The Spectrum of Prefabrication

The single most useful idea in off-site construction is that prefabrication is not one thing you either do or do not do, but a ladder of degrees - a spectrum running from the smallest factory-made part all the way to a complete building delivered as a box. Place a building anywhere on that ladder and you have made a strategic decision about how much of its making happens in a factory rather than on the open site. Learn the rungs and you have a language for the whole field.

At the bottom sit components - the single factory-made parts that almost every modern building already contains: a roof truss jigged up in a yard, a pre-hung door, a window unit, a precast lintel, cut-to-size joinery, a ready-made reinforcement cage. These are so ordinary we rarely call them prefab, yet they are exactly that - parts made before they reach their position. One step up are sub-assemblies: several components combined into a larger unit in the factory - a pre-glazed curtain-wall cassette, a plant skid with pumps and pipework mounted and tested, a riser cartridge of bundled services. A sub-assembly is still a *part* of a building, not a piece of enclosure you could stand up by itself.

Climb further and you reach 2D panels - whole walls, floors and roofs made flat in the factory and stood up or laid down on site (panelised construction). A panel is a *plane* of building: it encloses, but it arrives flat and is assembled into three dimensions on site. Above that is 3D volumetric - entire three-dimensional chunks of building, whole finished rooms or groups of rooms built as boxes in the factory, complete to varying degrees with structure, services, finishes and fittings, then stacked and connected on site. Beyond single modules sits the whole building delivered as one or a very few volumetric units - a cabin, a kiosk, a small clinic - craned into place almost complete.

Running alongside the pure rungs is the hybrid: the real-world mixture that combines degrees on one project - panelised walls on a site-built concrete frame, a volumetric bathroom pod dropped into an otherwise panelised flat, a modular superstructure on an in-situ podium. Hybrids are not a compromise or a failure to commit; they are how competent projects actually use the spectrum, concentrating off-site where repetition pays and leaving site-built where adaptability is needed.

Within panelisation sits a distinction that matters more than its modest name suggests, because it is the clearest small example of the spectrum's central logic: the difference between an open panel and a closed panel. Both are flat 2D elements made in a factory, but they differ in how much is finished before the panel leaves the line.

An open panel is essentially a structural frame - timber studs or light-gauge steel - sheathed on one or both faces and delivered as a skeleton. The insulation, services, internal linings, windows and finishes are installed later, on site, after the panels are erected. The panel gives you a fast, accurate, factory-made structure, but much of the fit-out work - and much of the weather exposure and trade sequencing - still happens on site. A closed panel carries far more: insulation already in the cavity, a breather membrane and sometimes the cladding fitted, windows and doors installed, services routed, and the internal lining boarded - all completed in the controlled factory before the panel is wrapped and delivered. A closed panel arrives as a near-complete piece of wall, and the site work is reduced to lifting, positioning and connecting.

The closed panel is further along the spectrum than the open panel, even though both are '2D panels', and the consequences are exactly the spectrum's consequences in miniature. The closed panel delivers more factory value - more of the work done in controlled conditions, more quality held at the bench, less site labour and weather exposure, a faster watertight shell. But it costs more to make, is heavier and more fragile to transport and lift, is harder to inspect once closed (the services and insulation are buried behind the lining), and it freezes more decisions earlier - the window positions, the service routes, the insulation build-up are all fixed when the panel is made, not negotiable on site. The open panel keeps more flexibility and is lighter and cheaper to move, at the cost of more site work and less factory quality.

This open-versus-closed choice is a real design decision on real projects, and it rehearses the judgement the whole spectrum demands: how much to finish off-site. Push everything into the closed panel and you maximise factory advantage but commit early and carry weight; keep the panel open and you stay flexible but give much of the work back to the site. There is no universally right answer - only one that fits the project's repetition, programme, logistics and appetite for early certainty.

Step back from the individual rungs and a single trade-off governs the whole ladder: the more you finish off-site, the more you gain in factory value - and the more you pay in weight, transport and early commitment. Understanding this trade is the core of choosing where to sit on the spectrum.

On the gain side, moving further off-site means more of the building is made in controlled conditions, so more quality is held at the bench and inspected before it is concealed; more work is done under cover, unaffected by weather; more is produced by repetitive, jig-assisted operations that are faster and less reliant on scarce site trades; and more of the programme can run in parallel with site groundwork. A closed panel or a finished volumetric module captures far more of these advantages than a bare component does. In principle, the further right you go, the more of the factory's virtues you collect.

On the cost side, the same rightward move makes every unit bigger, heavier and more three-dimensional, and that has hard consequences. A finished 3D module must fit on a truck, clear bridges and gates, and be lifted by a crane of sufficient capacity - so its size and weight are capped by transport and lifting limits, not by the design alone. It is more fragile to move and handle, and transporting a finished room means transporting mostly *air* (a room is largely empty volume), which is expensive compared with shipping flat panels or loose components. And crucially, the more finished the unit, the earlier and harder the design must be frozen: a closed module locks in its structure, services, finishes and dimensions when it is built, so late change is slow, costly or impossible.

So the spectrum is genuinely a trade, not a simple 'further is better'. Components are cheap to move and keep the design open but deliver little factory advantage; whole volumetric buildings deliver the most factory advantage but cost the most to move and demand the earliest, firmest commitment. The sweet spot for a given project depends on its repetition, its programme, the transport routes to its site, and how early its design can realistically be settled - which is why this is a strategic and design decision, not a default. And because those variables differ by element, the right answer is rarely a single point on the ladder for the whole building.

The last and most practical idea is that these rungs are degrees a single project mixes, not exclusive camps you must choose between - and deciding how far along the spectrum to go, element by element, is one of the central strategic and design decisions in off-site construction. A competent project does not ask 'are we modular or not?'; it asks, for each part of the building, 'how far off-site does this element deserve to go?' - and answers differently for different parts.

The logic that guides the mix is repetition and certainty. Concentrate the furthest-off-site effort where the building repeats and where the design can be settled early: the many identical hotel rooms become volumetric modules or closed pods; the repeated external wall becomes a closed panel; the standard riser becomes a service cartridge. Leave nearer the site-built end the parts that are bespoke, one-off, or likely to change - the ground-floor entrance, the irregular roof, the interface with an existing building, the tricky site boundary. This is why so many real schemes are hybrids: a volumetric superstructure on an in-situ concrete podium; panelised walls hung on a steel frame; factory pods inside a conventionally-built shell. The hybrid is not indecision; it is the spectrum used well.

For the designer this reframes the task. Choosing a position on the spectrum is not a late procurement tick-box; it is an early design move that shapes the grid, the module or panel sizes (set by transport and craneage), the interfaces between off-site and site-built work, and how much must be frozen and when. Go further right and you gain factory value but must commit earlier and design within tighter transport limits; stay left and you keep flexibility but hand more work back to the variable site. The skill is to read the building, find where repetition and early certainty live, push those parts off-site while keeping the bespoke and the uncertain nearer the site, and then design the interfaces between the two with care.

This also dissolves a common anxiety - that choosing prefab flattens architecture into identical boxes. It does not, because you are not forced to a single rung: you can take the repetitive accommodation far off-site for its efficiency while lavishing bespoke, site-built attention on the entrance, the roofline and the public face. Module 2 opens each rung in depth; the binding structural, fire, acoustic, transport and connection engineering stays, as always, with the qualified specialists, the manufacturer's tested system and the governing codes.