UK construction profit margins: what contractors make, and who does make money

Constructing Culture long read. Updated 30 September 2026. Every figure sourced inline, with full sources at the end.

Construction is not expensive because contractors are getting rich. The UK’s 100 largest contractors averaged a pre-tax margin of just over 3% in their latest accounts. A quarter made less than 2%. Thirteen made a loss (The Construction Index Top 100, 2026).

That is a better year than most. It is still a business where one bad package can take the whole profit. If you want to know where the money in construction actually goes, and who ends up keeping it, the published accounts tell a very different story from the one most clients believe.

What profit margin do UK construction companies actually make?

Around 3% before tax for the largest contractors, and less for many of them. The Construction Index Top 100 averaged just over 3% in its 2026 edition, up from 2.4% a year earlier, when 43 of the 100 were below 2% and 16 were loss-making. Construction News’s CN100 puts its 2026 average at 3.4%, the first time it has passed 3%.

The listed contractors sit in the same band. The segment figures below are the companies’ own underlying or adjusted numbers for their UK building businesses.

Segment operating margins from each company’s results; Top 100 average from The Construction Index (pre-tax).

The warning signs sit in the same numbers. ISG’s last filed accounts, for 2022, showed a pre-tax margin of 0.53% on £2.2bn of revenue. It went into administration in September 2024. At 3%, there is not much room between a good year and a bad one.

Where does the money go on a £10m project?

Almost all of it goes through the contractor, not to the contractor. On a typical £10m commercial job, the breakdown looks like this.

Illustrative £10m building project, built from published benchmarks: preliminaries at 11% (Spon’s, via SCL Paper 230; Turner & Townsend’s global survey puts them at 10 to 11%); overheads and profit from a 2.5% rule of thumb (SCL Paper 230) to the 4.5% tendered in Walter Lilly v Mackay; the retained profit range from The Construction Index Top 100, 2026. The kept profit sits inside the overheads and profit line, not on top of it.

The profit line is the one people argue about. It is also the smallest line on the page.

Do subcontractors make more than main contractors?

Not in the published accounts. The belief that specialist trades earn 10 to 15% does not survive contact with their results. The Construction Index puts the average for its 20 largest specialists at 3.33% in 2026.

Company results announcements; Top 20 average from The Construction Index Top 100, 2026.

Higher risk does not buy a higher margin. M&E, often described as the highest-risk package on a job, is among the thinnest. Steel swings from a strong year to a loss in two. And specialists fail more than anyone: specialised construction firms were 56 to 58% of all construction insolvencies in England and Wales in each year from 2023 to 2025 (Insolvency Service).

One honest limit on this evidence: it covers the larger firms whose accounts get reported. Smaller trade businesses often pay the owner out of what would otherwise show as profit, and no one publishes reliable margin data for them.

The overhead gap: why the tender says 2.5% and the business needs 10

Here is the part most clients never see. The overheads and profit line in a building tender is small: a commonly used rule of thumb is 2.5% (SCL Paper 230, citing Spon’s). In Walter Lilly v Mackay, the contractor tendered 4.5%. Its own 2006 accounts showed overheads and profit running at 9.6% across the business.

The listed contractors show the same gap every year. Their administrative expenses, the cost of running the business behind the sites, run at around 6 to 8% of revenue: Galliford Try 6.0% (FY25), Kier 6.2% (FY26), Morgan Sindall 7.8% (FY25). After that, they keep around 2 to 4%.

So a contractor tendering 2.5% for head office and profit is actually relying on the packages and the delivery to carry closer to 10%. When delivery slips, that is the money that goes first.

So who does make money in UK construction?

The people who own something. Land, a product, a quarry, a kiln, or a large advisory business. The published margins are consistently several times what the contractors make.

Underlying or adjusted operating margins from each company’s latest full-year results. Arcadis is a global figure on net revenue; the rest are UK or UK-led businesses.

The pattern is clear. Housebuilders own the land and the product. Manufacturers own the plant and the brand. Large consultancies own the expertise and carry little capital. Contractors, specialists, merchants and hire companies sell labour, coordination, stock or equipment, and they sit at 2 to 4%.

Two fair caveats. Manufacturers and hire companies carry heavy capital costs, and their margins rise and fall with the housing and building cycle. And the figures above are operating margins, while the contractor averages are pre-tax, which flatters the comparison slightly in the others’ favour. Neither caveat closes the gap.

Contractors own the risk. Everyone else owns an asset.

How many weeks of delay wipe out a contractor’s margin?

Fewer than most directors think. We call it the margin clock: the number of weeks of unrecovered delay it takes for site costs to eat the entire profit on a job.

Our Cost of a Week model puts a contractor’s own burn on a £10m job at roughly £21,000 a week, in time-related preliminaries and head-office overhead that is no longer being recovered. At a 3% margin, the profit is £300,000. The clock runs out in about 14 weeks. At 2%, it runs out in under 10.

Constructing Culture Cost of a Week model, central case: contractor burn is time-related preliminaries plus unrecovered head-office overhead. Model outputs from stated assumptions, not a published benchmark.

Error works the same way. The Get It Right Initiative puts the direct cost of avoidable error at around 5% of project value, with the full cost between 10 and 25%. On a 3% margin, error alone costs more than the entire profit. Its top-ranked cause is inadequate planning.

These are model outputs from stated assumptions, not an industry benchmark, because no such benchmark exists. You can run your own numbers in the Cost of a Week calculator.

If margins are this thin, why do construction costs keep rising?

Because the cost is being driven from outside the contractor’s margin. Building materials prices rose 25.3% in the year to May 2022, and the government’s materials index reached its highest level on record in July 2026, up 5.9% on a year earlier (Department for Business and Trade).

Pay is not the driver right now. Regular pay for construction employees was flat in April to June 2026, down 0.1% on a year earlier, the slowest of any sector (ONS). Productivity is the long-running problem: between 1997 and 2019, construction output grew 12% in real terms while hours worked rose 23% (ONS).

Add risk being pushed down the supply chain, and the Building Safety Act’s new approvals regime, and the bill rises while the margin stays where it is.

What should clients and contractors do with this?

Clients: stop buying on lowest price. A tender that looks cheap has usually missed a risk, and on a 3% margin there is nowhere for that risk to go except back to you as a claim, a delay or a failed subcontractor. Unrealistic programmes cost money. Relationship-based procurement outperforms adversarial contracting because it keeps problems visible while they are still cheap.

Contractors: the margin is not won at tender. It is won or lost in delivery. The planning, the sequencing and the people who run the jobs decide whether the 3% survives.

Construction does not have a profit problem. It has a delivery problem, and the profit is where it shows.

Sources

1. The Construction Index, Top 100 construction companies 2026. www.theconstructionindex.co.uk
2. The Construction Index, Top 100 construction companies 2025. www.theconstructionindex.co.uk
3. Construction News, CN100 breaks 3% margin barrier, 22 September 2026. www.constructionnews.co.uk
4. Balfour Beatty, 2025 full-year results. www.balfourbeatty.com
5. Morgan Sindall Group, results for the year ended 31 December 2025. www.investegate.co.uk
6. Kier Group, FY26 preliminary announcement. www.kier.co.uk
7. Galliford Try, FY2025 statement and financials. www.gallifordtry.co.uk
8. TClarke, results for the year to March 2026. www.tclarke.com
9. NG Bailey, results for the year to February 2026. www.ngbailey.com
10. Severfield, results for the year ended 28 March 2026. www.tradingview.com
11. Billington Holdings, results for the year ended 31 December 2025. www.investegate.co.uk
12. ISG Ltd (company 10081578), group accounts for 2022, Companies House. find-and-update.company-information.service.gov.uk
13. Insolvency Service, Company Insolvency Statistics, August 2026, Table 1c. www.gov.uk
14. Society of Construction Law Paper 230, R. Champion, The Hudson Formula: Death by Footnote? (2021). www.scl.org.uk
15. Walter Lilly & Co Ltd v Mackay [2012] EWHC 1773 (TCC). caselaw.nationalarchives.gov.uk
16. Turner & Townsend, Global Construction Market Intelligence 2025. publications.turnerandtownsend.com
17. Company results: Berkeley, Genuit, Howden Joinery, Persimmon, AtkinsRéalis, Arcadis, Forterra, Taylor Wimpey, Ibstock, Barratt Redrow, Breedon, Marshalls, Travis Perkins, Speedy Hire (latest full-year announcements)
18. Get It Right Initiative, A Strategy for Change. getitright.uk.com
19. Department for Business and Trade, Construction building materials statistics, August 2026. www.gov.uk
20. Office for National Statistics, Average weekly earnings in Great Britain, August 2026. www.ons.gov.uk
21. Office for National Statistics, Productivity in the construction industry, UK: 2021. www.ons.gov.uk
22. Constructing Culture, Cost of a Week model (Research Library, Economics of a Week, June 2026). constructing-culture.co.uk/free-tools/

Margins are as reported by each company for its latest full year at 28 September 2026. Operating, underlying and pre-tax measures differ; each table says which it uses. Last checked 28 September 2026.