The constraint is the field itself.
Productivity stagnation has four named causes, all of them about how work happens on site rather than what is being built. Each one is, in software terms, a tractable problem.
Variability.
Every job is a prototype. Sites differ in size, slope, soil, and utilities, across roughly 20,000 permitting jurisdictions. Volumes are tiny by manufacturing standards: about a million single-family homes a year against fifteen million cars and a hundred million smartphones.
Tacit knowledge.
Expertise lives in a few heads. The trade runs on informal standards, common practice, and worker expertise. That makes every good new method expensive and risky to introduce, because it demands extra design work and training upfront.
Coordination.
Loosely joined subcontractors work from sparse drawings that never fully define the building. Coordination breaks easily, and the breaks become delay and dispute.
Rework.
The hidden productivity tax. Defects found at final inspection mean tearing out finished work. Rework is the quiet line item that eats the gains every other improvement tries to create.
Construction is only about 4.5% of GDP, but it sits under the single largest line item in most family budgets.
The biggest cost inside the biggest expense of most Americans runs straight through field-labor productivity.
This is a software-shaped problem hiding inside a hard-hat industry.
Variability looks like prototype-by-prototype scoping. Tacit knowledge looks like the absence of a system that captures and re-serves what the senior crew already knows. Coordination looks like incomplete drawings and trades that never share state. Rework looks like the quality gate landing at the wrong end of the build. All four are routable with the right field-resident software.