It was a missing audit trail.
In 2018, Boeing engineers expanded the 737 MAX's flight control system without re-running the safety analysis. The MCAS pushed the nose four times harder than specified, relied on a single sensor, and kept pilots out of pilot training.
Twenty months later: two crashes, a global grounding, a $225m shareholder settlement, and a federal judge ruling the board "publicly lied" about how safety was monitored.
The deeper failure wasn't in the engineering. It was in the discipline nobody was keeping: configuration management.
Named by the U.S. military in the 1950s, it runs on five steps:
- Plan the configuration
- Identify the baseline
- Control changes
- Account for status
- Verify and audit
Skip any step, and you haven't removed the work. You've only postponed it, with interest.
And the same pattern is now repeating in software. SolarWinds. Log4j. MOVEit. XZ Utils. Each one was an audit trail nobody had until a breach forced them to go find it. That's why regulators now require a Software Bill of Materials by law.
Here's the part that travels beyond aviation and cyber:
Most failures aren't new failures. They're old failures nobody bothered to map.
The supplier was onboarded last quarter. The SOP nobody refreshed. The variation someone "took care of." Most business failures are configuration failures wearing a different name.
So the audit trail isn't the paperwork. The audit trail is the product.
Our job as leaders isn't to cut that trail to move faster. It's to make sure the trail can survive the next surprise.
What's the one change in your operation right now that nobody has written down, and what happens the day it breaks?
