News & insights

The Construction Site Is a Factory. We Just Don't Treat It Like One.

29 Jul 2026, by Vasilii Mishchenko

A construction site is a factory. We just pretend it isn't.

It takes in raw materials, transforms them, assembles them and delivers a finished product. The difference is that the manager of a factory, for example, a car plant, knows the cost of every minute of every operation. On a building site, the developer is guessing.

Manufacturing has that knowledge because the baseline is already in place. Supply chains are stable, internal logistics run smoothly, and teams are trained for long, continuous runs. A construction site has almost none of that. So much time goes into absorbing disruptions - design changes, access issues, late deliveries - that it never gets down to the finer detail. Nobody asks how long it takes to fix a particular component, what a minute of that worker's time actually costs or what changes if a robot does the same job in twenty seconds.

These are the types of questions manufacturers have been asking for a hundred years. Construction is only starting to ask them. The tools that could answer these questions already exist. The real gap is how the work is structured and understood.

Breaking the site into measurable units

Every manufacturing operation comes down to three types of action. Material gets handled, transformed into something usable, and joined to other materials - by cutting, welding, glueing or bolting. Every product, however complex, is a sequence of those steps.

Look at a construction site the same way, and the inefficiency becomes immediately visible. Most construction companies still assess their sites from the top down, using broad averages, because the infrastructure for more accurate data isn't in place. Without a consistent way to describe, measure and route each task across factory and site, they never reach the level of detail where real decisions get made.

Those decisions matter more than they first appear. Take fixing a component. If the process lasts five minutes at a given labour cost, and a robot does it in one minute at little more than depreciation cost, it changes the economics of the build. But the decision can only be made with the data in place and with reliable assumptions about how that one change affects everything around it. Most contractors don't have that data.

A catalogue of skills

Treating the process this way builds up a library of manufacturing skills: discrete, repeatable steps that map onto both the factory floor and the site. Cutting is a skill. Joining is a skill. Assembly is a skill. Once they're catalogued, you can ask straightforward questions. Which skills does this building need, which do we have and where is the gap? That analysis has traditionally taken teams of engineers weeks. Working from a structured library of skills and product descriptions, All3 can now do it in a fraction of the time.

The factory itself is built on the same principle. Each robotic production cell is an infrastructure unit built around power, safety and conveyance, with specific skills, tools and robotic arms slotting into it. One product range needs one set of skills, while another needs a different set. When a new design variant arrives, we add one capability to one cell, we don’t rebuild the factory.

The same logic on site

If the factory is tightly controlled and the site remains chaotic, the bottleneck has simply moved. The site is an assembly operation too, with its own skills in installation, sealing and connecting systems. Described the same way, they complete a single cost model that runs from raw material to finished building.

That model lets us ask questions construction rarely can. Which on-site operations take the most time? Which cost the most? What happens to the programme and the budget if we automate one of them? The answers guide real decisions about where money and effort should go. It also settles a question usually answered by habit. Should a given operation happen in the factory or on site? With the same cost model across both, that becomes a calculation - labour cost, timing, sequencing, transport risk - rather than a matter of what the foreman has always done.

Why it compounds

The thread running through all of this is consistency. Every operation, in the factory or on site, is described, measured and costed the same way, which means what we learn on one build carries directly into the next. Fragmented construction resets with every job. This approach compounds.

For a developer, that is the real shift. It is the difference between a cost that can only be estimated and one that can be known, and between a process that starts from scratch each time and one that gets faster, cheaper and more precise with every building it puts up.

Vasilii Mishchenko is Head of Manufacturing at All3.

Politicians Keep Promising Homes the Construction Industry Can’t Deliver. That’s Becoming a Political Trap

14 Jul 2026, by Rodion Shishkov

Insight: All3 co-founder and CEO Rodion Shishkov explains why Europe's housing crisis is fundamentally a production problem, and why the politicians promising to solve it keep failing.

The Constructivist Episode 3: Dr Simon Kempf on why housing projects stall before construction starts

01 Jul 2026, by All3

Video: Episode 3 of The Constructivist with Dr Simon Kempf, CEO of Periskop AG, one of Germany's leading real estate asset management companies.

The Constructivist Episode 2: Markus Weigold on the productivity challenge in housing

16 Jun 2026, by All3

Video: Episode 2 of The Constructivist with Markus Weigold, partner responsible for the Berlin–Brandenburg region at real estate investment specialist, Drees & Sommer.