Fable
After one fire alarm, the city still has to learn to manage all fire alarms
The bell tower in Huaimu City rang occasionally. A small fire, a broken bridge, a flooded warehouse, a forged token slipping into the accounts room—each time it was handled on the spot by whoever was on duty that night.
Most incidents were eventually contained, so the city long believed it was very good at emergency response. But no one had counted how long it took from discovery to stopping the bleeding, and no one knew why similar incidents kept recurring.
An automated assistant mistakenly sent an old customer list to a new campaign, and customer service, legal, sales, and security all intervened at the same time. Each group kept its own records, and the customer received three different versions.
The city lord's naive fix was to demand that everyone "respond faster." But faster chaos is still chaos: some people notify customers twice, some restore features early, and some overwrite records while the root cause is still being investigated.
The new duty officer treated incidents as a complete process, not a single-point firefight. He required incidents to have an ID, severity level, commander, timeline, impact scope, communication cadence, and closure conditions.
Every incident first enters the register, then is triaged. P1 immediately pulls in a cross-functional war room, P2 is led by the business duty officer, and P3 enters normal scheduling. All decisions, assumptions, and operations are written into the same timeline. The city also learns to preserve the scene. Key records, operational traces, and notification times are no longer casually overwritten. Only by preserving the incident process can the retrospective avoid mutual blame and instead find ways to suffer less next time.
The end of an incident does not mean the meeting is over. The duty officer requires a retrospective, action items, owners and deadlines; recurring incidents must be escalated to systemic problems rather than continuing to blame a particular night shift. The duty officer also unified the external messaging. Customers first hear confirmation of impact, then interim remediation, and finally root cause and preventive measures; until there is evidence, no team may rush to conclusions.
A few months later, incidents in the city had not dropped to zero, but the average detection time, containment time, and quality of customer communication had clearly improved. More importantly, every incident drove improvements in queues, permissions, monitoring, or release processes. Only then did the city lord distinguish clearly: response is how to stop the bleeding in a single incident, while management is enabling the organization to continuously reduce the business losses caused by incidents.