Quicker measures mean quicker results
For accurate statistical analysis, you typically require a sample size of 30 or above. If you were to use metrics later in the funnel, it would take a considerable amount of time to identify whether your safety measure worked or not. For example, if you were measuring safety using fatality data, it would take 1500 years to get 30 samples. However, if you were measuring risky behaviors, it would take just 90 minutes. The faster you can measure safety, the faster you can improve it.