Capital programme leaders across the GCC review Earned Value Management dashboards every reporting cycle, and many of them have learned to be quietly skeptical of what those dashboards say. Cost Performance Index and Schedule Performance Index readings can sit inside an acceptable band for months while completion dates slip and contingency reserves are drawn down behind the scenes. The gap between what those dashboards report and what is actually happening on site is not simply a data entry problem. It is a structural characteristic of the two indices themselves, one that becomes more pronounced as a programme approaches substantial completion, and one that most steering committees have never been shown how to read around.
The region’s giga-programmes make this gap expensive rather than academic. A 2025 PwC Middle East capital projects survey found that 44 percent of respondents attributed cost increases on Middle East projects to regulatory compliance pressure, and that organizations are increasingly judging project health by financial return and quality rather than strict adherence to the original schedule baseline, a shift that raises the stakes for any metric meant to flag trouble early. Earned Value Management was built to be that early warning system. When CPI and SPI are read without correction, the warning arrives too late to matter, often after the float on the critical path has already been consumed and the recovery options have narrowed to acceleration measures that carry their own cost penalty.
The same survey found that digital investment is accelerating even as the underlying discipline lags behind it: cloud platforms are now used by 89 percent of respondents, yet site-based data integration sits at only 48 percent against 88 percent for office-based systems, a gap that leaves the field data feeding those calculations less current than the dashboards built on top of it. A platform can only be as predictive as the data reaching it, and closing that gap is a project controls problem before it is a technology one.
The fix is not a new metric. It is a disciplined way of reading the two metrics that are already available, paired with the kind of project controls infrastructure that keeps cost and schedule data actually connected rather than reconciled after the fact: resource-loaded schedules integrated with cost baselines, so that outputs reflect what the critical path is actually doing, not just what has been spent against what was planned. That is the discipline Kairos’s Project Control Solutions practice is built to support.
Earned Value Management only earns the word predictive when its two headline indices are corrected for the mathematical behavior that makes them converge toward comfort as a project nears completion.
The Schedule Performance Index Converges Toward One Regardless of Real Delay
SPI is calculated as earned value divided by planned value, and that ratio has a built-in ceiling. As a project approaches its final activities, the denominator, total planned value, and the numerator, cumulative earned value, both approach the same fixed total, so SPI mathematically drifts toward 1.0 no matter how late the programme actually finishes. A project controls team relying on Earned Value Management without adjustment can watch SPI climb back toward a healthy reading in the final quarter of a programme that is, in real terms, months behind. This is not a data quality failure. It is the formula doing exactly what it was designed to do, which is why practitioners increasingly pair SPI with a time-based measure rather than treating the dollar-based ratio as sufficient on its own.
The practical consequence on a GCC capital programme is a false sense of recovery precisely when leadership attention has already moved elsewhere. A steering committee that watched SPI dip to 0.85 mid-programme and then climb back toward 0.97 in the final reporting cycles will often read that trend as evidence the team clawed back lost time. In most cases nothing of the sort happened. The formula simply ran out of remaining planned value to divide against, and the metric, taken at face value, rewarded that arithmetic with a reassuring number.
Earned Schedule Corrects What SPI Cannot See Near Completion
The earned schedule technique, first published by Walter Lipke in 2003 and developed to address this exact weakness, restates schedule performance in units of time rather than currency. Instead of asking how much value has been earned against how much was planned, it asks how far along the planned schedule curve the current earned value would fall, then compares that point to the actual time elapsed. The resulting time-based schedule performance index, denoted SPI(t) and calculated as earned schedule divided by actual time elapsed, does not converge toward 1.0 as completion nears, because it is measuring calendar distance rather than a ratio of two totals racing toward the same endpoint. For GCC capital programmes running multi-year construction phases, adopting earned schedule alongside traditional Earned Value Management gives project controls teams a second reading that stays honest through the final stretch of the works, precisely when the traditional SPI is least trustworthy.
Implementing earned schedule does not require replacing existing Earned Value Management systems. It requires an additional calculation layer, drawing on the same cumulative planned value and earned value curves already produced for CPI and SPI reporting, and a reporting template that presents the time-based figure alongside the traditional index rather than in place of it, so that stakeholders can see where the two readings diverge and ask why.
The additional calculation layer also gives project controls teams a natural place to record forecast completion dates derived from actual time elapsed rather than a ratio-based estimate, which matters most on contracts where liquidated damages are tied to calendar dates rather than percentage progress. A contractor working under a fixed completion date cares less about a favorable percentage than about the actual month a milestone will be achieved, and earned schedule speaks in exactly that unit.
Cost Performance Index Needs the Same Scrutiny as Schedule
CPI carries its own version of the same problem. Because it compares earned value to actual cost cumulatively, a period of cost discipline early in a programme can mask a deteriorating trend later, since the cumulative average smooths out recent performance. Research published through PMI’s knowledge library on integrating scheduling and EVM metrics demonstrates that schedule variance expressed in dollars is intuitively meaningless to most stakeholders, since teams relate to schedule in units of time, not currency, and recommends converting variance into a duration figure by dividing it against average planned value rates. The same discipline applies to CPI: a rolling, recent-period CPI calculated over the trailing reporting window catches a cost trend the cumulative figure will not show for several more cycles.
A programme that has run efficiently for eighteen months and then enters a difficult procurement phase will show that difficulty in a rolling three-month CPI calculation well before the cumulative figure moves enough to trigger a governance review. Reading the two figures on both timeframes, cumulative for the historical record and rolling for the current trend, closes that detection gap without adding a new data source.
Rolling CPI is also easier to reconcile against subcontractor and package-level performance than the cumulative figure, since a single problematic package cannot hide inside eighteen months of otherwise healthy averages the way it can in a portfolio-wide cumulative reading. Project controls teams that report both figures side by side give commercial managers an earlier and more specific signal about which package, rather than which month, needs intervention.
Consider an illustrative programme with a total budget of AED 800 million, eighteen months into a thirty-month construction phase. A cumulative CPI of 0.98 looks close enough to break-even to pass most steering committee reviews without comment. The same programme’s trailing three-month CPI, calculated only on work performed in the most recent reporting quarter, reads 0.89, a gap wide enough to represent roughly AED 14 million in cost growth concentrated in recent work rather than spread evenly across the programme to date. The cumulative figure alone would not surface that concentration for several more reporting cycles, by which point the underlying cause, whether a subcontractor default or a late design change, is harder and more expensive to correct.
Only Critical Path Variances Should Drive Forecast Revisions
Not every task contributing to a negative schedule variance threatens the completion date, and treating them as equally urgent is one of the more common misuses of Earned Value Management on capital programmes. The PMI research referenced above documents a case where five late-running tasks produced a substantial dollar variance, yet only two of those tasks sat on the critical path and put the finish date at real risk. A capital project management software platform capable of tagging each work package against the live critical path turns this from a manual triage exercise into a standing filter, so that project controls staff spend their attention on the variances that matter rather than reviewing every red flag with equal weight. The Kairos insight on capital project management software explores how integrated software platforms remove exactly this kind of manual reconciliation burden from project controls teams working across multiple contract packages.
This distinction matters most during commercial and financing reviews, when a lender or joint venture partner asks why a variance report shows five red items but the forecast completion date has not moved. Being able to answer that question with a critical-path-weighted view of EVM, rather than a flat list of every task behind schedule, is what separates a credible project controls function from one that generates noise.
Predictive Forecasting Requires SPI and CPI to Talk to Each Other
A single index in isolation, whether cost or schedule, tells only part of the story. The estimate at completion calculated from CPI alone assumes cost performance holds steady for the remainder of the programme, while a forecast built purely from SPI ignores whether the organization can actually accelerate enough to recover lost time within remaining float. Combining the two, using a formula such as estimate at completion adjusted by both cost and schedule performance, produces a forecast that reflects the operational reality of the remaining works rather than a single dimension of it. This is where artificial intelligence and predictive analytics increasingly earn their place inside Earned Value Management practice, correlating historical CPI and SPI trends against thousands of comparable work packages to flag which current readings are likely to worsen rather than self-correct. The Kairos insight on AI project management argues that predictive models built on this kind of historical pattern recognition catch deterioration earlier than a human reviewer scanning a monthly report ever will.
The combined forecast also gives project controls teams a defensible number to bring into a commercial negotiation, since it draws on both cost and schedule evidence rather than whichever index happens to look more favorable that month. Earned Value Management produced this way stops being a compliance artifact and starts functioning as the basis for an actual decision about resequencing, resourcing, or recovery planning.
That promise is only as good as the data reaching the model. The same PwC survey cited earlier found site-based data integration at just 48 percent against 88 percent for office-based systems, so a predictive layer trained on lagging site inputs produces a faster, more confident wrong answer rather than a better one. Closing that data gap has to happen before adding a predictive layer to EVM, not after.
Project Controls Governance Determines Whether the Correction Happens
None of the technical corrections above matter if the organization running the programme has no governance mechanism requiring their use. Earned Value Management becomes truly predictive only when the project controls function has the mandate, the resource-loaded schedule integration, and the reporting cadence to apply earned schedule, rolling CPI, and critical-path-weighted variance as standard practice rather than optional analysis reserved for troubled projects. That mandate has to sit above individual project teams, since a controls discipline applied inconsistently across a portfolio of GCC capital programmes produces exactly the false comfort that uncorrected Earned Value Management already creates on its own.
This is the governance model Kairos’s Project Control Solutions practice is designed around, so that the corrected reading, not the raw cumulative CPI and SPI figures, is what reaches the steering committee. A programme that standardizes on the corrected view from mobilization onward avoids having to relearn this lesson the hard way at final account stage.
A programme that trusts uncorrected Earned Value Management through its final quarter is not measuring performance. It is measuring how convincingly the formula can disguise a delay that has already happened.
Conclusion: The Metric Was Never the Problem
The recurring failure on GCC capital programmes is not that Earned Value Management produces bad numbers. It is that project controls functions read CPI and SPI as though both indices behave the same way across the life of a project, when in fact one of them is engineered to flatter itself as completion approaches. Correcting for that behavior, through earned schedule, rolling-period CPI, and critical-path-weighted variance, turns Earned Value Management from a compliance artifact produced for monthly steering committees into a genuine forecasting tool. The organizations that make this correction early, before the final quarter when the distortion is strongest, are the ones whose forecasts still mean something when the programme needs them most. For capital delivery in the GCC, where giga-programme timelines and third-party financing scrutiny leave little room for a forecast that quietly stopped being true months before anyone noticed, that distinction is the difference between Earned Value Management as theater and Earned Value Management as an early warning system worth acting on.