The snagging list we did not need

Jebel Vista handed over with nothing on the defects list. That was not luck, and it was not unusually good workmanship. It was four mock-ups and a bad-tempered meeting in month three.

The argument

The meeting was about a line. On the drawings, the travertine floor met the oak joinery in the main living space at a single pencil stroke — no threshold, no reveal, no expressed junction. Two materials, two trades, two rates of movement, and one line.

The stone subcontractor's position was that they set out and lay to the joinery. The joiner's position was that they scribe to finished stone. Both are standard positions. Both are reasonable. Together they mean nobody has designed the junction and each has assumed the other will resolve it on site, which is how you get a 4 mm gap that opens to 9 mm at one end and a strip of silicone that was never on any drawing.

The meeting was unpleasant because we made it be about who owned the detail rather than who would fix it later. It ran long, two people left annoyed, and it saved us the entire snagging list.

Four mock-ups

What came out of it was a decision to build the difficult junctions before building the building:

• The stone-to-joinery threshold, full width, in the real materials.
• The shadow-gap skirting return where it meets a door lining — the detail everyone draws and nobody has built.
• The ceiling-to-glazing pocket, viewed with the blind both up and down, because it is a different detail in each state.
• The external soffit junction where the slab meets the fascia and the waterproofing turns.

Each was built on site, looked at in daylight, adjusted, and signed by the architect, the client and both trades. Two of the four changed materially between first build and signature. That is two details that would otherwise have been resolved by a fitter, alone, in month fifteen.

Snagging is a design failure, not a workmanship failure

This is the part we would argue for hardest. Go through a typical defects list and very little of it is genuinely bad work. It is overwhelmingly junctions: where two materials meet, where two trades meet, where a drawing stopped short. The fitter who filled that gap did the only thing available to them at the time, with the next trade already waiting.

Treating snagging as a quality-control problem puts the effort at the end, where it is expensive, adversarial and too late. Treating it as a design coordination problem puts the effort in month three, where it is cheap.

What we do now

1. Every trade-to-trade junction has a named owner in the subcontract. Not implied by scope — named.
2. Four mock-ups minimum on any project with bespoke joinery or stone, built on site in the real materials, under the real light.
3. The month-three meeting is scheduled. It is in the programme, it has an agenda, and its purpose is to find the junctions nobody has designed while there is still time to design them.
4. Sign-off is a photograph and a signature against each mock-up, held with the drawings, so the standard travels to whoever is actually on site in month fifteen.