The First Management Failure: Who Investigates Matters More Than How
The identity of the investigator matters more than the method
If you want to know whether leadership truly wants to find out what happened, look at who they put on the investigation team. The team composition is the message.
Why the most consequential decision in any investigation is made before the investigation begins
The most consequential decision in any incident investigation is made before the investigation begins: who leads it, who is on the team, and what authority they carry. In most organisations, this decision is not made at all — it defaults. The site HSE officer picks up the file. The supervisor joins because they were there. A manager sits in because it happened in their area. Nobody asks whether this team has the competence, independence, or mandate to challenge the management system that produced the incident. This article argues that team selection is the first — and often the most revealing — management failure in incident investigation.
THE NUMBERS THAT MATTER
INVESTIGATION TEAM INDEPENDENCE: THE REALITY
Direct reporting line to implicated managers
Almost never formalised
Usually mandated by site or region
Investigators juggling operational duties
100% local teams are the norm
THE GAS TERMINAL WHERE THE INVESTIGATOR REPORTED TO THE SUSPECT
At an LNG processing terminal, a process safety incident resulted in an uncontrolled hydrocarbon release. Nobody was injured, but the potential consequences were catastrophic. The site's HSE manager was assigned to lead the investigation. He was competent, experienced, and diligent. He was also a direct report of the operations manager whose team had authorised the work that led to the release.
The investigation was thorough in its evidence collection but shallow in its causal analysis. It identified a failed isolation valve and a missed step in the permit-to-work procedure. It recommended retraining for the operators and a revision of the isolation checklist. It did not examine why the operations manager had approved a procedure that bypassed a critical safety step. It did not question the staffing levels that forced the permit authority to process 14 permits in a single shift. It did not address the production pressure that had been escalating for months.
The HSE manager did not suppress these findings. He simply never looked for them. His reporting line, his professional relationship with the operations team, and his own position within the organisation's hierarchy meant that the investigation's ceiling was structurally predetermined.
■ THE STRUCTURAL CEILING The first management failure was not the failed valve. It was appointing an investigator whose independence was compromised before the first interview. The investigation could never go higher than the investigator's reporting line allowed.
WHY TEAM SELECTION IS A GOVERNANCE DECISION, NOT AN ADMINISTRATIVE ONE
The Conflict-of-Interest Principle
In corporate governance, financial regulation, and judicial proceedings, the principle is unambiguous: no person with a material interest in the outcome of an inquiry should participate in conducting it. Auditors cannot audit their own divisions. Judges recuse themselves from cases involving personal connections. Directors declare conflicts before board votes.
In incident investigation, this principle is routinely violated. Supervisors investigate incidents in their areas. HSE officers who approved the risk assessment examine why it failed. Managers who set the production targets that created the pressure review whether pressure was a contributing factor. The conflict is structural, not personal — and that makes it invisible to those inside it.
KEY TAKEAWAY Structural conflict of interest is not about dishonesty. It is about the limits of what someone can see when their career, relationships, and professional identity are embedded in the system they are investigating. The investigator is not biased — they are structurally blind.
Aviation's Model: Structural Independence
The UK's Air Accidents Investigation Branch (AAIB) operates on a foundational principle: the investigation body must be structurally independent from the operator, the regulator, and the manufacturer. This independence is not optional. It is the precondition for investigative integrity.
Industrial safety has no equivalent. Investigations are conducted by employees of the organisation whose system failed. The investigator's career, salary, and professional relationships are all embedded in the organisation they are investigating. This does not make them dishonest. It makes them structurally incapable of seeing what an independent team would see.
THE FIVE REQUIREMENTS FOR AN EFFECTIVE INVESTIGATION TEAM
Requirement 1 — Competence in Advanced Causation Methodology: At least one team member must be trained in a structured causation methodology (Tripod Beta, AcciMap, STAMP, BowTie). Without this, the investigation defaults to linear cause-effect thinking, which consistently stops at the worker level.
Requirement 2 — Structural Independence: The team leader must not report, directly or indirectly, to any manager whose area, decisions, or systems are within the scope of the investigation. This is non-negotiable. If the team leader's career is influenced by the person whose decisions may have contributed to the incident, the investigation is compromised from day one.
Requirement 3 — Conflict-of-Interest Declaration: Every team member must declare any relationship, reporting line, or prior involvement with the incident, the area, or the people involved. This is standard practice in financial auditing and legal proceedings. It is virtually unheard of in incident investigation.
Requirement 4 — Mandate from the Highest Level: The investigation team must receive its mandate from the CEO, COO, or board — not from the site manager or regional VP. The level of mandate determines the ceiling of the investigation. If the mandate comes from the operations manager, the investigation will not examine the operations manager's decisions.
Requirement 5 — Dedicated Time and Resources: Investigators must be released from operational duties for the duration of the investigation. A part-time investigation conducted between production meetings and safety walks is not an investigation. It is a documentation exercise that will produce documentation-quality conclusions.
■ THE MANDATE TEST Ask: who issued the investigation mandate? If the answer is anyone below C-suite level, the investigation has a ceiling. The level of the mandate is the level of accountability the organisation is willing to examine.
THE TEAM SELECTION AUDIT: A DIAGNOSTIC FOR COOs
| # | Question | Failure Indicator | |---|----------|-------------------| | 1 | Does the team leader report to anyone whose decisions are within investigation scope? | Any direct/indirect reporting line | | 2 | Has every team member declared conflicts of interest? | No formal declaration process | | 3 | Does at least one member hold advanced causation methodology credentials? | No member trained beyond basic RCA | | 4 | Who issued the investigation mandate? | Mandate issued below C-suite | | 5 | Are team members fully released from operational duties? | Any member balancing dual roles | | 6 | Does the team include external or cross-site members? | All members from same site/unit |
“If you want to know whether management truly cares to find out what happened, look at who they put on the investigation team. The team composition is the message.”
— Bruno Hounkpati
References
ILO (2024). Nearly 3 million people die of work-related accidents and diseases.
Lloyd's Register Foundation (2024). World Risk Poll 2024: Engineering Safer Workplaces.
UK AAIB (2023). Annual Safety Report. Air Accidents Investigation Branch.
Reason, J. (1997). Managing the Risks of Organizational Accidents. Ashgate.
Leveson, N. (2011). Engineering a Safer World: Systems Thinking Applied to Safety. MIT Press.
Helmreich, R. L. & Merritt, A. C. (1998). Culture at Work in Aviation and Medicine. Ashgate.
MacLean, C. L. (2022). Cognitive bias in workplace investigation. Applied Ergonomics, 105, 103860.
Rasmussen, J. (1997). Risk management in a dynamic society. Safety Science, 27(2-3), 183–213.
This article is published by HSESKILLS Ltd for educational and informational purposes only. Composite scenarios illustrate common investigation patterns and do not refer to any specific organisation unless explicitly named.
