Continuity on Paper, Fragility in Practice: What a Three-Day Drill Reveals That a Binder Cannot
Photo: Kevin Rofidal, United States Coast Guard, Public domain, via Wikimedia Commons
Every organization has one. A business continuity plan, approved by a committee, stored in a shared drive, reviewed annually by someone who confirms the contact list is current. It lists scenarios — fire, flood, cyber incident, severe weather — and it lists responses, usually in the form of a call tree and a designated alternate site. It is a document written to satisfy a question. It has almost never been used to answer one.
That is the central problem with continuity planning as most mid-market firms practice it. The plan describes intent. It does not demonstrate capability. Those are different claims, and only one of them survives contact with an actual disruption.
Why 72 Hours
The first 72 hours of a disruption are where organizations fail, and they fail for reasons that are almost never in the plan. Not because the building is unavailable, but because the person who knows which purchase orders were open is on a plane. Not because the data center went down, but because the workaround requires a spreadsheet that lives on a laptop that was in the building.
Seventy-two hours is the right window for a drill because it is long enough to exhaust improvisation and short enough to be survivable. Most teams can absorb a bad afternoon through sheer effort. By day two, the workarounds start generating errors. By day three, the compounding effects appear — missed shipments, missed payroll inputs, missed filings — and those are precisely what the exercise exists to surface.
Designing Three Injections
Three scenarios cover the majority of realistic failure modes, and they can be run in sequence or in parallel.
Supplier failure. Select a vendor whose absence would be felt within a day — not your largest supplier, but your least substitutable one. Inform only the facilitator. The exercise is to place an order and carry it to fulfillment without that vendor's involvement, documenting every point where a person had to make a judgment call because no documented alternative existed.
System outage. Take down a system that transactions actually flow through — the order management platform, the warehouse management system, the billing engine — and require the team to process a defined volume of real transactions manually. The objective is not to see whether they can. It is to measure how long it takes and what the error rate is.
Key-person absence. Remove, for the duration, the person who holds the most undocumented knowledge in the operation. This is the injection teams resist most and learn the most from. In most mid-market firms, between four and eight individuals carry a disproportionate share of procedural memory, and the plan rarely identifies them by name.
Score Decision Latency, Not Heroism
The instinct after a drill is to declare success. Someone found a workaround, the shipments went out, the team pulled together. That conclusion wastes the exercise.
The useful metric is decision latency: how long it took, at each decision point, to determine who was authorized to act. In a well-functioning organization the answer is minutes. In most, it is hours, and the delay is rarely caused by the disruption itself. It is caused by ambiguity about authority — whether a supervisor can release an order without the plant manager, whether a buyer can exceed a threshold when the primary approver is unreachable, whether anyone other than the CFO can authorize emergency freight spend.
Record four things at every decision point: the time the issue was identified, the time the correct authority was located, the time the decision was made, and whether the decision required an undocumented workaround. Those four data points produce a fix list that is specific, fundable, and small enough to execute.
The Thirty-Day Rule
A drill without a deadline produces a report. Set thirty days from the exercise for every finding to be resolved or formally accepted as residual risk with a named owner and a date. In practice, findings cluster into three categories, each with a different remedy.
Authority gaps are resolved by a written delegation matrix naming alternates for every approval that matters, reviewed when people change roles rather than annually.
Documentation gaps are resolved by recording the workaround the team actually used, not the process the plan assumes. The most valuable continuity documentation in most companies is a set of one-page procedure notes written by the people who improvise daily.
Supply gaps are resolved commercially: a qualified second source, a standing agreement, or a confirmed and tested expedite path. Identifying a backup vendor on a list is not the same as having one that can ship.
Keeping the Drill Honest
The failure mode of any recurring exercise is that it becomes theater. Two practices protect against it. Rotate the facilitator, so the person designing the scenario is not the person defending their own function. And change at least one injection every cycle, because teams optimize for the test they have already seen.
The goal is not a better binder. It is a straight answer to a plain question: if three of your assumptions failed on the same Monday morning, would the business still be shipping by Thursday? A plan that has never been tested can offer only a hypothesis. A drill offers evidence — and evidence, unlike a binder, can be acted upon.