Prevention, Detection and Recovery Measures
Using a Bowtie perspective, root cause analysis checks the objective, the threats, the planned safeguards, and whether those safeguards worked. Prevention, detection, and recovery measures can also be examined before a failure occurs.
THE IDEA
Prevention, Detection and Recovery Measures
With the Bowtie perspective, we can start root cause analysis (RCA) if we fail to achieve an objective. We can ‘unpack’ causes, in a step‐by‐step way:
- First, was the objective we had clear and did we understand the threats to achieving it?
- Second, what measures were planned, or designed, to prevent things from going off track, detect a deviation, or recover afterwards?
- Third, were the measures implemented in practice and working effectively?
Considering how things are planned or designed provides an illustration of how RCA techniques can be used to anticipate and engineer away potential problems before they occur, increasing our chances of success.
When to use it
Analyzing a safety incident in an oil rig using the Bowtie framework to identify missing or ineffective measures.
Improving software deployment reliability by planning detection (monitoring alerts) and recovery (automated rollback) measures.
Conducting a post-mortem for a marketing campaign that missed targets, evaluating clarity of objectives and threats.
FROM THEORY TO EVERYDAY LIFE
In practice
Manufacturing Machine Failure
A manufacturing plant experiences a machine failure. Using the Bowtie perspective, the team first checks if the objective (operational uptime) was clear and if the threats (e.g., wear, lack of lubrication) were understood. They then examine planned measures: preventive maintenance schedules, detection sensors for vibration, and recovery procedures like backup parts. Finally, they verify if these measures were implemented correctly.
Software System Outage
After a critical system outage, the engineering team applies the framework: Was the objective (99.9% uptime) clear? Was the threat (e.g., memory leak) known? They review planned measures: automated monitoring (detection), stress testing (prevention), and rollback scripts (recovery). They then confirm if these were in place and effective before the incident.
A MOMENT TO REFLECT
Put it to work.
Think of a recent project failure in your work. Was the objective clearly defined? What were the known threats beforehand?
How could you design better preventive and detection measures for a recurring problem you face in your organization?
Examine a successful outcome you had. Which planned measures contributed most to that success?
CONTINUE THE CONVERSATION
Source texts
- Beyond the Five Whys Root Cause Analysis and Systems Thinking (James C. Paterson)