Hi,

We have a new managing director at work and they are wanting to set a monthly KPI for near misses, this seems fairly normal however I'm unsure of what to set the number at. Reviewing the near misses we have I would say almost 40% aren't actually near misses and instead would be classed as facilities issues i.e. lights not working.

We are small manufacturing company with mostly office based staff and I would rather tackle 5 important near misses than set it at something like 15 and struggle to hit it or just put nonsense fillers in. 

Thanks for any advice. 

The whole KPI thing is a very unecessary distraction for accidents / near misses / RA creation / RA review etc.

In a business the key performance is to be making (legal) money for the shareholders or owner.

Health & safety is a cost coming off the bottom line as initiatives, equipment, training, salaries and resources - at best this is off-set activity to negate fines and prosecutions.

Do you really want to be pushing what is likley limited resource in to capturing, collating, investigating and reporting "near miss"?

Personally I would promote active engagement in resolving situations considered near miss. 

Its like chasing the dream of zero accidents - if you employ humans there will be accidents ergo zero accidents should never be a KPI merely an aspiration.

Hi lasersafetyguy

I am all in favour of tryinig to come up with one or more USEFUL leading indicators as a KPI for H&S etc.

But NOT whatever you might define as "near misses" mostly for very similar reasons you describe.

Suppose you set a target of say 100 reports per month.

Good chence that the target will be met. Even better chance if it is incentivised in whatever way.

But what you will likely end up with will be mostly not particularly helpful reports.

So worker spots something out of place on the shop floor and creating an obstruction and/or tripping hazard.

No KPI and hopefully worker picks that something up and moves it to a suitable location and DOESN'T fill in any paperwork.

As soon as you have a KPI, the focus is on making a report so worker reports and leaves it someone else to decide what to do and the problem continues to lie around until someone makes an executive decision.

I would much rather have 5 quality reports a month than 100 which are of little use.

A quality report is one that makes people think EVEN if they decide not to take action (and as long as the rationale for not taking action is explained to whoever has made the report).

As soon as you set a target you end up playing numbers games and there are often already too many numbers games when it comes to H&S. So somebody insists on counting numbers and rates of accidents etc and produces pretty graphs which for most organisations will have little, if any, proper statistical significance. 

Even the HSE falls for this type of behaviour, particularly when it comes to the headline statistics for fatal accidents. So e.g. here Work-related fatal injuries in Great Britain, 2025

Construction DOWN year on year from 40 to 35. That's a 12.5% reduction so perhaps we should have popped the corks?

From a statistical perspective this is not a "significant" change. Could easily be a fluke and the number the following year could easily be back to 40 or higher. 

Unless you have sufficient data, all the pretty graphs are meaningless.

If you have sufficient data, you probably have a much bigger problem than counting the numbers - UNLESS you are looking at trend analysis over a very long period.

Edited by user 03 December 2025 16:16:53(UTC)  | Reason: Typo