What Is a Risk Matrix?

A typical risk matrix sets out likelihood on one axis and severity on the other, each usually on a five-point scale from rare to frequent and from minor to catastrophic. Every hazard is placed at the intersection of its likelihood and its severity, and that single cell, not either dimension alone, is what the matrix actually reads out as a risk level.

Two hazards can share the same likelihood and still deserve very different responses, because a frequent but trivial event and a frequent but catastrophic one are not the same risk. Severity is read first precisely because a rare event with a severe outcome can still land in a cell that requires immediate action, while a frequent but minor one may not.

The mistake that undoes a risk matrix is treating its output as a final judgement rather than a tool for ordering priorities: a rare, severe risk, an engineering failure that happens once in years but kills when it does, deserves an engineering control even though its likelihood score alone looks low. The matrix orders the work; it does not excuse skipping a control just because the cell looks acceptable at a glance.

Different contractors on the same industrial site frequently use different risk matrices with different scoring conventions, which makes comparing a risk level from one contractor's assessment against another's harder than the shared vocabulary of “high” or “low” suggests.

• Risk assessment
• JSA
• HAZOP

Actual frequency data from the site improves how likelihood is estimated in the first place, rather than leaving it to judgement alone. Records from cameras already installed give a documented count of how often a given condition actually occurs. Read about the platform.