A Test Is Not a Truth

There is a persistent misunderstanding in fire safety that continues to surface in projects, reports, and, increasingly, disputes.

A fire test result is often treated as a definitive statement of performance. It is not.

Standards such as BS 6336 were clear on this point decades ago. Fire tests are tools. They are controlled, simplified representations of specific conditions, designed to generate comparable data. They are not predictions of how a product, system, or building will behave in fire.

Yet, time and again, we see test evidence being stretched beyond its limits.

A product achieves a classification under a defined test configuration and is then assumed to perform similarly in materially different conditions. Interfaces change. Fixings differ. Assemblies are incomplete. And still, a degree of assurance is inferred that simply is not evidenced.

This is not a marginal issue. It sits at the centre of many of the systemic failures identified in post-Grenfell investigations. Misinterpretation of test data, selective use of evidence, and over-reliance on classification without understanding the underlying conditions have all contributed to unsafe outcomes.

The problem is not the test itself. The problem is how it is used.

Fire testing is inherently reductionist. It isolates variables. It simplifies geometry. It standardises exposure. That is its strength. But it also means that application requires judgement. The question is never just “what was the result?”, but “under what conditions, and how does that relate to the real building?”

This is where competence becomes critical.

A fire engineer must be able to interrogate test evidence. Not simply accept it. That means understanding the test method, the boundary conditions, the failure criteria, and the extent to which the result can, or cannot, be applied to the design in question.

It also means being prepared to say “this does not demonstrate what is being claimed”.

In practice, this is where tensions arise. Commercial pressure, programme constraints, and the desire for simple answers all push toward over-interpretation. Test data becomes a proxy for assurance. Sometimes, it is presented as such.

This is precisely where professional judgement must intervene.

The industry does not need more test data. It needs better use of the data we already have.

And perhaps a simple reframing would help. A fire test is not proof of safety. It is one piece of evidence in a wider assessment.

Treating it as anything more is where problems begin.

This insight reflects on the use and interpretation of fire test evidence in practice. It is not intended to provide design guidance, but to highlight recurring issues observed across projects and investigations.