Data Science

Seven Deadly Sins of Safety Data

After years of working with safety data, we have seen it all. Failing to get your safety data collection right can cause injuries or fatalities.

Seven deadly sins sign

After years of working with many organisations collecting their safety data, we have seen it all. Trust us. We see the same mistakes over and over. We don’t want to pretend we are the angels in this scenario – we have made our fair share of errors too, don’t worry. In order to help you avoid the blunders that all of us have made, we have distilled these common mistakes into 7 deadly sins. Are these Cardinal sins? Well, failing to get your safety data collection right will result in missed opportunities to identify trends that cause injuries or fatalities. So yes, indirectly, bad data collection could be deadly.

Gluttony (Quotas)

Some firms have a tendency to overindulge in good things like hazard identification reports, near misses and observations. I’m sure everyone will agree that from an injury reduction perspective there is gold in these reports. Collecting large numbers of these “good things” makes safety professionals look like stars! However, some firms attempt to gorge themselves on things like hazard IDs, near misses and observations by setting minimum reporting standards (quotas). This is a weak attempt to shortcut the real solution of creating a culture of reporting and it results in gobs of reports that are largely useless from a data analytics perspective.

This dilutes your safety data pool and largely negates the benefits of recording these types of records. It also erodes your employees’ confidence in your safety program. They are forced to come up with reports to fill their quota. They don’t see beneficial results from their efforts and they conclude that their efforts are a waste of time. We call this safety fatigue.

The cure to this gluttony and the way to get maximum benefit is to emphasize quality over quantity. Once workers are familiar with reports and have developed the habit (more quantity) then the focus can shift to quality, which in the end will yield more quality data for analysis.

Greed (Requiring users to enter too much data)

We all want to slice and dice our data six ways from Sunday and the simplest way to enable this slap chop style of data analysis is to collect every data attribute you can think of. What is the predominant eye colour of seriously injured people? Make your users record eye colour! Adding fields to your forms is easy…filling them out? Not so much.

Users are daunted by large forms and usually they’ll do a poor job of filling them out or find a way to shortcut the onerous task by providing brief or inaccurate data. What you’ll end up getting is either a bunch of records where the user has selected the “other” category all over the place or worse yet…nothing.

error with data

Creating data entry forms is a fine art. Forms must strike a balance between getting good quality data and keeping a user’s attention. This is difficult but very important. Ideally your application will walk the user through a series of questions and intelligently guide the user through the data collection process.

Machine learning can help here too. We have had success extracting valuable data from a user’s description using machine learning

Pride (Refusing to Adapt)

Pride in one’s safety program is a good thing but when that pride gets in the way of improvements, it can be dangerous – especially when it comes to your safety data recording program. Large firms are the most likely to be resistant when it comes to refusing to adopt industry standards. We hear refrains like “this is how we’ve always done it”.

There are always opportunities to learn from other organizations, industries and individuals. It’s a sin to ignore the things others have learned. The devil you know isn’t always the best way to go.

Participating in industry data standardisation initiatives will benefit all workers.

As leaders in your organization, remember that workers depend on you. Set standards that reward action and positive choices in safety.

Sloth (Change is too hard)

We see this all too often. Organizations (or rather certain people in organizations) are unwilling to change because it’s “too hard” or similarly, “we don’t have a problem”. There are really two possible reasons for this.

The first is that the person responsible for the change is just lazy and doesn’t want more work. If you are this person, you’re the worst.

The second reason is lack of management buy-in. Anecdotally, this is the strongest leading indicator of a company’s injury rates we’ve encountered. Time and time again the companies with lack of management buy-in have significantly worse injury rates than their peers. When it comes to sins, those “up above” have the most influence. Translation: if management isn’t on board, the change will not happen.

As leaders in your organization, remember that workers depend on you. Set standards that reward action and positive choices in safety. Dump those that say it is “too hard”.

Lust – (Lusting after low injury rates)

Way too many organisations are singularly focused on reducing injury rates. Bonuses are tied to it as is their job security or career advancement opportunities. Having an injury rate a notch or two above your previous quarter or worse higher than your competitor can result in punitive measures from, and to, management and senior leadership.

This creates strong motivation to misclassify and underreport incidents.

Lagging metrics are important to measure but they should not be the sole focus. It is a sin to ignore the host of other metrics which can be used to measure and predict performance. That array of metrics should be liberally populated with positive safety activities. 

The righteous path is to evolve to leading indicators for prevention!

Envy – (Theirs is bigger than ours)

Most of us are familiar with sibling dynamics. Little brother always has to be just like his big brother. A similar phenomenon exists within sectors. You’ve got big companies, small companies and then the mediums in between. Sometimes we see a medium company that’s trying to punch above their weight class. They want to be just like the big companies so they emulate and ACT big even though they’re not.

Sometimes, firms choose a safety management application that is too big for them because a big player in the industry purchased it. Big systems can come with many costs. They are expensive, difficult to configure, involve extensive change management, and require powerhouse technical capabilities and deep pockets.

Choose a data recording safety management application that’s appropriate for your company’s size and your maturity level. If you don’t, chances are you’ll pay for it in high software costs, an expensive implementation, failed adoption and worst of all…bad data.

Wrath – (Safety is a choice)

We see slogans like “Safety is no accident” or “all accidents are preventable” everywhere in industry. The problem with this paradigm is that it blames the worker. It ignores the fact that humans are fallible and places the blame on the shoulders of the injured worker. They suffer the wrath of management and co-workers.

The attitude of blaming the worker creates a culture of non-reporting. This will rob your organisation and your sector of valuable data that can be used to make your workplace safer and prevent injuries and potential loss of life.

stairs to heaven

Keep these guiding principles in mind when you’re designing, changing or enhancing your safety data collection to avoid these deadly sins. Your company and your sector will reap the rewards in the form of increased and actionable insight into problem areas.

Good data is a blessing: once you have it, you will be better positioned to implement truly impactful injury reduction programs.

Similar posts

Get the newest insights on safety analytics

Looking for up-to-date information AI for safety? Want best practices for your safety data? Subscribe to our newsletter for the newest updates!





Subscribe Now