Why Is UK Construction Productivity Stuck, and How Do You Get Site Teams to Use the Tech?
UK construction productivity is stuck because site teams don’t adopt the tools bought for them. Between 1997 and 2019, construction output grew 12% in real terms while hours worked rose 23% (ONS). The gap isn’t the technology. It’s fear, complexity and the lack of a feedback loop. Fix those with five steps and the tools start paying back.
Last reviewed: November 2026
Let me start with an uncomfortable truth.
UK construction has spent the last decade buying productivity: software licences, digital platforms, AI dashboards and BIM mandates. And productivity has barely moved. Between 1997 and 2019, construction output grew 12% in real terms while the hours worked to produce it rose 23% (ONS, Productivity in the construction industry, UK: 2021). We are working more hours for each unit of output, not fewer.
We bought a project management platform. We mandated BIM. We ran a three-hour software demo in the site canteen, handed everyone a login and called it going digital.
It wasn’t.
I’ve spent 23 years in construction, 18 of them at a major contractor, latterly as a multi-disciplinary head covering supply chain, planning and business improvement. I’ve seen this pattern on jobs from £5m school extensions to £40m schemes. The gap between what the boardroom believes is happening and what’s actually happening on site is where productivity goes to die.
How far behind is UK construction productivity?
Far enough to notice. In 2025, construction produced about £37 of output per hour worked against about £43 for the whole economy, roughly 86% of the average (our calculation from ONS, Output per hour worked, UK, August 2026). And it has been studied to death. McKinsey’s 2017 report Reinventing Construction ranked construction among the least digitised industries in the world and estimated that closing its productivity gap could add $1.6 trillion a year to the global economy (McKinsey Global Institute, 2017).
Everyone in construction leadership read that number. Most responded by increasing the technology budget. That’s the wrong response, because the productivity gap isn’t a technology problem. It’s a people problem wearing a technology costume.
Here’s what actually happens. A contractor wins a job. Pre-construction sets up a new project platform, configures the workflows and connects it to the ERP. There’s a launch meeting, logins and, almost certainly, a laminated quick-start guide.
Six months in, half the site team are using WhatsApp and a shared spreadsheet. The platform has seventeen documents in it, none of them current. The project manager has stopped logging issues because it takes longer to log them than to deal with them. The dashboard meant to give the leadership team live visibility shows data three weeks old that nobody trusts.
The software works. This is an adoption failure, and adoption failure is a cultural problem.
Why do site teams resist new technology?
I’m not anti-technology. I’ve championed digital tools throughout my career, including a performance dashboard that gave leadership live visibility across a whole portfolio when the norm was a monthly PDF that was out of date before it was printed. But technology is only ever as good as the behaviour it changes. When adoption fails, there are usually three culprits, and none are on the vendor’s spec sheet.
The first is fear. Not the dramatic kind. The quiet, rational fear that comes from an industry where admitting a problem too early is career-limiting and being caught with one later is career-ending. So people learn to manage information rather than share it. I once asked a site manager why he wasn’t logging non-conformances. His answer was honest and devastating: “Because once it’s in there, everyone can see it, and I haven’t got a plan to fix it yet.” That’s not laziness. That’s transparency without psychological safety.
The second is complexity. Most platforms are built by people who love construction technology. They’re feature-rich and deeply configurable, and overwhelming for a site manager running six trade partners against a programme under pressure. If it takes longer to log something than to deal with it, people won’t log it.
The third is the absence of a feedback loop. If a site manager logs a delay and nothing happens, no response, no conversation, they won’t bother next time. Data only flows if the people entering it believe it matters, and it only matters if something visibly changes as a result.
And here’s the point leaders miss: the old guard is right to resist. The average site manager is a pragmatist. They don’t hate technology; they hate waste. If new software adds thirty minutes of admin a day and doesn’t get the concrete poured any faster, they shouldn’t use it.
Is going digital really 20% technology and 80% people?
That’s my rule of thumb, and it’s not far from BCG’s. Its long-held rule for AI is that leaders put “10% of their resources into algorithms, 20% into technology and data, and 70% into people and processes” (BCG, Where’s the Value in AI?, 2024). However you split it, most of the work is people.
If you spend £100,000 on a new project controls platform and £5,000 on training and changing how people work, you haven’t invested in change. You’ve bought an expensive way to fail.
Technology is a multiplier. If your site culture runs on transparency, collaboration and high standards, technology multiplies your success. If it runs on lip service, siloed information and firefighting, technology multiplies the confusion.
Adoption comes down to two things. Trust in the data: if the model doesn’t match the ground conditions, the site team will go back to paper in a heartbeat. Usefulness of the output: if a dashboard tells a project manager they’re behind programme but offers no route back, it’s just noise. More on the 20/80 split in boardroom data versus site reality.
What are the five steps to genuine site adoption?
After two decades of watching this play out, and being the person responsible for making it work across multi-project programmes, I’ve landed on a framework that changes behaviour. It isn’t complicated. It needs consistency, which turns out to be the rarest commodity in construction leadership.
Step 1: Measure five things, not twenty
Before you adopt any technology, be clear about what you’re measuring and why. Five KPIs, maximum. In my experience, the five that drive productivity are:
- programme performance (earned value against plan, not just a RAG status);
- commercial performance at package level, not just overall;
- new defects per week, not cumulative;
- workforce attendance and stability, especially in the specialist supply chain;
- safety leading indicators, such as near-miss reporting rates.
The platform is the vessel. The KPIs are the fuel.
Step 2: Run the Live Loop
The Live Loop asks two questions of every planned task: did it happen, and if not, why not? The single biggest cause of failure is fixed first, then the next, in a repeating cycle. Crucially, the person who logged the problem sees something change.
In one business, the time it took for a problem on site to reach the leadership team was 60 days. Within 18 months it was 3 days. Not through better software, but through measurement, escalation protocols and a weekly discipline that had everyone from site to board looking at the same data. When people see their input create a visible output, adoption goes up. When it disappears into a dashboard nobody checks, it goes down. The method in full: the Live Loop.
Step 3: Remove the them-and-us
The them-and-us between site and head office, and between contractor and trade partners, is one of the most persistent barriers to productivity in our industry. Head office mandates a reporting system. The site sees overhead with no benefit and complies minimally. The data is poor, so head office makes decisions on a false picture. Those decisions feel wrong on site, trust falls, and compliance falls further.
The way out is co-design. Involve the people who’ll use the system in how it works. Be honest about what the data will and won’t be used for. And show, repeatedly, that information shared will be used to solve problems, not to allocate blame.
Step 4: Simplify until it passes the Friday Afternoon Test
My test for any system: if a site manager on a difficult week, running late and dealing with a supplier crisis, can complete the required input in under five minutes on a Friday afternoon, it’ll get done. If it takes longer, needs more than three screens or asks for information they’d have to look up, it won’t. Remove the fields nobody reads. Cut report frequency where daily data isn’t used daily. Build for the phone first.
Step 5: Intervene without blame
Every high-performing job I’ve worked on had a clear, non-punitive answer to one question: when the data shows a problem, what happens next? Like a flight recorder, the record captures everything without judgement, and it’s only valuable if someone uses it to learn rather than to find someone to blame. A project manager who flags early pressure should get a problem-solving conversation within 48 hours, with an offer of help, not a board escalation two months later. More in why construction needs a black box.
Five steps to genuine site adoption
- 1Measure five thingsYou know it’s working when…Everyone on the job can name the five and their current numbers
- 2Run the Live LoopYou know it’s working when…The person who logged a problem sees a response within days
- 3Remove the them-and-usYou know it’s working when…Site teams suggest changes to the system without being asked
- 4SimplifyYou know it’s working when…The Friday input takes under five minutes
- 5Intervene without blameYou know it’s working when…Early warnings arrive earlier every month
The five steps in order, each with the test that tells you it’s working.
What mistakes do smart contractors keep making?
Confusing implementation with adoption. A platform is implemented when it’s live. It’s adopted when people use it consistently and honestly. Go-live dates and user counts are vanity metrics. Measure data quality over time, the ratio of issues logged to issues resolved, and how early the warnings arrive. Measure adoption before results: the ability arrives long before the habit.
Mandating without modelling. If the leadership team doesn’t use the platform, and the director’s reviews still run off a hand-built slide deck, the message on site is that the system is for other people. Compliance follows behaviour, and behaviour follows leadership.
Solving the wrong problem first. When productivity stalls, the instinct is to accelerate the programme. More often the programme problem is a symptom of something upstream: a broken supply chain relationship, a dispute gone underground, or a design gap nobody flagged because it didn’t feel safe to. Fixing the programme without fixing the information is painting over damp.
Stopping when it gets hard. Around months two and three of any real change, it feels like it isn’t working. The early energy has faded and the quiet resistance has found its voice. That’s when most programmes are quietly dropped, leaving an expensive platform and a workforce more sceptical of the next change. Consistency is the strategy. In my experience, real change can embed in 12 to 18 months if you get through the difficult middle.
What does good look like?
On one multi-project programme, leadership had almost no live visibility, and by the time issues surfaced in formal reports they were already expensive. We built a live dashboard, updated daily and usable on a phone. That was the easy part. The harder work was earning the trust that made people fill it in honestly.
We ran site visits, not inspections. We asked project managers what they needed from head office, not what head office needed from them. We redesigned the input around their answers, and built a 24-hour response commitment into escalation. Then, critically, we kept it.
Within three months, the quality and honesty of the data had changed completely. We were intervening in weeks two and three of a problem rather than months four and five. The site teams, sceptical at the start, became the platform’s strongest advocates. It was the same technology that had sat largely unused before. What changed was the culture around it.
The bottom line: construction won’t close its productivity gap by buying better software. It will close it by building the conditions in which software gets used. If that’s the work you need, it’s what we do.
FAQs
Why is construction productivity so low in the UK?
Output has grown more slowly than the hours worked: between 1997 and 2019, output rose 12% while hours rose 23% (ONS). Fragmented delivery, low adoption of the tools bought, and late, hidden problems all play a part.
Why do construction workers resist new technology?
Because new tools often add admin without helping the work, data can be used against them, and nothing visibly changes when they report a problem. Fix those and resistance falls.
What does going digital mean in construction?
Using data and digital tools to plan, build and hand over better. It succeeds when people use the tools consistently and honestly, which is mostly about culture, not software.
How do you get site teams to use new software?
Make it save them time on day one, keep inputs under five minutes, respond visibly to what they report, and have leaders use it themselves.
Sources
- ONS, Productivity in the construction industry, UK: 2021 (output +12%, hours +23%, 1997 to 2019).
- McKinsey Global Institute, Reinventing Construction: A Route to Higher Productivity, February 2017 ($1.6 trillion).
- BCG, Where’s the Value in AI?, October 2024 (10-20-70, described by BCG as its long-held view). The 20/80 split is the author’s rule of thumb.
- ONS, Output per hour worked, UK, August 2026 (2025 values; the 86% comparison is our calculation).
- 60 days to 3 days, the dashboard programme, the site manager’s answer, the five KPIs and the Friday Afternoon Test are the author’s own experience and method.
Andy Pritchard MCIOB
CIOB Gold Medal, Construction Manager of the Year 2019. Founder of Constructing Culture and author of Going for Gold. About Andy →