The Supply Chain Already Exists

Not the bricks. They came from a factory, palletised and shrink-wrapped, on a lorry. Not the roof: the trusses arrived pre-cut, nailed and plated in a shed somewhere off the A-road network. Not the windows, the doors, the staircase, the boiler, the kitchen, the insulation, the plasterboard, the felt, the fixings. All of it manufactured, all of it delivered. What gets made on site is the assembly: several thousand factory products brought together outdoors, in weather, to tolerances no factory would accept, on a programme that exists mainly as a shared act of optimism.

A new house is a collection of manufactured products assembled in a field. The most handmade thing on a modern housing site is the programme.

This is worth holding on to, because the industry's long argument about modern methods of construction is usually staged as a choice: building versus manufacturing, craft versus factory, the old way versus the leap. The staging is wrong. The choice was made decades ago, product by product, and manufacturing won almost every round. The live questions are narrower and more interesting: where should the assembly happen, who decides, and when.

There is a published instrument for the first question, and it is not a door you walk through but a dial you set. In March 2019 MHCLG issued the definition framework that gave modern methods of construction its seven categories. Read in order, they run from the most pre-manufactured to the least, and then past it. At one end, Category 1, whole three-dimensional modules. Then two-dimensional panelised systems. Then structural components. Then additive manufacturing. Then the non-structural assemblies, the bathroom pods and service cassettes. And at the far end, Categories 6 and 7, which are not pre-manufacturing at all but site-based approaches: better products and better process that reduce the labour a site consumes. A scheme can sit anywhere on that dial, and different parts of one building can sit at different points. Since 2025 there has been a British Standard for how to run the decision: PAS 8700 is deliberately material and technology agnostic, a specification for process rather than product. The sector's own standards body, asked what good looks like, declined to name a winning technology. It described a sequence.

The measure that goes with the dial is pre-manufactured value: the share of a project's gross construction cost that arises through pre-manufacturing. Homes England uses 55 per cent as the qualifying threshold for pre-manufactured value in its funding framework, as one limb of an either/or test alongside Categories 1 and 2. Most of the argument about that number misses what it is for.

It is not a loyalty test for factories. It is a way of saying, in one figure, where a project has decided to add its value, so that the decision can be examined.

Now turn the telescope around, because the interesting view is from the factory's side. What does the manufacturing half of the supply chain actually need from a project? Across the principal guidance and industry reviews of the last decade, three recurring requirements emerge.

It needs the decision made in time. The DfMA Overlay to the RIBA Plan of Work puts it in one sentence: there is a finite window of opportunity for deciding on DfMA processes that lead to offsite solutions, after which their benefits quickly recede. The window has published edges: a Category 1 solution needs the commitment by Stage 2, Category 6 holds open to Stage 4, and every category needs a design freeze before manufacturing begins.

It needs demand it can bank. The Farmer Review said this plainly in 2016: the higher the sunk cost in machinery and equipment, the greater the pressure to drive volume against it, in a market whose demand is cyclical. Its recommendations asked government to stimulate demand for pre-manufactured housing and to publish a pipeline of housing demand along the lines of the National Infrastructure Pipeline. The recent history of firms that bet factories against pipelines that never firmed up is well enough known not to need retelling here.

And it needs information it can build from. The Construction Playbook made early supply chain involvement policy and was specific about depth: formally engaging the tier 1 contractor alongside tier 2 and 3 suppliers in the pre-construction phase. ISO 19650 exists so that the information handed over is something a production line can trust.

A decision inside the window. Demand that can be banked. Information fit to build from. None of the three is a technology. All three are decisions, and each one is made, or missed, in the months before a factory ever sees the job. Which is the quiet finding underneath the whole MMC debate: the supply chain is not waiting to be invented. It exists, it is broad, and it responds to exactly the disciplines the industry's own guidance has already specified. What it cannot do is survive being asked to price late, against unfrozen design, with no committed volume, and then absorb the blame for the premium.

Policy has started to notice. The Social and Affordable Homes Programme for 2026 to 2036, at least £27.3 billion of it outside London, carries no equivalent of its predecessor's 25 per cent MMC requirement and asks bidders instead to state their MMC numbers, against Categories 1 or 2 or a pre-manufactured value of at least 55 per cent. Compulsion out, disclosure in. A bidder now has to know their own number and stand behind it, which moves the dial-setting decision from the compliance file to the bid strategy, where it always belonged.