A stock condition survey is unusual among inspections in that its output is rarely read as a document. It is consumed as data, aggregated across thousands of properties, and turned into capital budgets.
That places a requirement on it that most inspection types do not carry: comparability between properties surveyed years apart by people who never met.
01
Fix the taxonomy before the programme starts
If surveyors can describe components in their own words, the resulting dataset cannot be aggregated. The component list, the condition grading and the attribute set have to be defined in the template and closed to interpretation.
This is the single decision with the largest effect on whether the survey produces a usable plan, and it is very expensive to correct once fieldwork has begun.
- Close the component list
- Use a fixed condition scale with defined grades
- Record age and condition as separate values
- Photograph each graded component
02
Separate observed condition from remaining life
Condition is an observation. Remaining life is an estimate derived from condition, age and type. Recording them as one value destroys the ability to revisit the estimate when assumptions change.
Keeping them apart also lets the plan be re-run against new replacement assumptions without re-surveying the stock.
03
Keep the template stable across cycles
Every change to the component taxonomy breaks comparability with everything captured before it. Amendments should be batched and timed to a cycle boundary rather than applied continuously.
04
Sample deliberately if not surveying everything
Many programmes survey a proportion of stock and extrapolate. That is defensible when the sample is designed and recorded as a sample, and misleading when it is whatever properties happened to be accessible.
