By Phil La Duke
The use of incentives is something of the Great White Whale of safety. Safety practitioners often find mounting pressure to use incentives to reduce workplace injuries. Incentives are popular because they seem to make sense—and I am not against incentives, provided that they encourage the right things. Sadly, incentives too often create unintended consequences, chiefly because the incentives are for the absence of injuries instead of the presence of safety.
There is a gulf between the apparent absence of injuries and the presence of safety and unfortunately neither of these are particularly easy to measure. One can’t measure the absence of injuries because one must depend on the injuries being discovered—either via self-reporting or discovery by the organization. Effectively zero injuries (or any number of injuries for that matter) is zero reported injuries; it’s a case of “we don’t know, what we don’t know”. But measuring safety is just as difficult because we don’t really have a hard and fast definition of exactly what constitutes safety. What we describe as “safety” is more accurately “safe enough” and ask seven Safety Practitioners what Safe Enough means and you are likely to get 19 answers. Safety is a continuum and is relative so it cannot be accurately (in and of itself) measured. Safety can only be measured as a state relative to another state. Something can be said to be safer than something else, but as long as any risk exists something can never be pronounced completely and utterly safe. If we can’t pronounce something qualitatively “safe” we have to rely on indicators, unfortunately, incentives are often misapplied or misinterpreted. It’s impossible (well at least foolish) to talk about incentives without considering indicators, and if we are going to provide incentive for the right things we need to understand, first and foremost, what indicators are telling us.
The Absence of Evidence is Not Evidence of Absence
The most commonly used indicator of safety is the extent to which injuries occurred. If people were injured it’s appropriate to say that they weren’t safe; that’s intuitive and people like it because it’s a simple calculation to make, provided that people report injuries. But as I’ve said, safety isn’t just the absence of injuries; it’s also the presence of things that drive a safer workplace, and that is the crux of the issue with indicators and incentives. Let me illustrate: the opposite of injuring workers (i.e. an indicating a lack of safety) is the absence of injuries. What does the absence of injuries indicate? Safe work habits? I know many people with incredibly unsafe work (or driving) habits who don’t get hurt, so while it’s possible that a lack of injuries indicate safe work habits it’s equally (perhaps more) likely that a lack of injuries indicates luck. Could it indicate that no one has been hurt? Possibly, but here again it could also indicate that people have been concealing their injuries. Could it indicate overly zealous case management? It might. In fact, there are numerous things that a lack of injuries could be indicative of so we can’t really use them as a good indicator.
Look For the Things That Produce Safer Outputs
I’ve come to realize that “safety” is really an output of sound business practices in five areas (there are many subsets within these areas, but five is a nice manageable number):
- People who are incapable of doing their jobs—whether it be because of a lack of training, or physical incapacity or insufficient intellectual ability—are less likely to work safely than the workers who possess these attributes.
- Process Capability. Work environments that lack a standard way to do the job that contains minimal variation are safer than work environments where workers half to figure out how to do the job each time they repeat a task. Similarly, workplaces with weak process discipline (the practice of following the prescribed process) are less safe than environments with strong process discipline. In other words if your jobs and tasks are poorly defined or your people are working out of process you are at greater risk of injury than if you have a well-defined process that people don’t follow.
- Risk Management. Organizations that appropriately assess and mitigate their risks are far safer than organizations that don’t manage hazards.
- Accountability. From the CEO to contractors, it is important to hold people appropriately accountable for doing their jobs correctly. Accountability systems must reflect corporate justice (in Just Culture parlance console human error, coach risk taking, and discipline recklessness).
- Engagement. Workers who are actively trying to improve the safety of the workplace because they believe that it’s the right thing to do are more likely to produce safe outcomes than those who aren’t engaged.
If we can accept that these five processes, if managed appropriately, will produce safe outcomes (and for the record, there are others, but like I said, they can be managed within these categories, but if you choose others I won’t gripe.) than we can look for things that indicate the presence of well-managed processes in these areas.
Indicators of Well Managed Processes
Indicators of well-managed processes may differ from industry to industry, even from site to site, but in broad strokes we can measure indicators of success in these areas.
Indicators of competency
How do you measure competency? If you don’t know ask your training department; you are likely to find that they are adept at measuring competency, but here are some suggestions:
- % trained. Personally, I wouldn’t limit this to safety training, although the percentage of people who have successfully completed training on time is a good indicator of competency. Of course it’s not the only indicator and the more indicators you use the stronger your confidence can be that whatever you are measuring is true. Since we are only looking at five areas we can use several indicators for each and have a much stronger correlation between the indicator and reality.
- % hired with all required/desired skills. We all know that job postings are essentially wish lists and there is seldom a new hire that hits ALL the requirements. The greater the percentage qualified the higher the likelihood that the person will be able to perform safely.
- Those individuals with higher skills tend to have higher productivity than those who don’t, so while productivity is an indicator for more than just competency it can be useful in conjunction with other indicators.
So how do we create incentives around these factors? Simple: reward people (at all levels) for completing their training on time, for hiring more skilled workers, and for maintaining high productivity.
Indicators of Process Capability
This is the easiest area for which to develop indicators because in many organizations there are already measurements that we can use to gage safety:
- Unplanned downtime. Unplanned downtime tends to indicate process breakdowns and the greater the frequency of unplanned downtime the higher the likelihood that workers are at risk of injury.
- Like unplanned downtime, scrap indicates a process that is out of control. Workers who are working in a process that is out of control are by definition working out of process. Since we tend to see more people hurt while they are working out of process this is a good indication of the level of safety.
Indicators of Risk Management
For our purposes we will define risk management as how the organization identifies, contains, corrects, and communicates hazards (including injuries). In this area there are a lot of things from which we can choose:
- % of walk-throughs completed on time. Whether you have BBS audits, Safety Observation Tours, Layered Process Audits, you probably have some formal requirement for the supervisor to identify hazards. Your requirement should have a frequency requirement that is easy to measure. The indicator here is mathematical—the less time someone is exposed to a hazard the less risk of injury. Meeting the requirement to complete these tasks on time is a strong indicator of safety.
- Number of hazards per tour. Hazards (especially behavioral) are dynamic so the number of hazards a person finds each tour correlates to the safety of the workplace.
- Number of overdue hazards. The priority assigned to the correction of a hazard should have a corresponding deadline and when that deadline isn’t met it indicates an increase in the time of exposure and perhaps a degradation of the containment measures.
Of course there are a lot more indicators you can use in this area, but I think you get my point.
Indicators of Accountability
Accountability should be just; the punishment should fit the crime. Justice is largely circumstantial—not every situation can be treated according to the same standard of accountability. Would you discipline a worker who mistakenly used the wrong we chemical and caused property damage as you would someone who engaged in sabotage? Or would you react the same way to a worker who faced with two pretty bad choices (after careful analysis) decided to choose the lesser of two evils as the worker who engages in clear recklessness? Of course not. Unfortunately we can’t feasibly measure the justice of a decision, but we can of course measure the number of write-ups, improvement plans, and similar efforts. We should also be looking at the number of times we “caught them doing something good”. Some examples I can think of off-hand include:
- Number of disciplinary actions. Clearly the number of disciplinary actions directly correlate to accountability; the more disciplinary actions the higher the accountability. But what if there are few disciplinary actions simply because there are less people who are acting inappropriately? Clearly this indicator cannot be interpreted alone and should be paired with an indicator that people are being recognized and rewarded for desirable behavior.
- Number of employees recognized for exemplary service. The number of people who are recognized for doing things like identifying a serious hazard, participating in safety efforts, leading a safety event, or something similar is also an indicator of accountability—rewarding desired behaviors. By pairing this with the number of disciplinary actions one can get a better picture of the overall performance of accountability.
Indicators of Engagement
Engagement, like process capability, is likely already being measured by your organization, but you can use some combination of the following to ascertain the level of worker engagement:
- Number of Grievances. Unhappy workers tend to have more “performance inhibitors”; that is, the things like stress, preoccupation, anger, frustration, etc. that increase the likelihood of human error. Not to mention unhappy workers may make poor choices rooted in frustration.
- Number of Suggestions. The flipside of grievances is suggestions. The greater the number of suggestions for approval the higher engagement tends to be.
- Participation in continuous improvement efforts. People who care about their work tend to get involved in making it better and this tendency is a good indicator of engagement.
- Participation in safety meeting. Participation in a safety meeting, like so many other indicators, cannot be seen as an absolute indicator of engagement; it could indicate that someone would rather sit in a meeting than do what they are paid to do. But when taken with these other indicators it can provide insight into the level of engagement of workers in an organization.
- Here again is an indicator of more than a lack of engagement, but this is a strong indicator of the relative safety of a workplace. High absenteeism is linked to poor morale, unhealthy working conditions, workers not managing their performance inhibitors (drinking to excess, drug use, sleep deprivation, etc.), but more than that, high absenteeism means more replacement workers who tend to be less skilled at performing the job. This ties into competence, process capability, and perhaps even risk management.
- Moral is also an indicator of many factors, but low moral does correlate to higher incidence of human error and risk taking.
- Turnover is a good indicator of an overall healthy or unhealthy workplace. As people are churned it lowers competency and impedes process capability.
Okay, but what about incentives?
One can only effectively set incentives that reduce the chance of unintended consequences after one has appropriate indicators of safe outputs. Once one has determined the best measures for the desired state one can then create appropriate incentives. When developing incentives:
- Look for (and avoid) potential unintended consequences. Too often incentives create an environment where the desired behavior isn’t rewarded and people game the system.
- Don’t provide incentives (or hold people accountable) for things they can’t control. When you provide incentives for things people can’t control the only real incentive is to lie, cheat, and steal. Takes sales incentives for example. While a salesman can control how many meetings he has with prospects (which is necessary to MAKE sales) he can’t really control whether or not a sale is made. This leads to bickering between sales professionals, stealing of clients and leads, undermining competitor’s success, and generally stabbing one another in the back. It actually diminishes teamwork, collaboration, and ultimately the likelihood of success for the company.
- Make it meaningful. Not everyone likes to be recognized or rewarded in the same way. Be sure to consider different people’s needs.