The formulas aren't the hard part. Knowing what to do with the numbers is. This page walks through a complete earned value management example on a construction project - month by month, with actual figures, showing exactly how the story unfolds when a project starts well, hits problems, and claws its way back. If you need the underlying EV formulas, read those first.
The Project: Highway Junction Improvement
Worked ExampleA £12.4M highway junction improvement scheme procured under NEC4 Option C (target cost with activity schedule).
| Parameter | Value |
|---|---|
| Contract form | NEC4 Option C |
| Target Price (BAC) | £12,400,000 |
| Planned duration | 12 months |
| Pain/gain share | 50/50 above and below target |
| Main works | Junction realignment, slip roads, drainage, signals, surfacing |
| Assessment periods | Monthly (last working day) |
Work Breakdown Structure
| Work Package | Description | Budget (£) | % of BAC |
|---|---|---|---|
| WP1 | Preliminaries & site establishment | 1,860,000 | 15.0% |
| WP2 | Earthworks & drainage | 2,480,000 | 20.0% |
| WP3 | Structural works (retaining walls, bridge deck) | 3,100,000 | 25.0% |
| WP4 | Road construction & surfacing | 2,976,000 | 24.0% |
| WP5 | Traffic signals & electrical | 1,240,000 | 10.0% |
| WP6 | Landscaping & finishing | 744,000 | 6.0% |
| Total | £12,400,000 | 100% |
Each work package needs measurable deliverables - metres of kerb laid, tonnes of capping placed, linear metres of drainage installed. Vague progress estimates are how EVM falls apart.
Months 1-3: The Honeymoon
Everything looks fine. Site establishment's on track, early earthworks are progressing. Nobody's looking hard at EV metrics. Why would they?
| Month | PV (£) | EV (£) | AC (£) | CV (£) | SV (£) | CPI | SPI |
|---|---|---|---|---|---|---|---|
| M1 (Apr) | 620,000 | 651,000 | 632,000 | +19,000 | +31,000 | 1.03 | 1.05 |
| M2 (May) | 1,364,000 | 1,395,000 | 1,358,000 | +37,000 | +31,000 | 1.03 | 1.02 |
| M3 (Jun) | 2,232,000 | 2,294,000 | 2,201,000 | +93,000 | +62,000 | 1.04 | 1.03 |
What the numbers say: CPI 1.04 - getting £1.04 of value for every £1 spent. SPI 1.03 - 3% ahead of programme. EAC at M3: £12,400,000 / 1.042 = £11,900,192. A forecast saving of £500K. On a 50/50 pain/gain share, that's a potential £250K gain-share. Looks lovely. Won't last.
What the commercial team should actually think: Don't celebrate. Months 1-3 are prelims and bulk earthworks - the easy bits. Structural works haven't started. A CPI of 1.04 built on muck shifting tells you almost nothing about where the job's heading.
Months 4-6: The Crack Appears
Month 4 is when structural works start. The retaining wall foundations hit unexpected ground conditions - clay with pockets of peat not in the ground investigation report. A compensation event under NEC4 clause 60.1(12). The CE quotation process takes time, and meanwhile the project bleeds money while the team waits for redesigned foundations.
| Month | PV (£) | EV (£) | AC (£) | CV (£) | SV (£) | CPI | SPI |
|---|---|---|---|---|---|---|---|
| M4 (Jul) | 3,348,000 | 3,100,000 | 3,224,000 | -124,000 | -248,000 | 0.96 | 0.93 |
| M5 (Aug) | 4,588,000 | 4,092,000 | 4,423,000 | -331,000 | -496,000 | 0.93 | 0.89 |
| M6 (Sep) | 5,952,000 | 5,146,000 | 5,710,000 | -564,000 | -806,000 | 0.90 | 0.86 |
Three months. That's all it took to swing from green to red. CPI dropped from 1.04 to 0.90. SPI from 1.03 to 0.86. EAC at M6: £12,400,000 / 0.901 = £13,762,486 - a £1.36M projected overrun.
About £800,000 of that overrun is the compensation event, which should adjust the target once agreed. Strip that out and the underlying overrun is closer to £562,000 - still bad, but manageable. This is why EVM and contract management can't live in silos. If you're tracking EVM without adjusting for compensation events, you're making distorted decisions.
Month 7: The Crisis Point
Three things collide in October: (1) the ground conditions CE quotation is still being assessed, (2) a statutory undertaker diverts services three weeks late blocking slip road construction, (3) rainfall is 180% of seasonal average causing 11 lost working days.
| Month | PV (£) | EV (£) | AC (£) | CV (£) | SV (£) | CPI | SPI |
|---|---|---|---|---|---|---|---|
| M7 (Oct) | 7,316,000 | 6,076,000 | 7,068,000 | -992,000 | -1,240,000 | 0.86 | 0.83 |
CPI: 0.86. For every pound spent, only 86p of value is being delivered. EAC: £12,400,000 / 0.8597 = £14,424,490. A £2M+ projected overrun.
But here's what separates experienced analysis from just running formulas. You don't just report the numbers. You decompose them.
EVM Decomposition at Month 7
| Component | CV Impact (£) | SV Impact (£) | Recoverable? |
|---|---|---|---|
| Ground conditions CE (pending) | -620,000 | -480,000 | Yes - CE adjustment to target |
| Statutory undertaker delay | -135,000 | -310,000 | Probably - clause 60.1(5) delay |
| Exceptional weather (October) | -87,000 | -250,000 | Partial - clause 60.1(13) weather |
| Base contract inefficiency | -150,000 | -200,000 | No - contractor's risk |
| Total | -992,000 | -1,240,000 |
Of the £992,000 cost variance, only £150,000 is genuinely the contractor's problem. The rest is either recoverable through CEs or partially recoverable through weather clauses. Without this breakdown, you'd be firefighting a £2M overrun. With it, you're managing a £150,000 inefficiency and processing three legitimate CEs.
Months 8-10: The Recovery
The ground conditions CE is agreed on 14 November at £843,000 (time and cost). The target price adjusts from £12,400,000 to £13,243,000. This is the new BAC.
Once a CE adjusts the target, the EVM baseline must update. Keep reporting against the original £12.4M and the numbers will never make sense - you'll show an artificial overrun that doesn't reflect actual performance. The commercial team implements recovery measures: weekend working on structural elements, accelerated drainage, and a revised construction sequence.
| Month | PV (£) | EV (£) | AC (£) | CV (£) | SV (£) | CPI | SPI |
|---|---|---|---|---|---|---|---|
| M8 (Nov)* | 8,622,000 | 7,560,000 | 8,310,000 | -750,000 | -1,062,000 | 0.91 | 0.88 |
| M9 (Dec)* | 9,945,000 | 9,150,000 | 9,780,000 | -630,000 | -795,000 | 0.94 | 0.92 |
| M10 (Jan)* | 10,872,000 | 10,478,000 | 10,890,000 | -412,000 | -394,000 | 0.96 | 0.96 |
*Months 8-12 use adjusted BAC of £13,243,000. CPI has recovered from 0.86 to 0.96. Look at the trend: 0.91, 0.94, 0.96. Three consecutive months of improvement. That trajectory matters more than any single reading. Revised EAC at M10: £13,243,000 / 0.962 = £13,766,112 - a £523K overrun against the adjusted target.
Months 11-12: The Finish
The final push. Surfacing completes in a dry window. Traffic signals installation runs ahead of schedule. Landscaping is deferred by agreement with no cost impact.
| Month | PV (£) | EV (£) | AC (£) | CV (£) | SV (£) | CPI | SPI |
|---|---|---|---|---|---|---|---|
| M11 (Feb)* | 12,187,000 | 11,940,000 | 12,095,000 | -155,000 | -247,000 | 0.99 | 0.98 |
| M12 (Mar)* | 13,243,000 | 13,243,000 | 13,462,000 | -219,000 | 0 | 0.98 | 1.00 |
Final outturn: Project completes on time. Total Defined Cost: £13,462,000 against adjusted target of £13,243,000. Final CPI: 0.98. Final SPI: 1.00. Final overrun: £219,000. Pain-share to contractor: £109,500.
Compare that to the Month 7 panic when numbers suggested a £2M+ overrun and a million-pound pain-share hit. This is why EVM exists - not to generate reports, but to give the commercial team enough visibility to intervene at the right moment with the right actions.
The Full EVM Dashboard: 12-Month Summary
| Month | PV (£) | EV (£) | AC (£) | CV (£) | SV (£) | CPI | SPI | EAC (£) |
|---|---|---|---|---|---|---|---|---|
| M1 Apr | 620K | 651K | 632K | +19K | +31K | 1.03 | 1.05 | 12,039K |
| M2 May | 1,364K | 1,395K | 1,358K | +37K | +31K | 1.03 | 1.02 | 12,039K |
| M3 Jun | 2,232K | 2,294K | 2,201K | +93K | +62K | 1.04 | 1.03 | 11,900K |
| M4 Jul | 3,348K | 3,100K | 3,224K | -124K | -248K | 0.96 | 0.93 | 12,917K |
| M5 Aug | 4,588K | 4,092K | 4,423K | -331K | -496K | 0.93 | 0.89 | 13,333K |
| M6 Sep | 5,952K | 5,146K | 5,710K | -564K | -806K | 0.90 | 0.86 | 13,762K |
| M7 Oct | 7,316K | 6,076K | 7,068K | -992K | -1,240K | 0.86 | 0.83 | 14,424K |
| M8 Nov* | 8,622K | 7,560K | 8,310K | -750K | -1,062K | 0.91 | 0.88 | 14,553K |
| M9 Dec* | 9,945K | 9,150K | 9,780K | -630K | -795K | 0.94 | 0.92 | 14,088K |
| M10 Jan* | 10,872K | 10,478K | 10,890K | -412K | -394K | 0.96 | 0.96 | 13,766K |
| M11 Feb* | 12,187K | 11,940K | 12,095K | -155K | -247K | 0.99 | 0.98 | 13,377K |
| M12 Mar* | 13,243K | 13,243K | 13,462K | -219K | 0 | 0.98 | 1.00 | 13,462K |
*Months 8-12 use adjusted BAC of £13,243,000 following CE agreement. EAC formula used: BAC / CPI (cumulative). For all EAC methods, see EAC, ETC and TCPI.
Five Key Lessons From This Project
Lesson 1: Early CPI Is Unreliable
The CPI of 1.04 at Month 3 was real, but it was measuring the easy work. Bulk earthworks and prelims have low complexity and well-understood costs. The structural works, where the real cost risk lives, hadn't started. Always weight EVM confidence by the complexity of work completed, not just the percentage.
Lesson 2: Decompose Before You React
Month 7's CPI of 0.86 looked catastrophic. But £842,000 of the £992,000 cost variance was recoverable through compensation events. Without decomposition, the team might have slashed resources to "save money" - exactly the wrong move when the problem isn't efficiency but uncompensated change. Understanding cost and schedule variance at work package level is what makes this decomposition possible.
Lesson 3: Adjust the Baseline or the Numbers Lie
Once the ground conditions CE was agreed at £843,000, the target moved from £12.4M to £13.243M. If the team kept reporting against the original target, every month from November onwards would show an artificial overrun that doesn't reflect actual performance. Update BAC every time a CE is implemented.
Lesson 4: CPI Trend Matters More Than CPI Value
At Month 8, CPI was 0.91 - still below 1.0, still technically overspending. But the trend - 0.86, 0.91, 0.94, 0.96 - showed clear recovery. A single CPI reading is a snapshot. The trend is the film. Always report both. An S-curve makes these trends visible at a glance.
Lesson 5: EVM Without Contract Management Is Just Arithmetic
The real value of this example isn't the formulas. It's the integration with NEC4 contract mechanisms. The compensation event process, target adjustment, and pain/gain share calculation - these are what turn EVM numbers into commercial outcomes. If your EVM system doesn't connect to contract administration, you're doing half the job.
How EVM Connects to NEC4 Target Cost Management
Under NEC4 Option C, the target price is a living number. It changes every time a compensation event is implemented. Most EVM textbooks ignore this, which is why teams struggle to make earned value work on NEC contracts.
When a CE is implemented, it adjusts the Prices (the target). Your BAC must update. If it doesn't, you're measuring performance against a contract position that no longer exists.
Under NEC4, the Project Manager can disallow costs that aren't in accordance with the contract (clause 11.2(26)). Disallowed costs get stripped from Defined Cost used in the pain/gain calculation, but they still show up in actual project spend. Your EVM system needs two AC figures: AC inclusive of all costs (for project control) and AC net of disallowed costs (for commercial outturn). Most spreadsheet-based EVM setups can't do this.
On this project, the final Defined Cost was £13,462,000 against an adjusted target of £13,243,000. Overrun: £219,000. At 50/50, the contractor absorbs £109,500 and the client absorbs £109,500. If EVM had flagged the efficiency problems at Month 5 instead of Month 7, two extra months of intervention could have saved £100-150K - the difference between a small pain-share and a gain-share.
Gather's AI agent reads daily site diary entries, extracts quantities and progress data, and feeds them into an earned value model automatically. On one highways package, it flagged a CPI dip below 0.95 in Week 3 of Month 4 - three weeks before it would have surfaced in a traditional monthly review. See how it works.
Common Mistakes in Construction EVM
- Measuring progress by spend, not output. "We've spent 60% of the budget, so we must be 60% complete." No. EV must be measured by physical output: cubic metres excavated, metres of pipe laid, tonnes of asphalt placed. If you've spent 60% but only completed 40% of the physical work, CPI is 0.67.
- Never adjusting the baseline. On NEC4 Option C, the target changes with every implemented CE. If you're still measuring against the original target six months and four CEs later, your metrics are fiction.
- Ignoring SPI after 80% complete. SPI becomes unreliable in the final 20%. A project can finish four months late and show SPI of 1.0 at completion because EV converges on BAC. Use earned schedule analysis in the final stages instead.
- Treating CPI as fixed. CPI changes. A Month 3 CPI of 1.04 doesn't mean you'll finish under budget. Report the trend, not just the number.
- Not decomposing variances. A negative CV of -£500,000 could mean five completely different things: genuine inefficiency, unpriced change, a pending CE, front-loaded subcontractor payments, or a timing difference. Report the headline without breaking it down and you'll get the wrong reaction every time. For more, see common EVM mistakes in construction.
Quick Reference: All EVM Metrics
| Metric | Formula | What It Tells You | Healthy Range |
|---|---|---|---|
| CV (Cost Variance) | EV - AC | Over or under budget | > 0 |
| SV (Schedule Variance) | EV - PV | Ahead or behind programme | > 0 |
| CPI | EV / AC | Cost efficiency | > 1.0 |
| SPI | EV / PV | Schedule efficiency | > 1.0 |
| EAC | BAC / CPI | Forecast total cost | Close to BAC |
| ETC | EAC - AC | Cost to finish | Decreasing trend |
| TCPI | (BAC - EV) / (BAC - AC) | Efficiency needed to hit budget | < 1.2 |
| VAC | BAC - EAC | Forecast overrun/underrun | > 0 |
For full formula derivations, see the earned value formulas page. For the monthly reporting structure, see the EV report template.
.webp)




