Behavioral Observation: 7 BBS Failures to Fix

Behavioral observation fails when it becomes a counting exercise, because the number of forms can rise while exposure stays unchanged. This article shows the seven failures that turn BBS into theater and the practical corrections supervisors can apply on the floor.

When observations keep ending with awareness talks, leaders should test whether safety training is being used to avoid system repair.

Why behavioral observation often misses the real behavior

Behavioral observation is supposed to reveal how work is actually performed, not how a procedure imagines it should be performed. When the observer records only whether a rule was followed, the system may produce hundreds of monthly cards with little evidence of risk reduction, which is a warning sign for any EHS manager.

As Andreza Araujo argues in Safety Culture: From Theory to Practice, culture appears in repeated decisions, informal trade, and leadership tolerance, not only in written standards. That is why an observation program whose only output is a percentage of safe acts rarely changes the conditions that shape behavior.

The practical test is simple to audit. Select ten recent observations, then ask whether each one changed a condition, clarified a risk, corrected a work method, or opened a real conversation with the person doing the task. If the answer is mostly no, the program is documenting activity instead of changing exposure.

1. Observing compliance instead of exposure

The first BBS failure is treating compliance as the same thing as safety. A worker may wear gloves, goggles, and a harness while still being exposed to stored energy, falling objects, or a line of fire that the checklist did not capture.

Across 25+ years leading EHS at multinationals, Andreza Araujo has seen that the most dangerous gap is often the gap between a controlled-looking task and the unspoken workarounds that make production possible. This is where serious risk can stay hidden, because the matrix says medium while the field situation is one failed control away from a severe event.

Supervisors should ask three exposure questions during every observation: what could release energy, what could put the body in the line of fire, and what control would have to fail for the event to become serious. The answer matters more than the safe-act percentage.

2. Counting cards instead of improving conversations

The second failure is using observation cards as proof of engagement. A site can reach 95% completion on its observation target while employees still avoid naming production pressure, weak planning, or missing tools because the conversation feels punitive.

The Vamos a Hablar? method, developed in Andreza Araujo's behavioral observation work, treats the card as the last artifact, not the main event. The main event is the dialogue, where the observer understands why the behavior made sense to the worker at that moment.

Replace the question "why did you do that?" with "what made this the easiest way to do the task today?" The first question often sounds like accusation, although the second one exposes design, time, training, and supervision factors that a checklist usually misses.

3. Training observers without training supervisors

Behavioral observation loses strength when the observer is trained but the supervisor does not know how to act on the finding. In that model, the observer reports exposure, the supervisor receives another item, and the worker sees no visible change after speaking honestly.

In Make The Difference: Be a Leader in Health & Safety, Andreza Araujo connects leadership to visible care, because the leader's reaction after a safety conversation tells the team whether speaking up is useful. This link explains why speak-up metrics belong close to behavioral observation indicators.

Each observation that identifies a real barrier should have one owner, one due date, and one return conversation with the worker or crew. The return conversation is where credibility is built, since people stop reporting weak controls when nobody closes the loop.

4. Turning BBS into blame with polite language

BBS becomes blame when the system treats the observed person as the main cause of the unsafe act. The language may sound respectful, but the action plan still says retrain, remind, or discipline, which means the investigation stopped too soon.

James Reason's Swiss Cheese Model helps separate active errors from latent conditions, while Andreza Araujo's Portuguese title Sorte ou Capacidade, or Luck or Capability, reinforces that serious events are rarely explained by one person's choice. The same logic applies to RCA that avoids the operator error trap.

A better observation closes with two records: the behavior seen and the condition that made the behavior likely. If the condition is absent from the record, the company may correct the worker and preserve the exposure.

5. Ignoring the quality of the sample

Observation data becomes weak when the sample is convenient rather than representative. Day-shift office walks, low-risk areas, and scheduled visits create a clean report, although the real exposure may sit in night maintenance, contractor work, start-up, shutdown, or temporary tasks.

In more than 250 cultural transformation projects supported by Andreza Araujo's team, observation quality has depended less on volume and more on where leaders choose to look. A program that never enters uncomfortable work rarely earns trust from the people whose tasks carry the most risk.

Build a simple sampling map by task, crew, shift, contractor status, and SIF exposure. The EHS manager should review the map monthly and ask which high-energy tasks received no observation, because absence of data is often the most important data point.

Each month without sample discipline allows the safest-looking areas to dominate the dashboard, while the highest-risk tasks remain invisible until an incident forces attention.

6. Separating behavioral observation from leading indicators

Behavioral observation should feed leading indicators, not sit beside them as a separate activity. When observation quality, closure time, repeated barriers, and supervisor return conversations are tracked together, the program starts to show whether exposure is moving.

The problem with many dashboards is that they put BBS cards next to TRIR as if both signals had the same meaning. The better question is whether observation detects weak controls before injury occurs, which connects directly to leading indicators TRIR will never show.

Use four measures for the first 90 days: percentage of observations linked to a critical task, percentage that names a condition, median days to close the condition, and percentage with a return conversation. These measures are harder to inflate than card volume.

7. Leaving culture out of the observation design

Behavioral observation cannot repair a culture that punishes bad news. If employees believe that every observation may become evidence against them, they will show the expected behavior while the observer is present and return to the practical workaround as a repeated safety habit afterward.

In Safety Culture Diagnosis: Learn how to do your own, Andreza Araujo presents diagnosis as the starting point for cultural work because perception predicts participation. A team that does not trust leadership will not offer the information that makes observation useful, which is why safety culture diagnosis should precede any serious BBS redesign.

The cultural question for leaders is not whether employees know the rule. The question is whether employees believe the organization wants the truth about how the work is really done, especially when that truth exposes weak planning or production pressure.

Behavioral observation that works vs BBS theater

Conclusion

Behavioral observation works when it exposes why a behavior made sense, which control was weak, and what leadership changed afterward.

For companies ready to move beyond card volume, Andreza Araujo's work connects BBS, safety culture diagnosis, and visible leadership into one practical system. If your operation needs that transition, start with the books and diagnostic resources at Andreza Araujo.