How do we run a tabletop that tests decisions rather than rehearsing a script?

Start

The objectives are approved and the injects are written

A tabletop becomes a rehearsed script when participants know what is coming, when the facilitator steers them to the answer, or when nobody records who decided what. The pack guards against each. The exercise tests the plan and the organisation, never the people in the room.

• Then: Send the plan and the ground rules, never the injects

Step

Send the plan and the ground rules, never the injects

About 5 working days before, in the pack's default, send each participant the parts of the plan their role uses, the playbook for the scenario family, the ground rules and the times. Say which kind of incident it concerns only if the sponsor agrees. Never send the injects or the expected answers. The scenario pack stays in a folder only the facilitator and the scenario checker can open.

• Then: Are the objectives approved and an independent observer confirmed?

Decision

Are the objectives approved and an independent observer confirmed?

The day before, the exercise lead runs the final check. Without approved objectives the exercise cannot be rated. Without an observer independent of the team, nobody outside it judges the decisions. If the facilitator holds a role in the scenario, their deputy plays it or another facilitator runs the session.

• Yes: Pre-brief: objectives and decisions, never injects or answers
• No, or not known: Postpone, or have the sponsor approve a replacement observer

Outcome

Postpone, or have the sponsor approve a replacement observer

Postponing is better than running an exercise that cannot be rated. With no independent observer, the sponsor decides between postponing and a named replacement. If a postponement would take the audience past its interval, executive management decides whether to accept the delay with a date or find another slot.

Step

Pre-brief: objectives and decisions, never injects or answers

The first 15 minutes: why this exercise, the objectives with the decision each tests, the ground rules read out in full, who is who, and the two clocks. The ground rules include no blame; play your own role with what you know at that moment; decide, because 'I would check the plan' is not an answer; accept the scenario. Agree a stop phrase that ends the exercise if a real incident needs someone.

• Then: Give each inject at its minute, read as written

Step

Give each inject at its minute, read as written

The facilitator reads each inject at its minute on the timing card and states the scenario time when it jumps. Then asks for a decision and who takes it: 'so what do you do now, and who does it?' Questions about the scenario are answered from the facilitator's card. The facilitator never says whether a decision is right and takes no decisions.

• Then: Has the room decided, and named who takes the decision?

Decision

Has the room decided, and named who takes the decision?

Discussion without a decision is not an answer. A decision says what is done and who does it. If the room is still on the previous decision when the next inject is due, the pack's default allows up to 5 more minutes, asks for the decision, then gives the new inject.

• Yes: Was the decision wrong, or taken by someone not entitled to it?
• No, or not clear: Climb the prompt ladder, one rung at a time

Step

Climb the prompt ladder, one rung at a time

Wait first, 2 minutes in the pack's default: silence is often thinking. Then restate the inject and the decision; ask the person whose role holds it; ask what the plan says; give the phase's spare inject; finally state the fact the scenario needs, note that the decision was not reached and move on. The scribe records every rung after the first, because a decision reached only after a prompt is rated at most Partly met.

• Then: The scribe logs the decision, who took it and any prompt

Decision

Was the decision wrong, or taken by someone not entitled to it?

The best findings often come from a decision taken by the wrong person, or a rule nobody remembered. A facilitator who corrects it in the room hides the finding. The facilitator's job here is to leave the judgement to the observers and the report.

• Yes, or not sure: Let the scenario run on the decision as taken
• No: The scribe logs the decision, who took it and any prompt

Step

Let the scenario run on the decision as taken

Do not correct it and do not say it was wrong. Let the scenario run on what the room decided, and let the scribe and observers record who decided. If it would end the exercise, ask 'who holds that decision under the plan?' If the room turns to blame, restate the no-blame rule and move to the next role.

• Then: The scribe logs the decision, who took it and any prompt

Step

The scribe logs the decision, who took it and any prompt

One line per decision: exercise time, scenario time, the inject, the decision in the words used, who took it, and whether a prompt came first. Observers write what happened as fact, with the time and the objective, without names or opinions, and note strengths as well as gaps. The log is the record every rating rests on.

• Then: Is the session more than 10 minutes behind the timing card?

Decision

Is the session more than 10 minutes behind the timing card?

The facilitator compares the clock with the timing card, and at the break decides what to do if the session is late. Rushing the last injects to cover everything produces ratings nobody can trust. Late, in the pack's default, means more than 10 minutes behind.

• Yes: Drop an inject whose objective another inject also tests
• No: Are there injects left to give?

Step

Drop an inject whose objective another inject also tests

Choose it at the break, and tell the observers which one. A phase that overruns loses an inject, never the hot debrief. An objective the session never reaches is rated Not tested and carried to the next exercise. It is never counted as met.

• Then: Are there injects left to give?

Decision

Are there injects left to give?

Keep going to the end of the last phase. If a phase finishes early, use its spare inject or a discussion question, and do not start the next phase early: the observers need the break. If a real incident needs anyone in the room, stop with the stop phrase; untested objectives are rated Not tested.

• Yes: Give each inject at its minute, read as written (back to an earlier step)
• No: Start the hot debrief on time, the same day

Step

Start the hot debrief on time, the same day

In the pack's run of show it starts at +145 on the session clock, whatever has happened, and takes 30 minutes. Each participant gives one thing that worked and one to change; observers give a first view per objective on the evaluation scale; the facilitator reads back the likely findings. Nobody defends a decision. A gap that would break a legal clock or a decision right today goes to the sponsor the same day.

• Then: Rate each objective from the log, not from how the room felt

Outcome

Rate each objective from the log, not from how the room felt

Observers rate against the approved success statement, in order. Not reached is Not tested. Wrong, not taken, or breaking a rule, a legal clock or a decision right is Not met. The right person, in time, without a prompt and in full is Met; anything less is Partly met. Where observers disagree, take the lower rating. The after-action report fixes the rating.