Tabletop Exercises Fail Because You Script the Ending
Most tabletop exercises are theater with a script everyone has already read. The scenario is announced weeks ahead. The right people clear their calendars. The injects are gentle, the timeline is generous, and everyone knows the exercise ends at noon with coffee and a slide that says “areas for improvement.” We call it a test. It tests nothing except whether people can perform competence in a room where the stakes are zero.
Real incidents don’t work that way, and the gap between the rehearsal and the event is exactly where programs fail.
You can’t rehearse decisions you already know the answer to
The point of a tabletop is not to confirm you have a plan. It’s to find the moment where the plan breaks — where two people think someone else owns the call, where the decision to pull the site offline needs an executive who isn’t reachable, where legal and comms want opposite things and nobody has the authority to break the tie. Those moments only surface under pressure and ambiguity. Script the ending and you engineer them out of existence.
An exercise where everyone knew what to do isn’t a success. It’s a sign the scenario was too easy. If nobody was uncomfortable, nobody learned anything.
Inject the things that actually go wrong
Real incidents are defined by their friction, so put the friction in. The person who owns the decision is on a plane. The monitoring you were counting on is the thing that’s down. The information is contradictory and arrives late. A regulator’s clock started before you noticed. A journalist emails for comment while you’re still deciding whether it’s real. None of these are exotic — they happen in nearly every serious incident — and almost no tabletop includes them, because they make the room squirm.
Change the injects mid-exercise. Take a key player out of the room without warning. Give the incident commander bad information and see how long it takes anyone to challenge it. The discomfort is the deliverable.
Measure decisions and time, not attendance
A good tabletop produces a short, painful list: the decision that had no clear owner, the authority that wasn’t delegated, the contact who couldn’t be reached, the step that took forty minutes because one person had to do it manually. That list is the whole value. If your after-action report says “the team responded well and demonstrated strong coordination,” you didn’t run an exercise. You ran a morale event.
Write scenarios you’re not sure the team can handle. Let people fail in the room, where failure costs nothing, so they don’t discover the gap for the first time at 3am when it costs everything.