Industrializing the Construction Process

Construction has been struggling for years with productivity flatlining at a low level, poor build quality, high cost, and uncontrollable time schedules. Building is seen as high risk, and the industry has been unable to produce good quality housing stock at the rate it is required. The output that hits the market is often of questionable value.

Ideas of volumetric modular construction, standardized kits of parts, and buildings prefabricated off-site have returned in various guises over the past hundred years yet not been able to change how we build, let alone revolutionized construction. While compelling in theory, the question remains why these concepts have been so unsuccessful.

The simplest answer is probably that most types of buildings simply aren’t suitable for being built using an approach to industrialization whose concepts are at best early Fordian: putting together simplified assemblies of standardized repetitive parts, unsuitable for reacting to the conditions of the site, brief, and market. But we have also found that the solutions are highly tuned to proprietary ideas, systems, and assembly methods, which makes them incompatible with the scale the construction industry requires.

Paradoxically, at the same time, significant parts of the industry are highly industrialized. Building envelopes are prefabricated and installed to the highest tolerances. Large scale MEP installations usually come in pre-assembled elements, ready to fix on-site.

Timber construction is inherently relying on being kits of parts, and the steelwork industry is almost entirely based on sophisticated shop finish and site assembly. But on-site it still takes forever to bring the parts together; site adjustments and last-minute solutions seem to be the norm.

Upon closer inspection, two things are tripping us up on every project, and both are closely linked. Designers— architects and engineers—draw up a design intent, entirely relying on that industrialized capacity of the specialist supply chain, leaving the detail to them. Each resolving what matters to their industrialized product, none of these specialists owns or resolves the connections to any adjacent trade. The necessary physical coordination happens late in the process, or not at all. It is easy to understand the importance of these interfaces and, therefore, not entirely logical why this isn’t addressed.

The answer here isn’t immediately obvious. But it all comes down to how we deal with risk. The construction process is riddled with it.

We need to resolve interfaces and design issues, develop the necessary level of detail before we start building

And we have developed highly refined methods to package it up and pass it on. From client to design team, from the design team to the main contractor, from the main contractor to the trades. Each on their own and the risk becomes concentrated at the point where players try to pass it on—at those interfaces, where it is not owned and not resolved.

And yet the solution is simple because that interface is simply a design challenge. As such, what we treat as a risk, is an unresolved piece of design, failure of thinking in trades, rather than in components. We look at the car industry and how pieces go together neatly on the production line.

It isn’t because it is all in a protected environment or because it is done a million times over. It is because the interfaces are also part of the design and not left to chance. The construction industry, by contrast, chooses, for instance, to leave the connection of the façade to the frame to chance—or resolve it last minute on every project—although we have done that a million times, too.

So, in response to the original question: “Why is modular not delivering?” the answer is: “Because we have been solving the wrong problem!”

We need to redefine risk, design out the unknowns, and engage the supply chain in doing so. The contractual vehicles for that are there – design-build, two-stage tenders, IPD– but we need to use them to resolve interfaces, resolve design issues, develop the necessary level of detail before we start building. That requires true collaboration rather than a false sense of gain out of competitive tender and old school ‘risk management.’ It must go hand in hand with integrated computational modeling to handle the complexity that comes from detailing the physical connections of the building components.

Industrializing the construction process needs to be the aim. Kits of parts, sub-assemblies, and industrialized components grow out of that.

Only then can we focus our attention on what matters: Designing our buildings for our users, for good performance, for a healthy environment, for longevity, and having the smallest impact on our planet.