Do you still use risk matrices?

Here’s our advice.

Are you still using risk matrices?

We’re not – here’s why.

In 2019, we removed risk matrices from our risk assessments after discovering they can do more harm than good. As you probably know, they are not mandatory, yet many professionals continue to use them regardless. However, after an ex-HSE inspector raised the subject with us, we took time to explore the benefits and drawbacks, to make sure we were using best practices. Our research helped us conclude that in many instances it makes sense not to use them.

The official line

Just to be clear, before we take you through our key reasons for not using risk matrices anymore, let’s talk about the law. There’s no mention of risk matrices in any health and safety legislation, the HSE’s risk assessment guidance, or HSE document indg163: www.hse.gov.uk/pubns/indg163.pdf. Therefore, you are not obligated to use risk matrices by the regulators.

Our findings:

After a thorough examination of the subject, these are the key reasons that we’ve decided to abandon the use of risk matrices.

  1. Risk matrixes do not usually take into account the findings of psychological research concerning the cognitive biases that impair most people’s ability to assess risk.
  2. The verbal labels used in ordinal scales (e.g. high/medium/low likelihood) are interpreted inconsistently among different users and by the same user. For example, one person’s perception of likelihood is different from another’s, based on personal experience or their level of competence.
  3. A multiplicative model, for example, will assess the likelihood and impact of several different risk activities as if they were entirely independent. When the scores are aggregated, three different activities may be considered to be medium or low risks. However, if the activities are correlated and have an impact on one another, three low-to-medium risks could together produce one very high risk.
  4. Hazards of low severity and high likelihood will receive the same risk value as hazards with high severity and low likelihood. Although the risk values may be the same, the response to these different hazards in terms of priority for correction may be very different (source: St John Holt, 1999).

Taken together, these four findings indicate that the scoring methods of risk matrices are likely to be poor tools for risk assessment.

A potential exception

Using a matrix may still be worthwhile for more complex situations, or those where a score might help provide context to stakeholders during discussions around a particular issue. However, it does require experience and expertise to judge the severity and likelihood accurately. Getting risk scores wrong can result in failing to implement important control measures, or implementing the wrong control measures

Moving forward

We advocate removing risk matrices for the most part. Instead, we advise concentrating your efforts on assessing what the risks to your event or activity are, as well as considering the control measures that are likely to reduce that risk.

If you couple this with being clear about who is responsible for actioning the control measures, what instructions or information persons need to receive to control the hazards effectively, and how you are recording what you are doing, you are likely to achieve a better overall outcome.