On Data-Driven Compliance

Image for post
Image for post

In our previous project we have determined the impact of workplace safety initiatives, to find out which management efforts are the most effective in preventing workplace accidents. We have further developed a predictive model that served as an accident forecast. Similar to a weather forecast, but instead of telling you where it’s gonna rain we’d tell you where you’ll have accidents. This allows the safety function to focus on locations with imminent risk instead of a blanket approach.

Now that we have determined what we need to do, and also where we need to do it, we still don’t know how???

Compliance remains dependent on the talent of individual safety managers, their ability to connect with and motivate employees under their responsibility. A difficult task, as we found that ~30% of employees do not complete any safety reports, despite having the same requirements as everyone else. How do we get them to do the right thing?

One approach to rule them all?

To solve this issue, we have followed an approach developed by Dr. Howard Moskowitz, inventor of Mind Types and Mind Genomics. The basic idea is that there can not be only one approach. Different people respond to different things. Think of all the different types of soft-drinks, candy bars, or cereals. There are literally hundreds of them. Why doesn’t a big company like Kellogg’s hire the best scientists and the best engineers to develop the best cereal that everybody would love? Because it doesn’t exist. Everybody likes something different, and everybody will respond differently to different approaches.

Safety management systems ignore this fact and try to find that one unified approach, one message that everybody will be happy with and comply to 100%. As we’ve seen in practice it doesn’t work. Some more talented managers are able to tune in to their staff and recognize everyone’s individual needs and motivate them better than others. This is why we often see a slight improvement upon change in leadership. A different message will address a different part of the employee population who become more motivated. Eventually employees that responded better to the previous approach will become de-motivated and we are back to the baseline.

With modern data science, it is possible to determine the optimal approach for each individual employee. In order to do so we ran a study on the general population, to see if we can find different groups of people that respond to different motivators with respect to workplace safety practices. We were trying to determine what motivates them to submit safety reports. The study worked better than expected and delivered excellent results.

One for all and all for one?

Image for post
Image for post

We were able to group our respondents according to three distinct attitudes.

  • One group cares only about the Visibility of the safety reports. Seeing reports from their manager or their colleagues motivates them to complete more themselves.
  • The second group is more interested in the Usefulness of the safety reports, knowing they don’t end up in a black hole and seeing positive examples of accident prevention. They can actually be turned off by messages addressing visibility.
  • The third group are the competitive people that care about getting the High-Score. They can be motivated with competitions and raffles, and will get their numbers in just to be the best.

Each of the groups represents a different school of thought amongst our population. We can define the types of messages that each person will feel most passionate about and approach each individual in the best possible way. This way we can increase compliance rates and reduce ambiguity on managing talent. Imagine giving the safety managers a cheat-sheet on how to approach each employee, without doubt.


In this example, we have shown an approach to data-driven compliance management. After determining the most effective actions, and where best to apply them, we have now outlined how they can be implemented in a systematic way that doesn’t depend on the talent of the manager.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store