The project profession in the UK
The scale of the profession.
Project work is a large and growing part of the UK economy. Pay rose in 2025 for the first time in four years. Yet the same research describes a workforce that feels short of skills, with a pay gap that is widening, so a lot of national value rides on work its own practitioners call under-resourced.
£186.8bn GVA
The project profession contributed £186.8 billion of Gross Value Added to the UK economy in 2024, employing 2.32 million full-time equivalents: 9.2% of total UK GVA and 8.5% of full-time employment.
UKAPM (2024), The Golden Thread, research by PwC. SourceTrade-body commissioned; built on ONS Labour Force Survey data, business surveys and interviews.
£52,500average pay
The average UK project manager salary rose just over 10% to £52,500 in 2025, its first increase in four years, against median full-time UK earnings of £37,430.
UKAPM (2025), Salary and Market Trends Survey; ONS for the national figure. SourceFieldwork late 2024. The report does not state its overall sample size. APM's figure is an average and the ONS comparison a median, so they are indicative rather than like-for-like.
35% pay gap
The gap between average pay for men and women in the profession widened from 24% to 35% between 2023 and 2025. Over the same period the share of female Project Directors fell from 28% to 18%.
UKAPM (2025), Salary and Market Trends Survey. SourceSample size not stated in the report.
54% work solo
More than half of UK SME project professionals, 54%, are the only project specialist in their organisation. In the same survey, 83% said their employer needs to improve project management skills across the workforce.
UKAPM (2024), survey of 500+ SME project professionals. SourceSelf-reported practitioner survey.
If that is the scale, the next question is how well the work gets delivered. Government is unusual here: it grades its own delivery in public.
UK government project delivery
The one client that grades its own work.
Government is the rare project client that publishes a delivery rating for its whole portfolio, alongside a Treasury study of its largest projects. Together they are the clearest primary evidence on this page that big projects are hard to deliver. The confidence ratings are in the panel at the top; here is what the Treasury found in the biggest, and what practitioners make of it.
Every UK mega-project overran.
HM Treasury's 2025 value-for-money study found that all of the UK's mega-projects since 1980 have run over budget, and most were delivered later than planned. The three current mega-projects, HS2, Sizewell C and Dreadnought, account for more than 10% of the government's entire capital budget.
UKHM Treasury / Office for Value for Money (2025), pp. 8, 34. Source
The gap between the first estimate and the latest cost
Three of the UK's largest recent projects, from first published estimate to latest figure. The size of the gap is a measure of how optimistic the original number was, the same problem the 1996 benefits study names later on this page.
| Project | Original | Latest | Change |
|---|---|---|---|
| London 2012 Olympics | £4.0bn | £9.3bn | +133% |
| Crossrail | £14.8bn | £18.9bn | +28% |
| HS2 | £25bn | £66bn | +164%* |
| Thames Tideway "super sewer" | ~£4.2bn | ~£5.0bn | +~19% |
UKSources: London 2012, Crossrail and HS2 from HM Treasury / OVfM (2025), pp. 6–7; Thames Tideway from Tideway and New Civil Engineer (2023–25). Sources
How the ratings changed between the last two reports
The government's delivery-confidence ratings from its two most recent published reports (the IPA's, and its successor NISTA's from 2025), on the same Red/Amber/Green scale, two years apart. No directly comparable figure was published for the year in between.
IPA · 227 projects
NISTA · 189 projects
UKComparison compiled by HotPMO from IPA (2024) and NISTA (2026) reports. Sources
55% not confident
Practitioners share the doubt. Asked about the UK government's ability to deliver projects over the next five years, 55% of project professionals said they were not confident, against 37% who were.
UKAPM (2025), Salary and Market Trends Survey, p. 22. SourceFirst year the survey asked this; sample size not stated.
Government is not unusual in this respect, and when you look more widely the same difficulty takes on a recognisable shape.
Project success and failure
Why delivery is hard, in numbers.
Cost overruns cluster by project type and grow with scale. On the more forgiving measure of self-reported success, only about half of projects clearly land well. And one project type carries a risk with no upper bound at all.
Average cost overrun, by project type
Real-terms average overruns. Road and rail come from the long-established per-mode study of projects completed between 1927 and 1998; the Olympics from a second study on the same basis, running to recent Games. (The 2025 dataset in the next card measures something different: tail risk across 23 project types.)
GlobalSources: Flyvbjerg, Holm & Buhl (2002), JAPA 68(3), for road and rail. Ansar, Flyvbjerg & Budzier (2026), Public Management Review, for the Olympics. Sources
453% IT tail overrun
IT sits in a category of its own. Across roughly 16,000 projects, IT was the only one of 23 project types the researchers found to have no statistical ceiling on its cost risk. The small share of IT projects whose costs land in the extreme tail overrun by an average of 453%. Reaching that tail is rare, but it gets its own line here because it is a different kind of risk from the steady averages above.
GlobalFlyvbjerg, Budzier, Aaen, Keil & Zottoli (2025), Project Management Journal, pp. 5–8. Source
73.8%
This is the most generous measure on the page: the average project performance rate in PMI's 2024 survey, the mean share of completed projects that met their business goals. Even so, fewer than half of project managers said their organisation offers any single supportive enabler such as mentoring.
GlobalPMI (2024), Pulse of the Profession, pp. 4, 21, n=2,588. SourceSelf-reported.
48–50% succeed
On a stricter measure, across PMI's large-sample research (n=5,751 in 2024), 48 to 50% of projects qualified as successful and 12 to 13% as outright failures. The rest sat in between. Projects with a clearly defined vision scored +41 on PMI's Net Project Success Score, against −18 for those without one.
GlobalPMI (2025), Step Up, pp. 4, 7. SourceThe success score is PMI's own index.
Where projects land, on a strict measure
PMI's large-sample split of project outcomes. About half clearly succeed, roughly an eighth clearly fail, and the rest sit in between.
GlobalSource: PMI (2025), Step Up. Success 48–50%, outright failure 12–13%; the remainder drew mixed views. The success score is PMI's own index. Source
36% on time
A separate practitioner survey found much the same. In its latest (2026) edition, only 36% of organisations mostly or always complete projects on time. Five years earlier the picture was similar: 34% on time, 34% on budget, and 36% delivering a project's full benefits.
GlobalWellingtone (2021, 2026), The State of Project Management, n=214 in 2021. SourceSelf-selected survey by a PMO consultancy.
If delivery is this hard, the useful question is what makes some organisations better at it. A few surveys have asked, and they lean the same way: the organisations that run a capable, established PMO tend to bring more of their projects in on budget and to handle complexity better than those without one.
PMOs
What the better performers have in common.
Several surveys associate PMO capability with better delivery, all self-reported and two vendor-published, so treat them as directional; none establishes cause. UK-only PMO data is scarce, so these lean on international surveys, and HotPMO sells PMO services, so weigh the section accordingly.
With a PMO, and without
Two self-reported measures from the same survey. The gaps are modest, but they point the same way.
GlobalSource: PMI (2026), Pulse of the Profession, p. 24. Self-reported; a modest gap. Source
69–82% have a PMO
Most organisations now run a PMO. Two independent surveys put prevalence at 69% (rising to 93% of large organisations) and 82%, with a quarter of those PMOs less than two years old. Most report high up: 82% to a VP or above.
GlobalPM Solutions (2025); Wellingtone (2021). SourcesTwo vendor-published surveys, cited together for triangulation.
72% lose a half-day+
Reporting is where much PMO time goes. 72% of project professionals spend half a day or more each month collating status reports, and 44% are dissatisfied with their PMO's reporting: a reminder that having a PMO and getting value from one are different things.
GlobalWellingtone (2026), The State of Project Management, 10th edition. SourceSelf-selected survey by a PMO consultancy; sample size not stated in the press release.
83 / 30% on budget
The best-rated fifth of PMOs deliver on budget 83% of the time; the weakest fifth, 30%. Each figure is a share of that group's own projects.
USPM Solutions (2025), The State of the PMO, pp. 5–6, n=134. SourceVendor-published, US-skewed. The groups are defined by rated performance, so some of the gap is close to circular. It shows association, and cannot show cause.
One failure mode is older than any of this, and it starts before a project does.
Benefits realisation
The problem that starts in the business case.
Some problems predate every current framework. The founding study of benefits management found that the numbers used to justify projects were often inflated from the outset, which means some overruns begin before delivery does.
47% admitted it
In the founding UK study of benefits management, 47% of the 200 largest UK companies surveyed openly admitted overstating the expected benefits of a project to win approval for an IT investment. The study is from 1996; we include it because it started the discipline, and because the modern figures echo it. Thirty years on, only about a third of organisations say they deliver a project's full benefits (36%, Wellingtone 2021), so the habit the study named has not gone away.
UKWard, Taylor & Bond (1996), European Journal of Information Systems 4(4). Paywalled. Source
Those failure modes are old and well documented, whereas the newest one is still being worked out.
AI and delivery
The newest variable, and the least settled.
AI is the fastest-moving topic here, so these are the numbers to read most carefully. Adoption among UK project professionals has climbed quickly and most report a benefit. On failure, the steadiest finding is a familiar one: the causes are mostly about leadership rather than technology. Several widely shared "AI statistics" did not hold up when we looked for a source, and those are in the next section.
36% → 70% of organisations
The share of organisations using AI, as reported by UK project professionals, rose from 36% to 70% between 2023 and 2025. A separate APM survey the same year found seven in ten professionals had brought AI into their everyday work.
UKAPM (2025), Censuswide surveys of ~1,000 UK project professionals. SourceThe 36% is APM's own two-years-earlier comparator, reported in the 2025 survey.
84%
When RAND interviewed AI and machine-learning practitioners about why AI projects fail, 84% named leadership-driven causes as the main reason, ahead of any technical factor. The widely quoted line that "more than 80% of AI projects fail" appears in the same report, but as a second-hand estimate that RAND quotes from other work.
USRAND (2024), Root Causes of Failure for AI Projects, pp. 1, 5, 65 interviews. Source
47% see an impact
In the first year APM asked, 47% of UK project professionals agreed AI is already having an impact on project management and their organisation. The single biggest impact they named was the automation of routine tasks (27%).
UKAPM (2025), Salary and Market Trends Survey, p. 20. SourceSample size not stated in the report.
~5% use it deeply
Adoption is wide but shallow. A joint KPMG and UT Austin study of 1.4 million real workplace AI interactions, from 2,597 users over eight months, found only around 5% consistently using AI in a sophisticated way.
USKPMG & UT Austin McCombs (2026). Source
The sections above cover different aspects of project delivery. Below, we see how the different perspectives converge.
What they agree on
Most delivery problems are set before delivery starts.
Taken one at a time, these are separate findings from separate fields. Read together, from a 1996 survey to 2024 research on AI, they keep pointing at the same place: the costly failures are shaped early, in the business case and the leadership around it, more than in the execution.
- The business case is optimistic by design. In 1996, 47% of large UK firms admitted inflating expected benefits to win approval.Benefits
- The pattern holds at the largest scale. Every UK mega-project since 1980 has run over its original estimate, and the Treasury's own study points to optimism at the approval stage as part of the reason.Government
- Even in AI, the causes are mostly human. 84% of the practitioners RAND interviewed blamed leadership before technology (65 interviews).AI
- And clarity at the start separates the winners. On PMI's index, projects with a clear vision score +41; those without, −18.Success
None of this proves cause, and the findings come from different fields. But they line up: the decisions that shape a project's outcome are made early, and by people, more often than at the technical coalface.
Numbers we left out
The figures we couldn't trace to a source.
We did not run an exhaustive investigation on every number in circulation. Where a widely shared figure would not resolve to a primary source in the time we gave it, we set it aside. The main ones are below, with what stopped us. If you can point us to the original, we will add it.
"72% believe AI will change their role" and "68% have had insufficient AI training"
Widely attributed to APM, but we could not find either figure at an APM primary source when we looked. The research page usually linked to is a small qualitative study that does not contain them. Left out until we can point to the original.
"80% of project management tasks will be eliminated by 2030"
A real Gartner line, but a 2019 forecast, and this page keeps to figures that have been counted. A prediction about the future sits outside its scope.
"Nine out of ten transport mega-projects run over budget"
Commonly quoted in that exact form, but the wording seems to have drifted from the original in retelling. We would rather cite the specific overrun studies above than a phrase we could not match to its source.
"Using ChatGPT reduces brain connectivity by up to 55%"
Circulated from one 2025 study, but we only found it in secondary write-ups and could not confirm the wording in the paper itself. Left out until we can check it against the source.
About these numbers
How this page is put together.
This page is maintained by hand. It is not an exhaustive audit. HotPMO sells PMO consultancy and a PMO assessment, so read the PMO section with the same eye you would bring to any vendor's figures.
Traced to a source
Each figure is taken from the original publisher's own document where we could reach it, with the year, the publisher and the stated method.
Tagged and caveated
Every statistic is tagged UK, Global or US. Where a source is self-commissioned, self-reported, gated or contested, that sits beside the figure.
Small studies flagged
A few small or single-industry studies appear where they are the best evidence available. Each is flagged so you can weight it yourself.
What we left out
Figures we could not trace are listed above, with the reason. Every number here keeps its original publisher's name and links to their source.
Citing this page
hotpmo.com/uk-project-management-statistics. Verified 18 July 2026.
Journalists, researchers and practitioners are welcome to cite this page. You may reuse individual statistics with their original citations as shown, and a link back helps us keep it maintained.
Changes: first edition, July 2026. Material updates will be logged here as sources publish new editions.
We maintain this page as a working reference.
HotPMO is a UK PMO consultancy. If you want a clear read on where your own delivery is losing time and money, a PMO health check is a structured way to start.
Read about the PMO health check →