Share:

We choose certain actions to improve results. But the act of measuring these actions, in lieu of measuring the results, can cause problems like trivialising the actions.

Back in the days when I used to personally consult to my clients, I worked with a team in a global mining company, to help them measure their process of discovering new deposits. Mining companies take safety very seriously; even in their city offices. There was, without fail, a mandatory safety conversation at the start of all my meetings there. If memory serves me, it was as simple as one participant in the meeting sharing an observation related to safe work practices, like a pile of rubbish that hadn’t been cleaned up yet.

Workplace safety experts say that safety conversations is a good lead indicator for reducing safety incidents. Does that mean we should measure it?

What is the risk in measuring actions as lead indicators?

Many companies measure the number of safety conversations. It’s observable, it’s seemingly relevant, and it’s much easier to measure than some of the more intangible causes that threaten overall safety performance.

Kerry Mayes, one of my connections on LinkedIn, sees the same risk. He recently shared this perspective on the risk of measuring safety actions:

Another reason people fall into this confusion is when they are used to the really problematic areas like Safety KPIs. It is so hard to effectively measure Safety that people look for “leading indicators”, things like Safety Conversations. They measure the number of safety conversations in the workplace as there is research showing that safety improves if there are more safety conversations. Seems smart (though I always wonder about correlation / cause and effect) but it certainly adds to that confusion between actions and results: “Our goal is to increase safety conversations to 9 per week.”

With a target like that, can you see the risk?

What we measure is what we focus on.

If we are trying to hit the target of 9 safety conversations per week, that’s what we’ll focus on. The risk is, of course, that to hit the target, we trivialise the action to jumping through a hoop. We unwittingly start having random and irrelevant safety conversations, so we can have enough of them to add to the count and hit the target.

Isn’t the point of a safety conversation to raise awareness, discuss current risks, and decide if actions are needed to reduce the threat to safety performance?

Measuring only the amount of action can too easily end up having no impact on the real result we care about.

There are two parts to solve the problem of measuring only activity.

The first part of the solution to trivial activity measures is to make sure we have a measure of the ultimate result that the activity is supposed to help improve. For safety, we need to know which specific result safety conversations are supposed to help with. Maybe it’s the reduction of safety incidents overall.

The second part of the solution to trivial activity measures is to make sure we’re not focused on the amount of activity, but rather on the effectiveness of the activity. The direct result of safety conversations is not that they are always happening; it’s that people become more aware of the safety risks so they can avoid, reduce or remove them.

It is important to measure both of the ultimate result and the direct result of our actions. When we measure both these types of results, we can more closely tie an action to its outcome:

1. We want to make sure we’ve chosen the right actions that really will influence our performance outcomes, so we need to know if there is a positive correlation between the activity and the outcome.
2. We want to make sure that we are doing the actions right, in the intended way such that their influence isn’t undermined by complacency or by changing conditions that go unnoticed.

Measuring how much we’ve done an action is never a substitute for result or outcome measures.

It’s not that we did the actions, it’s what the actions changed. And what the actions changed is what we should measure. [tweet this]