I don't want the details
Short summary
After incidents, leaders should skip the detailed explanations of what happened and instead focus on what systemic changes will prevent it from recurring. Understanding why something went wrong doesn't drive organizational change—only fixing the underlying system does.
- Detailed postmortems that explain reasonable behavior actually prevent change because once everyone agrees the decisions made sense, the urgency to improve vanishes
- Effective corrective actions must be system-based, not people-based—if the fix depends on someone remembering a conversation or being more careful, it won't work when that person leaves or circumstances repeat
- Not every failure needs a new process, but you need to consciously accept that risk rather than pretending vague commitments like "we'll communicate better" constitute actual fixes
Full summary
An SVP of engineering shut down a detailed explanation of an incident by saying "I don't want the details." Instead of hearing a timeline of what went wrong, they wanted to know what the organization would change to prevent it happening again. The SVP's point was sharp: assuming the people involved were competent, understanding why something happened won't fix anything. Once you explain the sequence of reasonable decisions that led to failure, everyone nods, feels satisfied that the behavior made sense, and walks away without changing anything. Six months later, the same failure repeats.
The real question after an incident shouldn't be "Why did this happen?" It should be "What are we changing so this class of failure is less likely next time?" The author gives concrete examples: if ownership was unclear when someone was on vacation, build a system that makes ownership unambiguous. If requirements shifted three days before launch, establish a process that forces a decision when that happens. If alert fatigue caused a miss, fix the signal-to-noise ratio. The focus is systemic, not personal—because people are usually making the best decisions they can given their constraints.
Most postmortems fail because they end with hopes dressed as action items: "we'll communicate better" or "we'll be more careful." These don't work unless the same people stay and remember the conversation. A real fix is one that would produce a different outcome even if everyone involved left tomorrow. Not every failure needs a new process—sometimes accepting a risk consciously is smarter than building bureaucracy. But you have to own that choice explicitly rather than pretending you'll try harder next time.