The £256 Pound

That is the return reported on one pound, between 2023 and 2026, across twenty-five live UK construction projects, when it was spent teaching people to stop errors before the errors were built. The Construction Industry Training Board funded the training. The Get It Right Initiative designed it. Kier, BAM Nuttall, VolkerStevin and Taylor Woodrow ran it across £942.5 million of work and reported what happened next: an estimated £92.6 million of error cost avoided. Call it roughly ten per cent of the project value, represented by estimated error cost avoided, identified through a training budget of roughly a third of a million pounds.

The anatomy of the number matters more than its size. Of the £92.6 million, some £71.2 million, over seventy-six per cent, came from one source: eliminating the errors caused by inadequate planning. Not workmanship. Not weather. Not the subcontractor everyone likes to blame in the wash-up meeting. Planning: the work done, or not done, before anyone was standing on anything.

Construction errors are not made on site. They are designed, priced and programmed in advance, and installed later.

If that distribution sounds surprising, it shouldn't, because the same organisation published the same finding a decade earlier and hardly anyone in housing read it then either. In 2016 GIRI ranked the root causes of construction error by financial significance. First: inadequate planning, from task level through to project level. Second: late design changes. Third: poorly communicated design information. The causes people instinctively reach for, poor interface management, poor communication on site, inadequate supervision, come eighth, ninth and tenth. The full ranking is printed below, and it repays a slow read, because its top half describes work done in offices, months before the site exists.

Numbers this round invite suspicion, so here is everything that qualifies this one. The £92.6 million is the participants' own estimate of what the training helped them avoid, not an audited measurement. The four contractors build infrastructure, not homes. The gain is attributed to a training intervention, on projects that chose to take part. The £256.48 is a programme return calculated against reported avoided cost, not an audited economic return. CITB has since published the final evaluation report, which carries that figure, the £361,306 of funding and the £92.6 million directly, so the primary document is now on the record rather than the trade press alone; the £71.2 million remains GIRI's own figure, reported in the trade press and corroborated in the evaluation only as a proportion, at over seventy-six per cent. Take the return with any amount of salt. What survives the salt is the shape: a decade of ranked evidence and a three-year measured pilot, from different methods and different years, pointing at the same place. The money leaks at the front.

Scale matters here too. The Farmer Review listed, among the ten symptoms of its 2016 diagnosis, low margins, adversarial pricing and financial fragility. Set a leak worth ten per cent of project value against margins that thin and error stops being an efficiency topic; it becomes the difference between a business and a casualty, several times over, on every scheme. What the pilot evidences, bounded as above, is that structured error-reduction training identified and avoided substantial potential error cost, with inadequate planning the largest reported category by value. What it does not evidence is a controlled estimate of the error that would otherwise have occurred. The reading it supports is that the leak responds to something as unexotic as planning done properly, early, by people trained to do it. Few industries are offered that trade by their own training board. Fewer still have spent sixty years declining it.

Which would be merely interesting if the front end of a project were unmanageable, a zone of fate. It is the opposite. It is the one part of a scheme where every option is still open and every change is still cheap, and the industry now has more published equipment for managing it than at any point in its history. Since 2019 there has been a shared language: the seven MMC categories, so that a delivery strategy can be named without a meeting about vocabulary. Since then, a shared measure: pre-manufactured value, with Homes England's 55 per cent as the qualifying threshold in its funding framework. Since 2025, a shared process standard: PAS 8700, which specifies how manufacturing-led delivery should be sequenced and evidenced, and pointedly declines to prefer a technology. Language, measure, process. What GIRI priced is the cost of front-end failure that those disciplines are designed to reduce.

Policy is now betting the same way. The Social and Affordable Homes Programme, at least £27.3 billion of it outside London, now asks bidders to state their MMC numbers rather than requiring a proportion of them. The supply-chain reading of that shift belongs to the previous Text; the bid-strategy reading belongs here. Reading it as retreat is the common take and probably the wrong one. A mandate buys adoption without conviction. A disclosure regime asks a bidder to know their number and defend it, which quietly converts front-end discipline from a compliance line into a competitive one. The bids that can show their reasoning will beat the bids that can only show their enthusiasm.

The finding is sixty years old. The receipt is new. The cheapest pound on the next scheme is the one spent before the site exists, and there is now a published number for what it buys. And what it buys was never really the ten per cent. It is the distance closed between what was promised and what is handed over: a home built once, properly, for the person waiting to live in it.