portfolio Total | £145,700,000 | £151,531,000 | 1.04 | Mild creep |
The portfolio-level CGI of 1.04 looks manageable. But it hides the M62 Bridge Replacement, which is 23% over budget on a relatively small package. That's £1.886M of overrun on an £8.2M job. Margin is gone. The commercial director's question isn't "why is the portfolio 4% over?" It's "what the hell is happening on the M62 bridge?"
Digging into the M62 Bridge (CGI 1.23):
- Original BAC: £8,200,000
- Two CEs implemented (+£600K) → Adjusted BAC: £8,800,000 (but EAC didn't reduce proportionally)
- Wait. If BAC should be £8.8M after CEs, then CGI = £10,086,000 / £8,800,000 = 1.146
- Still severe, but different from 1.23. The commercial team hadn't updated BAC for implemented CEs.
This is exactly the kind of error CGI surfaces when you look at it critically. Always check that BAC includes approved scope changes before panicking about the ratio.
CGI vs CPI: Different Questions, Different Uses
People often confuse CGI with CPI. They're related but they answer different questions.
| CGI | CPI | |
|---|---|---|
| Formula | EAC / BAC | EV / AC |
| Measures | Total forecast cost growth | Current cost efficiency |
| Timeframe | Whole project (forecast) | To date (actual) |
| Audience | Board, programme directors, sponsors | Project team, commercial managers |
| Updates | Monthly with EAC refresh | Monthly with progress measurement |
CPI tells you how efficiently you're spending right now. CGI tells you where you'll end up. A project can have a current CPI of 0.95 but a CGI of 1.02 if the team believes they'll recover efficiency on later works. Whether that belief is realistic is a separate conversation.
How to Calculate CGI When EAC Methods Differ
CGI is only as reliable as the EAC feeding it. Different EAC methods produce different CGI values:
| EAC Method | Formula | CGI Implication |
|---|---|---|
| EAC = BAC / CPI | Performance-based | CGI = 1 / CPI (mathematically linked) |
| EAC = AC + bottom-up ETC | Judgement-based | CGI reflects team's forecast, not just historical trend |
| EAC = AC + (BAC - EV) | Assumes future = plan | CGI only reflects overrun to date |
| EAC = AC + (BAC - EV) / CPI | Assumes future = past | CGI reflects both overrun to date and efficiency trend |
On UK construction projects, the bottom-up ETC method (where the commercial team forecasts remaining costs package by package) is the most common. It gives a more realistic CGI because it incorporates known future risks, not just historical trends.
Common Mistakes
Comparing CGI across projects without context. A CGI of 1.10 on a highways scheme with 200 unforeseen utility diversions is completely different from a CGI of 1.10 on a straightforward warehouse build. The number tells you the size of the problem. The root cause tells you whether it's recoverable.
Not separating scope growth from cost growth. If BAC increased by £2M through legitimate CEs and EAC increased by £2.5M, the real cost growth is only £500K. CGI should be calculated against the adjusted BAC (post-CE), not the original. I've seen programme reports where CGI was calculated against original contract value, making every project with scope changes look worse than it was.
Treating CGI as a leading indicator. CGI is a lagging metric. It tells you where you are, not where you're heading. A project with CGI of 1.05 today could be accelerating towards 1.15 if the CPI trend is worsening. Always look at CGI alongside the CPI trend line.
Reporting portfolio CGI without the distribution. A portfolio CGI of 1.04 could mean all six projects are at 1.04, or it could mean five are at 0.98 and one is at 1.23. The aggregate hides the outliers. Always report CGI at project level alongside the portfolio 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 a good CGI for a construction project?
Anything at or below 1.05 is generally healthy for complex UK construction projects. Between 1.05 and 1.10 warrants investigation. Above 1.10 means the budget approval is no longer valid and needs formal review. On NEC4 Option C, CGI above 1.00 means the Contractor is heading into pain share territory (depending on the share percentage in the contract data).
How is CGI different from percentage overrun?
They express the same thing differently. A CGI of 1.15 means 15% cost growth. Percentage overrun = (EAC - BAC) / BAC x 100 = (CGI - 1) x 100. CGI is preferred for portfolio dashboards because it's a ratio you can compare across projects of different sizes without conversion.
How often should CGI be reported?
Monthly at project level, quarterly at portfolio and board level. More frequent reporting doesn't help because EAC only changes meaningfully when new cost data or forecasts are produced. If your cost reporting is monthly, your CGI is monthly.
Can CGI be below 1.0?
Yes, and it's good news. A CGI of 0.95 means the project is forecast to complete at 95% of the approved budget. That's a 5% saving. On NEC4 Option C, this puts you in gain share territory. Just make sure the saving is real and not because the team hasn't updated their EAC to reflect emerging risks.
.webp)
