By Evgenii Garde
Everyone blames expensive labour for construction’s productivity crisis. Yet the real problem is not labour costs, but the fact that only a third of the paid day is spent on value-adding work. With labour shortages growing across Europe, digital transformation is becoming the industry’s most powerful tool for overcoming coordination challenges and boosting productivity.
Facts and myths
Construction is among the least productive of the major industries. Output per construction worker has grown at roughly 1% a year for decades, a fraction of the rate seen in manufacturing. The common explanations for this are mostly myths.
The first myth is that construction is unproductive because its workers are slow or unskilled. The second is that the route to a cheaper, more productive industry runs through cheaper labour. Both miss what site studies have shown for half a century: the problem is not the price or the pace of the work, but how little of the paid day is spent doing it.
The industry has a name for time spent actually building: “direct work”, or informally “tool time”. Decades of work-sampling studies converge on an uncomfortable range. The Construction Industry Institute puts direct work at no more than 35 to 45% of a craftsman’s day. A Danish study of renovation projects found value-adding work at just under 30%. A 2024 study of mechanical and electrical trades measured direct work as low as 13.6% for plumbing and heating crews, with workers interrupted 40 to 90 times a day. The consistent finding: a tradesperson spends little more than a third of paid time, and often far less, on the task they are paid for. The rest goes to planning and coordination, and to sourcing and moving materials.
This reframes the economics. Take a representative pound of construction revenue: roughly 40p to labour, 40p to materials, and the remainder split between overhead and profit. Because labour is around 40% of the bill, the instinct is to attack that cost by paying less. But only about a third of that labour is value-adding time. The wage buys a full day; the building receives a third of one.
That gap, between the average cost of an input and its marginal product, is the crux. The productivity lever is not cheaper labour but reallocation. Moving an hour from searching for materials to laying brick raises output without touching the wage rate, the headcount or the worker’s skill. As one practitioner put it in Engineering News-Record: if only 30% of the day is direct work, why obsess over improving that 30% instead of attacking the 70%?
Most of that 70% is what economists would recognise as transaction and coordination cost: the friction of getting the right materials, information and people to one place at one time. On a fragmented site with dozens of subcontractors it is expensive and largely invisible on the invoice. It shows up twice, as wasted labour (walking, waiting, searching) and as wasted materials, with at least 10% of everything delivered to a UK site lost to damage, loss and over-ordering.
Why the industry needs to digitise
Two forces make this coordination overhead newly urgent.
The first is that it has never improved on its own. A study tracking 98 projects from 1972 to 2009 found no rise in direct-work rates across nearly four decades, despite power tools, computers and better machinery arriving throughout. Better hardware in individual workers’ hands did not fix a system-level coordination problem. This is close to what economists call Baumol’s cost disease: sectors that cannot easily raise output per hour see their costs climb relative to more productive parts of the economy, which is one reason a new building costs what it does.
The second force is demography. Europe’s construction workforce is ageing and shrinking. The European Federation of Building and Woodworkers estimates the sector will need up to 1.5 million additional workers between 2023 and 2030, with the industry federation FIEC putting the figure closer to two million, on top of more than a million needed simply to replace retiring workers. When labour cannot easily be added, the only way to raise built output is to raise the share of each existing hour that reaches the work face. Coordination, not headcount, becomes the binding constraint: a textbook case of the theory of constraints, where output is governed by the bottleneck rather than the average.
Where digitisation has actually worked
The gains that matter are not robots laying bricks faster. They come from cutting the planning and sourcing overhead, the two-thirds, and the evidence is now specific.
In design and planning, Building Information Modelling linked to automated scheduling has compressed programming work that took a planner several hours into around twenty minutes; on one large tunnelling project an AI-assisted system cut the response time on a scheduling decision from 45 minutes to 15. The same model prevents waste before it is built. Rework, the cost of building something twice because two designs clashed or a part did not fit, runs at an estimated 4 to 12% of project cost. Automated clash detection runs the architectural, structural and services models against each other and flags conflicts before they reach concrete; it has been found to cut rework costs by 20 to 40%, with machine-learning versions catching more than 90% of collisions in advance. Taken to its conclusion this becomes design for manufacture and assembly, which McKinsey estimates could raise productivity five- to tenfold in the parts of the industry that adopt it.
On site, the clearest wins are in coordination rather than craft. A survey robot such as Boston Dynamics’ Spot walks a fixed route on a schedule, capturing laser scans and progress data straight back into the model: one contractor scanned a full floor in 20 to 60 minutes, and another cut inspection time by more than 95% by sending robots round overnight. Where robots do touch the build, they target the most repetitive tasks. Hilti’s Jaibot takes a drilling plan from the model and drills the overhead holes for mechanical and electrical fixings while a tradesperson supervises, removing one of the most strenuous and repetitive jobs on site.
The common thread is not the hardware but the data. Clash detection, scheduling, drilling and progress-checking all work because a single model flows from the designer’s screen to the field and back. The firms that pull ahead will be the ones that build that digital spine, not the ones that buy the flashiest machine.
Conclusion
The stopwatch settled the first question decades ago: give the value-adding third of the day a larger share and output rises, with no change to wages, headcount or skill. What is new is that the tools to shift that distribution now exist and pay for themselves. The remaining constraint is neither economic nor technological but organisational: how quickly an industry built on fragmented, relationship-driven projects can change the way it works. On that question, the stopwatch has yet to report.
About the Author
Evgenii Garde is Head of Marketing, Tool Services at Hilti GB, specialising in digital transformation and subscription-services growth across construction and heavy industry.
Sources:
-
Construction Industry Institute, craft productivity (“direct work” / “tool time”) research programme.
-
Gong, Borcherding and Caldas, “Effectiveness of craft time utilisation in construction projects”, Construction Management and Economics (2011), covering 98 projects, 1972 to 2009.
-
Görsch, Seppänen, Peltokorpi and Lavikka, study of mechanical and electrical trade productivity (2024).
-
Danish work-sampling study of value-adding work on renovation projects (Aalborg University).
-
Labour shortage: European Federation of Building and Woodworkers / ITUC Just Transition Centre, up to 1.5 million additional construction workers needed 2023-2030; FIEC position paper on labour shortages (2025), estimating around 2 million additional workers by 2030.
-
McKinsey Global Institute, Reinventing Construction: A Route to Higher Productivity (2017) and Delivering on Construction Productivity is No Longer Optional (2024).
-
Material waste figures: UK site over-ordering and loss estimates (Guthrie et al.) and 2025 material-flow analysis of on-site waste efficiency.
-
Technology examples: peer-reviewed 2024 to 2026 studies on BIM-based scheduling and procurement optimisation (MDPI Buildings and Designs).
-
Rework and clash detection: rework estimated at 4 to 12% of project cost, with BIM clash detection cutting rework 20 to 40% (MDPI Buildings, 2026; Kermanshahi et al., 2020).
-
Robotics on site: Boston Dynamics Spot case studies (Swinerton, Turner Construction, Foster + Partners) for autonomous reality capture and progress tracking.
-
Hilti Jaibot: semi-autonomous BIM-driven ceiling-drilling robot for MEP installation (Hilti, launched 2020; Engineering News-Record, 2020).
