Ask an architect where a project was decided and they will describe a review. Ask a site manager and they will describe a Thursday in February when a bracket did not fit and somebody made a call over the phone.
Both are true. Only one of them is in the record.
What actually gets decided on site
The formal position is that site visits inspect for conformity with the drawings. The practical position is that a site visit is where the drawings meet reality and one of them gives way.
The recurring categories are remarkably consistent across projects and building types:
- Tolerance stack-up. Three elements each within tolerance, adding to a gap that no detail covers. Usually at a junction between trades with different tolerances — steel at ±10mm meeting joinery at ±2mm.
- Substitution. A product is unavailable, and the alternative is dimensionally different in a way that matters two details away from where anyone is looking.
- Sequence damage. Something was installed correctly and then walked on, drilled through, or covered before inspection.
- The missing drawing. A condition nobody drew because it occurs once, and the person who has to build it decides how, at speed.
None of these appear in a design review, because in a design review the geometry is exact and the products exist.
The thirty-minute problem
Most site visits are too short and structured backwards. The visit starts at the site office, moves to whatever the contractor wants to show, and ends with a walk past the work in progress on the way out.
The practices that get the most out of a visit invert it. They arrive with a list of three things — usually the details they were least confident about at tender — and go to those first, before anyone can steer the conversation. Photographs are taken of the things nobody asked about. The site office comes last, when there is something specific to raise.
That is a discipline, not a talent, and it is the difference between an inspection and a tour.
Photograph what is about to disappear
The single highest-value activity on a construction site is photographing work that is about to be covered up.
Every service run before the ceiling goes in. Every membrane lap before the screed. Every fixing before the cladding. Every rebate before the plaster. These photographs cost nothing, take minutes, and are the only evidence anyone will have when a leak appears in year three and the question is whether the membrane was dressed up behind the upstand.
Practices that do this systematically — with a consistent naming convention and a location plan, not a phone gallery — find that their defects negotiations get shorter and their record drawings get better. It is also, quietly, the best construction education available to a young architect, because you cannot photograph a junction without understanding it.
From a site manager, quoted with permission “The architects I like are the ones who ask what I’m worried about. Most of them tell me what they’re worried about, which I already know, because it’s on the drawing.”
What to do with what you find
The failure mode after a good site visit is a report nobody acts on. Observations without a decision and an owner become a list that grows until handover.
The version that works is unglamorous: each item gets a decision, a person, and a date, and the ones that are design changes are separated from the ones that are workmanship. Conflating those two is how a defects list becomes a dispute, because a contractor asked to fix a design change for free will stop fixing anything.
There is also a category worth naming explicitly: things that are wrong and should be left alone. Every project has them. A junction built differently from the drawing that works better, a level that is out by 5mm and harmless, a finish that came out unexpectedly well. Marking these as accepted, in writing, saves someone six months of arguing about a non-problem.
Why it matters more than the review
A design review changes a drawing. A site visit changes a building — the actual one, the one that will stand there for sixty years with whatever bracket was available on the day.
The profession spends most of its self-examination on the front half of that process. The back half is where the buildings come from.
