What should an emergency access test prove?

The test should prove that the right people can recognise the event, obtain the right information, understand the runbook, and complete a safe recovery path. It should also prove that the team can stop a mistaken activation and rotate any material that was exposed.

Do not make the first test a production release. Start with a tabletop exercise using non-production systems and clearly labelled test credentials. Nominate a facilitator, an owner, a recipient, and an observer. Record the assumptions: which person is unavailable, which system is affected, what evidence exists, and what the team must accomplish.

What sequence should the drill follow?

Run the drill in six stages:

  1. Confirm the simulated activation criteria.
  2. Verify reminders, recipient details, and the safety period.
  3. Release a test delivery and confirm that only the intended recipient can access it.
  4. Follow the runbook against a sandbox or read-only environment.
  5. Rotate the test credential and close any temporary access.
  6. Compare the actual steps and timing with the written procedure.

Capture evidence without copying secrets into screenshots or tickets. Record timestamps, decisions, missing contacts, ambiguous instructions, and failed assumptions. If the test depends on a person improvising a missing step, update the runbook.

When is the plan ready?

A plan is ready for operational use when a qualified backup responder can execute it without the original owner, the business owner accepts the residual risk, and the organisation knows how to review and revoke the access created by the procedure.

Repeat the exercise after changes to identity providers, cloud accounts, billing, domains, staff, or vendors. Schedule a regular review, but also attach the runbook to change management so it does not wait for an annual date. A passing test is evidence about one version of the system at one point in time; it is not a permanent guarantee.