Tabletop Exercises That Work

A tabletop exercise is a discussion based rehearsal in which participants work through a scenario together, without touching live systems or customers. Its strength is that it can test judgement, coordination and communication at low risk, and its weakness is that it can drift into a pleasant conversation that changes nothing. A tabletop works when it is built around decisions rather than presentations, when participants must answer in role, and when disagreement is welcomed as evidence that a real choice is being examined. The session should end with named changes, not with applause for teamwork. Facilitation, preparation and follow up matter as much as the scenario itself.

Preparation starts with choosing participants who would actually decide or act in a real disruption, rather than a general audience. Each participant should receive enough background to play their role, including the service, its tolerance and the current plan, without receiving the scenario in such detail that surprise disappears. The facilitator should prepare injects, which are pieces of information or complications introduced during the session, and should know which decision each inject is meant to force. A short briefing at the start can set ground rules: speak in role, admit uncertainty, and treat a gap found in the room as a success. Without those rules, seniority can dominate and junior staff who know the workaround may stay quiet.

Facilitation is the difference between a discussion and an exercise. A good facilitator asks who decides, what information that person needs, and what happens if the information is unavailable, rather than allowing the group to agree quickly on a sensible sounding course of action. Silence after a difficult inject is useful, because it mirrors real uncertainty. The facilitator should also prevent the group from solving the scenario by assuming resources that have not been confirmed, such as a supplier response or a backup that works. When participants propose an action, the facilitator can ask how it would be carried out, who would carry it out, and how the firm would know it had worked. Those follow up questions turn general intentions into testable claims.

An inject is useful when it changes the decision, not when it merely adds drama. A useful inject might remove a key person, contradict an earlier report, introduce a customer complaint that changes priorities, or reveal that a workaround depends on an affected system. Consider a tabletop for a lending service facing a supplier outage. An inject stating that the supplier will not respond until the next working period forces the group to decide whether to pause new lending, to use a manual process, or to accept a backlog. An inject that simply states that the outage continues adds nothing. Injects should arrive in a sequence that builds pressure, with time for the group to record decisions before the next complication arrives.

Observers watch for gaps that participants, absorbed in the discussion, may miss. They note decisions that were assumed rather than made, information that was requested but never provided, and handovers that were unclear. They also watch for the pitfall at the heart of this cluster: a group that designs its own success by softening the scenario. Observers should be briefed to record evidence, not impressions, such as the exact wording of a decision or the time at which a customer message was agreed. Their notes become the raw material for the debrief, and they help prevent the common outcome in which everyone remembers a good discussion but nobody remembers what was decided.

A tabletop leads to change when its debrief is structured around actions rather than praise. The debrief should revisit each major decision, ask what information was missing, and identify the plan, contact, procedure or training that needs to change. Each action needs an owner and a check that it has been completed, and the next tabletop should test whether the change works. That loop is what separates rehearsal from theatre. A tabletop that produces no action may still have value as an introduction to a plan, but it should not be recorded as a test of resilience. Recording it honestly protects the firm from false reassurance, which is more dangerous than admitting that a plan is still untested.