The call came in at 3:14 a.m.: a key inbound carrier had gone dark mid-route — no answer, no tracking ping, a full trailer of time-sensitive freight somewhere on an interstate with no update. The night supervisor had two options, neither good: hold the dock crew on overtime waiting for freight that might not show, or send them home and scramble in the morning if it did. He had about four minutes to decide before the decision made itself by default.
Here's the uncomfortable truth about disruptions like that one: your actual disaster-response plan isn't the document in the shared drive. It's whatever the most tired, least senior person on shift does in the first ten minutes before anyone experienced is even awake. Every resilience program, every tabletop exercise, every contingency slide deck is really just an attempt to make that decision less lonely when it happens.
What separates a good 3 a.m. decision from a bad one
It's almost never about having more information — at 3 a.m., you never will. It's about whether the person facing the decision knows three things in advance: what the actual priority is when two good options conflict, how far their authority extends before they need to wake someone up, and that they won't be punished for a reasonable call that didn't work out. Take any one of those away, and even a capable person will freeze, escalate everything, or guess badly.
Resilience research consistently finds the same pattern: the organizations that recover fastest from a disruption aren't the ones with the most redundancy. They're the ones where frontline decisions get made quickly, because the people making them already know the boundaries they're operating in.
That framing comes from work at MIT's Center for Transportation and Logistics, which has spent years studying what actually determines how well supply chains absorb shocks — from port closures to supplier failures. Their consistent finding isn't "add more buffer inventory" or "add more backup vendors," though both help. It's that resilience is disproportionately a decision-speed problem, not a resource problem. The fastest-recovering networks are the ones where authority to act is already pushed down to where the disruption is actually happening.
Building the decision in advance, not in the moment
- Name the priority before the crisis, not during it. "If we ever have to choose between an on-time metric and a safety shortcut, safety wins" is a sentence that should exist before 3 a.m., not get improvised during it.
- Draw the escalation line in writing. Below this cost, this delay, this headcount — decide and inform. Above it — wake someone up. Ambiguity here is what actually causes both the paralyzed calls and the reckless ones.
- Debrief the call, not just the outcome. A reasonable decision that went badly and a reckless one that got lucky should get very different feedback — but only if you're actually reviewing the reasoning, not just the result.
The supervisor that night sent the crew home. The trailer turned up at 6:40 a.m. — a dead phone battery, nothing more dramatic than that. It could easily have gone the other way. What mattered wasn't that he guessed right; it was that he'd known, going in, exactly how far his call was allowed to go. That's the part you can actually build ahead of time.
Sources:
- MIT Center for Transportation & Logistics, supply chain resilience research.