Behavioral Observations Gone Wrong
On paper, the system looked solid. Every Operator was expected to complete three safety observations per week, enter them into the system, and discuss trends in team meetings. In practice, it felt like a numbers game grafted onto an already strained culture. Supervisors reminded people about “getting their observations in” more than they talked about actual risk. Techs started having fewer genuine safety conversations and started playing a game of “gotcha.” Instead of giving a quick reminder about a hard hat, peers silently submitted observations. The observed worker would return the favor the next day by writing up a failure to sign into the station in the morning. An initiative that was meant to prevent workplace injuries had devolved into “tit for tat” games and what felt like forced snitching.
As you may have guessed, this isn't a hypothetical. It's a real example that represents a pattern we've seen play out with clients across industries. A system that was intended to prevent incidents and build a culture of safety has, through a combination of poor implementation and gradual drift, become a source of resentment and distrust. But this outcome isn't inevitable. Behavioral observation systems, when built and maintained correctly, are one of the most powerful tools available for preventing workplace injuries. The problem isn't the concept. It's the execution. Understanding what these systems are supposed to do, and where they tend to break down, is the first step toward getting them right.
Why Are Behavioral Observations Important?
Safety observations are part of a proven approach to decreasing safety incidents by increasing safe behaviors on the job. Most organizations have a wealth of results or outcome data (e.g., incident rate, near misses) but little data on the behaviors that produce those results. If safety results improve, it may be tempting to assume that’s because people are working more safely. However, we know that’s not always the case. Safety outcomes may improve for several other reasons, like underreporting or simply having a lucky stretch. Behavioral observations pierce the veil to explain what safety data really mean and ensure that you get results the right way.
What Are Behavioral Observations?
Behavioral safety observations are a fundamental component of Behavior-Based Safety (BBS) systems. The process typically involves teams of frontline workers creating lists of specific critical behaviors that they would like to turn into safe habits. Peers observe each other doing the work and record the behaviors as either safe or at-risk on a scorecard. These observations focus specifically on objective behavior, rather than the physical condition of equipment or facilities.
In addition to observing and scoring behavior, there are other vital elements to conducting helpful BBS observations. Another critical component is a brief conversation after observing, where the observer provides feedback to encourage safer work practices. Where necessary, observations should uncover barriers to safe behavior too. The data gathered during observations are typically converted to graphic feedback which is periodically reviewed and analyzed. Ultimately, the goal of behavioral observations is to increase personal safety awareness and prevent workplace injuries by reinforcing consistent, safe habits.
Two Common Observation Errors
When done well, observations can raise awareness and build engagement, transparency, and ownership for safety. When done poorly they can undermine and erode safety. Consider the following ways safety observations go wrong and what to do about them.
Error #1: Using Observations Punitively
Using behavioral observations punitively is one of the most common ways a Behavior-Based Safety (BBS) system can drift into failure. Unfortunately, the default way of managing safety is through the use of negative consequences. We use the threat of negative consequences to get compliance (e.g., do it or else you could get hurt, fined, or fired). On the other end, we apply negative consequences when we observe at-risk behavior (e.g., reprimands, discipline, termination).
Thus, organizations default into using observations punitively, and there are several problems with that. First off, the effects of punishment are often immediate but temporary. I’m sure we’ve all had the experience of having a tough corrective conversation, seeing quick change, only to have the person revert to their old ways days or weeks later. Punishment also doesn’t teach the right behavior. Even if they stop the at-risk behavior, that doesn’t guarantee that they’ll start doing the safer one. Oftentimes people simply learn how to avoid getting caught. This can result in leaders creating an environment of constant surveillance, which undermines the foundation of trust that successful observation systems rely on.
Instead, leaders should be deliberate about pairing observations with positive consequences. Observations and the data they generate should never be used to blame or punish workers. That’s why the best observation systems are typically “no name, no blame.” Using a documented behavioral observation to discipline an employee will result in people refusing to do observations or simply pencil-whipping them, again undermining the spirit of open, honest communication. At-risk observations should instead initiate conversations about barriers to doing it safely. Leaders can use insights about barriers to avoid blaming workers and instead build systems that support safer alternatives.
Error #2: Conducting Observations Without Providing Feedback
Behavior change requires more than just observations and measurement. In fact, measurement by itself is unlikely to change behavior for long or at all. Behavioral observations do have the capability of producing lasting behavior when paired with feedback. Unfortunately, most organizations undervalue and underutilize feedback and allow observers to drift into conducting silent observations.
Observing and recording without feedback given during the observation misses the opportunity to encourage safe behavior or understand at risk behavior while it’s happening. To capture these valuable learning opportunities, peers must be willing to exchange some level of feedback. One consistent characteristic of effective feedback is that it is objective, which means it clearly describes the behavior in clear, observable, and specific terms. Verbally acknowledging a peer for wearing their Kevlar sleeves while in the shop will go a lot farther than silently observing them, or saying something vague like, “good job working safely.” It’s also important that the feedback comes across as sincere. A simple “nice handrail” from an introverted peer will go farther than saying nothing or adding so much floof that you sound like a walking chatbot.
Sustaining a “Want To” Culture of Observations
Grafting a behavioral observation system onto an already strained culture is a recipe for failure. Yet even when the cultural groundwork has been laid, it can be difficult to sustain a practice of conducting frequent, high-quality observations. This is when it gets tempting to pressure the frontline with negative consequences, which will produce quick but inconsistent observations of varying quality. Resist this temptation and instead consider how to connect the behavior of doing frequent, high-quality observations with more positive reinforcement (R+).
I'm thinking about this as using the right kind of fuel. R+ is high-performance fuel. Threats, surveillance, and negative consequences are regular unleaded. You can't sustain productive behavioral observations on regular unleaded. The highest performing safety cultures run on R+. It's what transforms safety from an agenda item to a core value and way of doing business.
So what does running on R+ look like in practice? It starts with leaders modeling the behavior they want to see, not just auditing it. When a supervisor stops to genuinely thank someone for following lockout/tagout procedure, that moment teaches more than any stated expectation ever will. It means discussing the data, highlighting progress, and asking questions about variation rather than calling out who's behind on their quota. It means giving workers real ownership over the critical behaviors list, the observation process, and the conversations that follow, so the system is genuinely something they helped build, not something that was built to monitor them.
None of this happens by accident, and it doesn't happen overnight. Cultures that "want to" do observations, rather than feel forced to, are the product of deliberate, sustained investment from leadership: training observers to give feedback well, protecting the system from being weaponized for discipline, and treating every observation as a data point about the work, not a verdict on the worker. Get that right, and behavioral observations stop being a compliance exercise and start becoming a daily habit of looking out for each other.
