Why construction ignored 47 years of its own productivity research and called the workaround “best practice”

Project management has a dirty secret it has spent four decades not discussing.

The single metric the discipline elevated to gospel Earned Value Management did not win because it is the most useful framework for running a construction project. It won because the U.S. Department of Defense required it. ANSI/EIA-748 standardized it in 1998 after DoD made compliance mandatory on major systems acquisitions. PMI codified it. The PMP exam tested it. An entire ecosystem of auditors, consultants, and software vendors built revenue around it. Two generations of project managers learned CPI and SPI as the native language of project health.

Meanwhile, the discipline that actually explains why construction bleeds money resource management has been sitting in plain view since the Carter administration. And the industry has spent 47 years pretending it doesn’t exist.

This is not an academic complaint. It is the reason your last megaproject ran over.

The research has been there since 1979

In the late 1970s, the Business Roundtable not exactly a radical body concluded that U.S. construction was in a productivity crisis serious enough to warrant a multi-year study. The result was the Construction Industry Cost Effectiveness (CICE) project, which ran from roughly 1979 to 1983 and produced over 20 reports. Its capstone, More Construction for the Money, identified labor productivity, contracting practices, and management of the work-front as the dominant levers of project performance.

Read that sentence again. Forty-seven years ago, the largest corporate users of construction services in America formally told the industry: the problem is not your cost variance reports. The problem is what your people are actually doing on site.

The industry responded by founding the Construction Industry Institute (CII) in 1983 at the University of Texas at Austin. Since then, CII has produced hundreds of research publications on workforce productivity, direct work rate sampling, foreman delay surveys, craft time-on-tools studies, and benchmarking across thousands of projects. The findings have been remarkably consistent across four decades:

Across published work-sampling research, direct work actual hands-on, value-adding craft labor has consistently landed between roughly 30% and 55% of paid craft hours on industrial construction projects. The rest is waiting, traveling, looking, fetching, reworking, attending unscheduled meetings, breaks, early quits, late starts, and the organizational friction that fills the gaps. CII’s RT 215 program documented this from the craft workers’ perspective across 1,996 craft respondents and 83 productivity factors (CII RS 215-1, Work Force View of Construction Productivity, 2006). CII’s RT 252 program the Construction Productivity Research Program that ran from 2007 through five formal phases defined construction productivity along three explicit dimensions: direct work rate, labor productivity (hours per unit), and rework rate (RS 252-1 through RS 252-1d, 2009–2013).

These are not contested figures. They are the most replicated findings in construction productivity research, sitting in plain view in the CII catalog for anyone who chooses to read them. A recent published analysis of 84 industrial projects across 2015–2023 found mean direct work rates of 46.4% on traditional projects and 52.9% on projects implementing WorkFace Planning (Neve et al., Frontiers in Built Environment, 2025). Even at the upper bound of best-in-class performance, roughly half of every paid craft hour does not produce installed work.

McKinsey’s 2017 Reinventing Construction report reaffirmed the broader stagnation: global construction labor-productivity growth has averaged just 1% per year over the past two decades, compared with 2.8% for the total world economy and 3.6% in manufacturing. U.S. construction productivity is lower today than it was in 1968. The opportunity cost of closing the gap, by McKinsey’s estimate, is $1.6 trillion annually.

So we have known for almost half a century that the labor we pay for produces installed work less than half the time.

And the industry’s flagship management discipline the one it spent four decades canonizing does not measure this, At all.

How EVM won, and why it wasn’t on merit

EVM is a good cost-and-schedule reconciliation framework. I will say that plainly so the EVM apologists can lower their pitchforks. It is a useful tool for what it does.

But it did not become dominant because project managers chose it after considering alternatives. It became dominant for five structural reasons that have nothing to do with whether it actually improves project outcomes:

1. It had a federal mandate. EVM’s lineage traces to DoD’s Cost/Schedule Control Systems Criteria (C/SCSC), formalized in 1967 under DoDI 7000.2. NDIA and the Electronics Industries Alliance jointly published EIA-748 in June 1998, codifying the modern 32-criteria EVMS standard, and DoD adopted it that August for Major Defense Acquisition Programs. EVMS compliance has been embedded in DFARS 252.234-7002 ever since. NASA, DOE, and civilian agencies followed. The standard is now maintained by SAE International, with EIA-748-D published in 2019 and EIA-748-E in development. Resource management never had a regulatory tailwind. Mandates drive adoption. Research papers do not.

2. It compresses to two numbers. CPI 0.92 and SPI 0.95 fit on an executive dashboard and require no explanation. Utilization deficit, mobilization tax, crew composition drag, and whitespace require context, a willingness to look at the work-front, and an executive who tolerates ambiguity. Executives reward the dashboard, not the truth.

3. It uses data that already exists. Every project has a budget and actuals. Earned value math runs on accounting data the CFO already collects. Resource utilization requires time-on-tools sampling, badge data, foreman delay surveys, and field observation. Most ERP systems don’t capture this. Most owners refuse to fund its collection. The data problem is real, but it is also a choice.

4. It serves the funder policing the spender. EVM is a procurement compliance instrument dressed up as a management discipline. The funder wants to know if the contractor is on budget. The contractor wants to defend their margin. EVM is the agreed accounting protocol for that dispute. It optimizes for contractual posture, not productive output.

5. It indicts the contractor. Resource management indicts the owner. This is the one nobody wants to say out loud, so I will.

A cost variance can always be blamed on scope creep, vendor performance, or change order disputes. A utilization deficit cannot. It points squarely back at the owner’s planning maturity, contracting strategy, work-front sequencing, permit-and-isolation processes, materials staging, and the silo structure that prevents trades from getting clean work fronts in the first place.

The owner is the only party in the contracting chain who has both the visibility and the authority to fix the 55/45 problem. The owner is also the party with the least incentive to surface a metric that says the work-front mess is theirs to clean up.

So nobody surfaces it. And the research literature sits on the shelf.

What PMI’s framework actually says about resources

Open the PMBOK Guide. Any edition. Look up “resources.”

You will find guidance on planning, estimating, acquiring, developing, managing, and controlling resources as inputs to project execution. You will find almost nothing on measuring whether those resources, once acquired, are actually producing installed work at a rate that approximates the rate you are paying for.

The framework treats resources as a procurement problem. The research has been telling us for 47 years that they are a productivity problem.

This is not an oversight. It is a category error baked into the foundation of the discipline. PMI’s center of gravity is knowledge work IT projects, product launches, business transformations where the assumption that “an assigned resource is a working resource” is roughly correct. In construction, that assumption is wrong by a factor of two.

A PMP-credentialed project manager can run a $500M industrial project, generate weekly EVM reports of immaculate quality, hit a CPI of 1.02 at handover, and never once measure the fact that the project paid for roughly twice the labor hours it received. Because nothing in the framework requires them to look.

The math when you actually measure it

Pick a number. Any number.

If your craft labor budget on a major industrial project is $200M and the most recent benchmarked data puts direct work rate at roughly 46% on traditional projects, then $108M of that budget is paying for time that did not advance the work-front. That is the good scenario.

Some of that is unavoidable. Travel time exists. Toolbox talks exist. Some delays are weather or genuine surprises. A realistic best-in-class direct work rate, sustained, sits in the low- to mid-50s and getting there is hard, deliberate work, not a poster campaign about safety culture. CII’s RT 252 Phase V research concluded that wrench time could be increased by approximately 15% on average through disciplined application of activity analysis tools.

The gap between a 46% direct work rate and a 55% direct work rate is $18M on a $200M labor budget. The gap between 40% and 55% is $30M. That is not a rounding error. That is a megaproject’s contingency. That is the difference between a profitable contractor and a litigated one. That is the difference between an owner who delivers on schedule and an owner who is in front of a regulator explaining the variance.

EVM cannot see this gap. By design. The earned value calculation treats every hour billed as an hour earned at the planned rate. The framework is mathematically incapable of distinguishing a productive hour from an unproductive one. You can have a CPI of 1.00 and a direct work rate of 28% on the same project, and the discipline will report it as healthy.

That is not project controls. That is bookkeeping in a project controls uniform.

Even the people who mandated EVM stopped believing in it

If EVM were the management discipline its defenders claim, you would expect its strongest advocates to be the agencies that pioneered it. The opposite is true. The federal apparatus that built and mandated EVM has been quietly stepping back from it for over a decade.

In January 2018, the Section 809 Panel, a 16-member advisory body established by Congress under the FY2016 National Defense Authorization Act to recommend acquisition reforms published Volume 1 of its Final Report. Section 4 was devoted entirely to the use of EVM on software programs running Agile methods. The Panel’s specific recommendation, #19, was bounded: eliminate the EVM mandate for Agile software development. But the language it used to justify the recommendation was not bounded. In what should give any defender pause, the Panel concluded that “another substantial shortcoming of EVM is that it does not measure product quality.” It went on to observe that a program could perform ahead of schedule and under cost on every EVM metric while delivering a capability that was unusable to the customer. The Panel’s broader implementation language read more sweepingly: program managers should be able to choose appropriate monitoring methods for their programs rather than be mandated to use EVM.

The contractor base agrees. In the 2010 Grant Thornton Government Contractor Industry Survey, of the federal contractors subject to mandatory EVMS, only 37% believed EVM was a cost-effective management approach. The 2013 18th Annual Survey, conducted with a different respondent pool, reached substantially the same conclusion: only 41% of EVMS-required contractors considered it a cost-efficient management tool, and only 34% reported ever receiving meaningful feedback from government personnel on the EVM data they were required to produce.

Read that again. After four decades of regulatory enforcement, fewer than half of the contractors forced to use EVM believe it is worth what it costs them to produce. This is not what a healthy management discipline looks like. This is what regulatory capture looks like.

The walk-back is no longer just sentiment — it is showing up in the regulations themselves. In September 2015, DoD Class Deviation 2015-O0017 raised the validated-EVMS threshold from $50M to $100M, meaning a contractor on a $99M cost-type contract no longer needed a formally validated system. The pace has accelerated. As part of the May 2025 Federal Acquisition Regulation Overhaul, DFARS Subpart 234.2 was rewritten and the EVMS reporting threshold was lifted from $20M to $50M (published in the Federal Register at 90 FR 5727, January 17, 2025, and implemented through DFARS Class Deviation 2026-O0011 in February 2026). For firm-fixed-price contracts of any dollar value, the current DFARS now explicitly discourages the use of EVM. The Pre-Award and Post-Award Integrated Baseline Review clauses were stripped out. DCMA has restructured its EVMS Compliance Metrics into a tiered system that explicitly deprioritizes most checks. EIA-748 itself is being reissued as Revision E, collapsing the 32 guidelines down to 27.

The mandate that built the EVM industry is being unwound in real time, by the same institutions that built it.

This does not mean EVM is going away tomorrow. It remains required for major defense development contracts, the IBR is still mandatory, and the audit ecosystem is too deeply entrenched to disappear. But the trajectory is unmistakable. The Pentagon believes in EVM with substantially less conviction than it once did. Civilian project managers in construction are roughly a decade behind that recognition.

If the federal procurement system that invented EVM is now publicly admitting that the metric does not measure product quality, the rest of the discipline should be asking what it has been pretending to manage for forty years.

Stop calling it best practice

EVM is a federally-mandated procurement compliance tool that the discipline has spent 40 years dressing up as a management philosophy. It is fine at what it does. It is not what construction project managers should be optimizing toward.

Resource management measured, instrumented, made visible, with the same rigor we apply to cost variance is the actual discipline of running construction. The research has been there since 1979. The frameworks exist. The metrics exist. The benchmarking data exists. None of it is novel. All of it has been ignored because the answer points at the wrong people.

I am tired of pretending this is a complicated debate. It is not. It is a courage problem.

If you are a project director on a program above $100M and you cannot tell me your craft direct work rate within five percentage points, you do not have a project controls function. You have an accounting function with a project controls letterhead.

The Business Roundtable said this in 1983. CII has been saying it ever since. McKinsey said it again in 2017. The next megaproject your organization underwrites will say it once more, in the form of a $40M overrun nobody can quite explain.

Forty-seven years of research is not an academic curiosity. It is a standing indictment. The only question is when the industry decides to read it.

Kyle Mussmacher is a Project Director with 30 years in industrial construction and the author of The Utilization Deficit: Why Construction Bleeds Money in the Gaps Nobody Measures. He writes at constructionresourceutilization.com.

---

Sources

• Business Roundtable, More Construction for the Money: Summary Report of the Construction Industry Cost Effectiveness Project, January 1983. (archived PDF)
• Business Roundtable, Construction Productivity Measurement, Report No. A-1, 1982.
• Business Roundtable, CICE: The Next Five Years and Beyond, 1988.
• Construction Industry Institute (CII), Work Force View of Construction Productivity, RS 215-1, 2006.
• CII, Construction Productivity Research Program — Phases I–V, RS 252-1 series, 2009–2013, including IR 252-2a Guide to Activity Analysis, IR 252-2b A Guide to Construction Rework Reduction, and IR 252-4d The Construction Productivity Handbook.
• CII Research Team 319, Advanced Work Packaging: Design through Workface Execution, 2013–2015 (designated CII Best Practice in 2015).
• Neve et al., “Examining the impact of Advanced Work Packaging and WorkFace Planning on direct work rates of on-site construction workers: a comparative analysis,” Frontiers in Built Environment, 2025.
• McKinsey Global Institute, Reinventing Construction: A Route to Higher Productivity, February 2017.
• EIA-748, Earned Value Management Systems, first published by NDIA and EIA in June 1998; currently EIA-748-D (SAE International, 2019), with EIA-748-E in development (reducing guidelines from 32 to 27). (Note: commonly referred to as “ANSI/EIA-748,” though ANSI dropped from the designation in 2005.)
• U.S. Department of Defense, DoDI 5000.02, Operation of the Adaptive Acquisition Framework; DFARS 252.234-7002, Earned Value Management System.
• Section 809 Panel, Report of the Advisory Panel on Streamlining and Codifying Acquisition Regulations, Volume 1, January 2018, Section 4 (Earned Value Management for Software Programs Using Agile), Recommendation #19. (DTIC — full report) (DTIC — Recommendation 19)
• Grant Thornton LLP, 15th Annual Government Contractor Industry Survey (2010), EVM cost-effectiveness findings; 18th Annual Government Contractor Industry Survey (2013), EVMS cost-efficiency findings (reported via Grant Thornton press release, January 16, 2013).
• DoD Class Deviation 2015-O0017, Earned Value Management System Threshold (September 28, 2015), raising the validated-EVMS compliance review threshold from $50M to $100M.
• Federal Register, 90 FR 5727 (January 17, 2025), updates to DFARS Subpart 234.2 — Earned Value Management System; DFARS Class Deviation 2026-O0011 (February 2026); FAR Subpart 34.2 (May 2025 FAR Overhaul revisions).
• DCMA Manual 2303-01 Volume 3, Surveillance: Earned Value Management; DCMA EVMS Compliance Metrics (DECM) v8.0 prioritization framework.