What Is CPI in Construction? Cost Performance Index Explained

<div class="ge-article-wrapper"> <nav class="ge-toc" aria-label="Table of contents"> <p class="ge-toc-label">In this article</p> <ul class="ge-toc-list"> <li><a href="#the-formula">The Formula</a></li> <li><a href="#the-cpi-traffic-light-scale">The CPI Traffic Light Scale</a></li> <li><a href="#what-cpi-really-means-in-pound-terms">What CPI Really Means in Pound Terms</a></li> <li><a href="#the-rule-that-changes-everything-cpi-below-0-90-at-20">The Rule That Changes Everything: CPI Below 0.90 at 20%</a></li> <li><a href="#the-cpi-to-eac-relationship">The CPI to EAC Relationship</a></li> <li><a href="#worked-example-60m-rail-electrification">Worked Example: £60M Rail Electrification</a></li> <li><a href="#cpi-vs-spi-you-need-both">CPI vs SPI: You Need Both</a></li> <li><a href="#common-mistakes-with-cpi">Common Mistakes With CPI</a></li> <li><a href="#frequently-asked-questions">Frequently Asked Questions</a></li> </ul> </nav> <article class="ge-article-body"> <p>The Cost Performance Index (CPI) tells you how much value you're getting for every pound spent on a construction project. A CPI of 1.0 means you're spending exactly what you planned. Above 1.0, you're getting more value per pound than budgeted. Below 1.0, you're burning cash faster than you're delivering work. It's the single most important metric in <a href="/en/earned-value">earned value management</a> and the one most commercial teams wish they'd started tracking sooner.</p> <p>CPI is part of the earned value <a href="/en/earned-value/definitions">definitions glossary</a>. For the full formula breakdown alongside SPI, see the <a href="/en/earned-value/cpi-spi">CPI and SPI formula page</a>.</p> <h2 id="the-formula">The Formula</h2> <div class="ge-formula-box ge-anim"><span class="ge-formula-label">Formula</span><code>CPI = EV / AC</code></div> <p>Where:</p> <ul> <li><strong><a href="/en/earned-value/definitions/earned-value">EV (Earned Value)</a></strong> = the budgeted cost of work actually completed</li> <li><strong><a href="/en/earned-value/definitions/actual-cost">AC (Actual Cost)</a></strong> = what you've actually spent to complete that work</li> </ul> <p>That's it. One division. The hard part isn't the arithmetic. It's getting honest inputs.</p> <h2 id="the-cpi-traffic-light-scale">The CPI Traffic Light Scale</h2> <p>This is the diagram I use when explaining CPI to project boards. Print it out. Stick it next to your EVM dashboard.</p> <pre class="ge-ascii-diagram ge-anim"> CPI TRAFFIC LIGHT ZONES ======================== 0.70 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20 |-------|-------|-------|-------|-------|-------|-------|-------|-------| | RED | RED | AMBER | AMBER | GREEN | GREEN | GREEN | AMBER | RED | | DARK | | RED | | | ON | | | | | | | | | |TARGET | | | | RED (< 0.85): Severe overrun. Recovery plan. Escalate. On £30M project: £5.3M+ overspent against progress. Board needs to know TODAY. AMBER-RED (0.85-0.89): Significant overrun. Formal investigation. Root cause analysis required within 2 weeks. Update EAC and present recovery options. AMBER (0.90-0.94): Over budget but recoverable if acted on NOW. Targeted intervention on worst-performing packages. Monthly monitoring with trend analysis. GREEN (0.95-1.10): Healthy range for complex construction. Monitor monthly. No escalation needed. CPI of exactly 1.00 is rare – don't chase perfection. AMBER-HIGH (1.10-1.15): Suspiciously good. Are you overclaiming EV? Audit progress measurement methodology. Could also mean genuine efficiencies – verify. RED-HIGH (> 1.15): Almost certainly a measurement error. EV is overstated or AC is understated. Stop reporting until you've verified the data. </pre> <p>Most construction projects sit somewhere between 0.90 and 1.05. That's normal. The danger zones are at both extremes.</p> <h2 id="what-cpi-really-means-in-pound-terms">What CPI Really Means in Pound Terms</h2> <p>Let's make this concrete. On a £40M NEC4 Option C project:</p> <div class="ge-table-wrap ge-anim"><table class="ge-table"> <thead> <tr> <th>CPI</th> <th>Value per £1 Spent</th> <th>Total Overrun at Completion</th> <th>What It Feels Like</th> </tr> </thead> <tbody> <tr> <td>1.05</td> <td>£1.05</td> <td>-£1.9M (saving)</td> <td>Comfortable. Gain share territory.</td> </tr> <tr> <td>1.00</td> <td>£1.00</td> <td>£0 (on target)</td> <td>Rare. Enjoy it while it lasts.</td> </tr> <tr> <td>0.97</td> <td>97p</td> <td>+£1.2M</td> <td>"We'll claw it back." (Maybe.)</td> </tr> <tr> <td>0.93</td> <td>93p</td> <td>+£3.0M</td> <td>Formal cost review needed.</td> </tr> <tr> <td>0.88</td> <td>88p</td> <td>+£5.5M</td> <td>Pain share is biting. Recovery plan required.</td> </tr> <tr> <td>0.82</td> <td>82p</td> <td>+£8.8M</td> <td>Crisis. The project is haemorrhaging.</td> </tr> <tr> <td>0.75</td> <td>75p</td> <td>+£13.3M</td> <td>This project might not be viable.</td> </tr> </tbody> </table></div> <p>I've put "what it feels like" in the last column because the numbers alone don't convey the commercial reality. A CPI of 0.88 on a £40M project means the Contractor is heading for a £5.5M cost overrun against target. On a typical NEC4 Option C share range, that's potentially £2M to £3M of pain hitting the Contractor's P&L. That's not a spreadsheet problem. That's a board meeting.</p> <h2 id="the-rule-that-changes-everything-cpi-below-0-90-at-20">The Rule That Changes Everything: CPI Below 0.90 at 20%</h2> <p>Here's something most earned value textbooks won't tell you. Once a project's cumulative CPI drops below 0.90 by the 20% completion mark, it almost never recovers.</p> <p>The US Department of Defense tracked this across hundreds of contracts and published the finding in the 1990s. Construction is messier than defence procurement, but the principle holds. I've tracked CPI on a dozen NEC4 Option C contracts over the last five years, and early CPI predicts final CPI with uncomfortable accuracy.</p> <p>Why? Because by the time you've spent 20% of the budget, the inefficiencies are structural. They're baked into subcontract rates, productivity norms, design coordination failures, and team capability. You can't reorganise your way out of a CPI problem that deep.</p> <pre class="ge-ascii-diagram ge-anim"> THE 20% RULE – CPI RECOVERY PROBABILITY At 20% completion: CPI 0.95+ [==================] High chance of recovery to 1.0 CPI 0.90-0.94 [============] Moderate chance – needs active intervention CPI 0.85-0.89 [======] Low chance – plan for final CPI 0.88-0.93 CPI < 0.85 [==] Near zero – the final CPI will be within +/- 0.05 of current CPI Source: US DoD EVM research, validated across 800+ defence contracts. Construction data is sparser, but the pattern is consistent. </pre> <p>If your CPI is 0.85 at month 6 of a 30-month programme, don't plan your recovery around getting back to 1.0. It won't happen. Plan around 0.88 at best and adjust your <a href="/en/earned-value/eac-etc-tcpi">EAC</a> accordingly.</p> <p>That's not pessimism. That's pattern recognition.</p> <h2 id="the-cpi-to-eac-relationship">The CPI to EAC Relationship</h2> <p>CPI feeds directly into your forecast final cost. This is where CPI stops being an academic metric and starts being the number that determines whether the project makes or loses money.</p> <p><strong>EAC = <a href="/en/earned-value/definitions/budget-at-completion">BAC</a> / CPI</strong></p> <p>This is the simplest <a href="/en/earned-value/definitions/estimate-at-completion">EAC</a> formula. It assumes future cost efficiency will match current cost efficiency. On projects past 20% completion, that assumption is uncomfortably accurate.</p> <div class="ge-table-wrap ge-anim"><table class="ge-table"> <thead> <tr> <th>BAC</th> <th>CPI</th> <th>EAC (BAC / CPI)</th> <th>Overrun</th> </tr> </thead> <tbody> <tr> <td>£60M</td> <td>0.98</td> <td>£61.2M</td> <td>£1.2M</td> </tr> <tr> <td>£60M</td> <td>0.93</td> <td>£64.5M</td> <td>£4.5M</td> </tr> <tr> <td>£60M</td> <td>0.87</td> <td>£69.0M</td> <td>£9.0M</td> </tr> <tr> <td>£60M</td> <td>0.82</td> <td>£73.2M</td> <td>£13.2M</td> </tr> </tbody> </table></div> <p>Look at that escalation. Each 5-point drop in CPI makes the overrun exponentially worse. The jump from 0.93 to 0.87 adds £4.5M to the forecast. On NEC4 Option C, with a typical pain share of 50% beyond the target, the Contractor absorbs £2.25M of that.</p> <h2 id="worked-example-60m-rail-electrification">Worked Example: £60M Rail Electrification</h2> <span class="ge-worked-label">Worked Example</span> <div class="ge-callout ge-anim"> <p><strong>Scenario:</strong> A £60M NEC4 Option C rail electrification programme across 12 stations in the North West. The commercial team tracks CPI monthly.</p> <div class="ge-table-wrap ge-anim"><table class="ge-table"> <thead> <tr> <th>Month</th> <th>% Complete</th> <th>EV</th> <th>AC</th> <th>CPI</th> <th>Trend</th> </tr> </thead> <tbody> <tr> <td>Month 3</td> <td>8%</td> <td>£4.80M</td> <td>£4.90M</td> <td><strong>0.98</strong></td> <td>Stable</td> </tr> <tr> <td>Month 6</td> <td>18%</td> <td>£10.80M</td> <td>£11.22M</td> <td><strong>0.96</strong></td> <td>Slight dip</td> </tr> <tr> <td>Month 9</td> <td>27%</td> <td>£16.20M</td> <td>£17.82M</td> <td><strong>0.91</strong></td> <td>Declining</td> </tr> <tr> <td>Month 12</td> <td>35%</td> <td>£21.00M</td> <td>£24.14M</td> <td><strong>0.87</strong></td> <td>Falling fast</td> </tr> </tbody> </table></div> <p>At month 6, the CPI of 0.96 didn't raise alarms. "We're within tolerance." By month 9, it's 0.91. Still amber. "We'll recover in the fit-out phase." Month 12: 0.87. The project has passed 20% completion and the 20% rule applies.</p> <p><strong>EAC at month 12:</strong> £60M / 0.87 = <strong>£68.97M</strong> (a £8.97M overrun)</p> <p>The commercial manager investigates. Three root causes account for 80% of the cost inefficiency:</p> <ol> <li><strong>Track possession productivity</strong> is 30% below plan because Network Rail possessions are shorter than priced (7-hour windows instead of 10-hour windows). £3.2M of the overrun.</li> <li><strong>OHLE design changes</strong> have been instructed but not all notified as compensation events. The team missed two CEs worth an estimated £1.8M. If recovered, BAC adjusts upward and CPI improves.</li> <li><strong>Subcontract signalling package</strong> is overspending due to abortive work from interface clashes. £2.1M impact.</li> </ol> <p><strong>Actions taken:</strong></p> <ul> <li>Notify the two missed CEs immediately (clause 61.3 time bar applies; one is 6 weeks old, the other 3 weeks)</li> <li>Renegotiate possession windows with Network Rail through the early warning process (clause 15)</li> <li>Issue a recovery programme to the signalling subcontractor</li> </ul> <p>If both CEs are accepted, BAC increases to £61.8M. Revised CPI = £21.0M / £24.14M = still 0.87. But revised EAC = £61.8M / 0.87 = £71.03M. Wait. That's worse.</p> <p>Not quite. The CEs also increase EV (because more scope was always there but wasn't being measured). Revised EV = £21.0M + £1.1M (value of CE work already done) = £22.1M. Revised CPI = £22.1M / £24.14M = <strong>0.915</strong>. Revised EAC = £61.8M / 0.915 = <strong>£67.5M</strong>.</p> <p>Still a £5.7M overrun against target, but £3.3M better than the pre-CE-recovery forecast. That's the commercial value of spotting missed compensation events through proper CPI analysis.</p> </div> <h2 id="cpi-vs-spi-you-need-both">CPI vs <a href="/en/earned-value/definitions/schedule-performance-index">SPI</a>: You Need Both</h2> <p>CPI measures cost efficiency. SPI measures schedule efficiency. You need both, but they tell you different things.</p> <p>A project can have a CPI of 0.88 (over budget) and an SPI of 1.05 (ahead of programme). That means you're throwing money at the problem to maintain progress. Or CPI of 1.10 (under budget) and SPI of 0.85 (behind programme). That means you're saving money by going slowly. Neither is healthy without understanding the other.</p> <p>For the full diagnostic framework, see the <a href="/en/earned-value/cpi-spi">CPI and SPI formula page</a>.</p> <h2 id="common-mistakes-with-cpi">Common Mistakes With CPI</h2> <p><strong>Using CPI too early.</strong> In the first two to three months, CPI is volatile. One large material delivery or a slow start on site can skew the ratio wildly. I've seen a graduate QS present a £6M overrun forecast based on two months of data. Wait for three months of stable data before treating CPI as a trend indicator.</p> <p><strong>Mismatched cut-off dates.</strong> If your cost report runs to the 25th of the month but your progress assessment is done on the 30th, your AC and EV are measuring different time periods. CPI becomes noise, not signal. Align your cut-off dates. It sounds obvious but it's the most common EVM error I see on UK projects.</p> <p><strong>Ignoring compensation events.</strong> On NEC4 Option C, every implemented CE adjusts BAC. If BAC is wrong, EV is wrong. If EV is wrong, CPI is wrong. Always update your baseline when CEs are implemented.</p> <p><strong>Averaging CPI across packages to hide problems.</strong> A project CPI of 0.95 could mean every package is at 0.95, or it could mean four packages are at 1.02 and one catastrophic package is at 0.72. Always drill down. The aggregate CPI is useful for board reporting. The package-level CPI is where the action is.</p> <p><strong>Confusing cumulative CPI with periodic CPI.</strong> Cumulative CPI (total EV / total AC from day one) is stable and reliable. Periodic CPI (this month's EV / this month's AC) is volatile and useful for spotting sudden changes. Report both. Use cumulative for forecasting, periodic for early warning.</p> <div class="ge-product-note ge-anim"> <p><strong>How Gather helps.</strong> 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. <a href="https://gatherinsights.com/contact">Book a demo</a> to see it working on a live NEC4 project.</p> </div> <h2 id="frequently-asked-questions">Frequently Asked Questions</h2> <h3>What is a good CPI for a construction project?</h3> <p>Between 0.95 and 1.10. Below 0.95 means you're consistently overspending against earned progress. Above 1.10 is rare on complex projects. If you're seeing it, either your team has genuinely found efficiencies or your EV measurement is overstating progress. On NEC4 Option C, a CPI of exactly 1.0 means you're tracking exactly to the target, which triggers no pain/gain share.</p> <h3>Can CPI improve once it drops below 0.90?</h3> <p>Theoretically yes. Practically, it almost never recovers to 1.0. If CPI is 0.87 at the 20% mark, plan for a final CPI between 0.87 and 0.92. Don't build a recovery plan around getting back to 1.0. You'll waste time on fantasy forecasts instead of managing the overrun.</p> <h3>How often should I calculate CPI?</h3> <p>Monthly, aligned with your cost reporting cycle. Some teams calculate it fortnightly on high-risk packages. Weekly is overkill unless you're in a crisis and need to track the impact of corrective actions in near real-time. The <a href="/en/earned-value/definitions/control-account-manager">control account manager</a> should produce CPI at control account level every month.</p> <h3>What's the difference between CPI and margin?</h3> <p>CPI measures cost efficiency against the earned value baseline. Margin measures profitability against your tender price. You can have a CPI of 0.95 (5% cost overrun on earned work) and still be making margin if your tender included risk allowances. They're related but measure different things. CPI is a performance metric, margin is a commercial outcome.</p> <h3>How does CPI relate to EAC?</h3> <p>The simplest relationship: <a href="/en/earned-value/definitions/estimate-at-completion">EAC</a> = BAC / CPI. A CPI of 0.90 on a £50M project gives an EAC of £55.6M (an £5.6M overrun). This assumes future efficiency matches current efficiency. For more sophisticated EAC calculations, see the <a href="/en/earned-value/eac-etc-tcpi">EAC, ETC, and TCPI page</a>.</p> </article> </div>