Running a snagging process that actually closes out

Snagging is the part of a project where good intentions meet a spreadsheet with four hundred rows, and the spreadsheet usually wins.

The problem is rarely that defects are not being found. It is that the list is not structured in a way that lets anyone close it out or learn from it.

Locate defects, do not describe them

“Chipped tile in corridor near lift lobby” requires whoever rectifies it to find it first, and requires whoever verifies it to find it again. On a large project that search cost, repeated hundreds of times, is significant.

Defects should be pinned to a location — level, room, zone — against a plan, with a photograph taken at the point of inspection. It is not a small change; it removes an entire category of wasted time and makes the data usable.

Distinguish snags from non-conformances

A snag is workmanship not yet to the required standard, normally identified at completion. A non-conformance is work that does not comply with the specification, often structural or systemic, and carries a formal process.

Mixing them produces a list where a scuffed skirting sits beside a failed pressure test. They need different routes, different approvers and different closeout evidence.

Closeout needs evidence, not a tick

A defect marked closed without evidence is a defect that will be raised again at handover, usually by the client, usually at the worst moment.

Require a photograph and, where relevant, a test result or an inspection sign-off against the original item. The extra thirty seconds at closeout removes the second and third inspection rounds.

Find the pattern before it repeats

This is where most defect data is wasted. A register showing four hundred closed items tells you very little. The same data showing that sixty per cent sit in one zone, on one trade, in a six-week window tells you what to do on Monday.

On repetitive work — residential towers especially — a detail that is wrong in unit twelve is wrong in every unit until somebody notices. A heatmap by floor and trade surfaces that after a handful of inspections rather than after four hundred.

Our QA/QC module maps inspections and NCRs to a 2D schematic for exactly this reason, linking each one to the drawing revision and specification clause it was checked against.

Link inspections to what they were checked against

An inspection record that does not state which drawing revision and specification clause it was carried out against is weak evidence. If the drawing changed after the inspection, or the inspection was carried out against a superseded revision, that needs to be visible.

This is the join between quality and document control, and it is the join most projects leave to memory.

Handover starts long before handover

• Agree the snagging standard and the closeout evidence required at the start, not at the end.
• Run trial handovers on early units so the process is tested while there is time to change it.
• Assemble certificates, test results and warranties as areas complete, rather than in the final fortnight.
• Track outstanding items by area, so phased release is realistic rather than aspirational.

The measurement that matters

Not the number of open defects — that number goes up when inspection improves, which punishes the right behaviour. Track repeat defects by trade and average time to verified closeout instead. Both reward fixing causes rather than closing rows.