Hand over the incident, not another mystery

It is 15:50, and the outgoing incident lead is tired. The new lead has five minutes before the next storage decision and ten before the customer update. What they need first is not the story from the beginning. They need to know what happens next.

In this fictional Dispatch incident, exports are still delayed into the second hour. A submission cap remains active, and storage specialists are investigating. A good handoff lets a fresh person continue that work without accidentally undoing a control, repeating a failed attempt or treating a proposed rollback as something already done.

Start with the controls still in place

Put current customer impact and active controls at the top of the record. Name their owners, the pending decisions and the next technical and customer updates. Link the detailed history below. The incoming lead can then inspect a disputed assumption without making the whole team retell the incident first.

Include the changes that never happened. “Rollback discussed, not authorized or executed” describes a different system from “rollback completed at 15:20.” The single word “rollback” could mean a proposal, a permission or a completed change. Failed attempts belong here too, along with the conditions under which they should not be repeated.

Maya has five minutes to give Rowan the right picture

At 15:50 UTC, outgoing lead Maya hands coordination to Rowan. Region A exports remain delayed. New submissions are capped, and Jules owns that cap. Jules will check queue age at 15:55.

Imani, the storage specialist, has proposed a regional repair. At 15:55, Imani will present its scope, compatibility evidence and rollback plan. The primary approver is the storage duty owner; the alternate is the storage incident delegate. Both are simulated role cards held by the facilitator, not real contacts. If the duty owner is unavailable, Rowan will obtain the delegate's explicit decision. The customer update is due at 16:00. Worker rollback is still unapproved and unexecuted.

Rowan reads it back: “I own coordination from 15:50. Jules retains the cap. The next decision is storage repair authorization after Imani's 15:55 evidence. I will contact the storage duty owner, then the incident delegate if unavailable, and send the 16:00 update.”

Maya confirms the transfer in the response record. In the exercise, the role cards supply the people to contact. In real use, those contact routes must be verified.

Everyone needs to know who is leading now

The outgoing lead names the replacement and the time coordination changes hands. The replacement states the next decision, the action owners who continue in place and the update deadline. Announce the transfer in the response channel so the rest of the team follows the same lead.

This transfers coordination, not credentials or permission to accept every risk. Check that the new lead can reach the authorized service and dependency owners. Outstanding approvals remain outstanding after the handoff.

“The repair is ready. Who can approve it?”

Add that news at the moment of transfer. The incoming lead should be able to name the contact, say whether the cap stays in place and identify the measurement that would show whether the repair helped.

A read-back can catch a consequential mistake while it is still only a sentence. Suppose Rowan says, “I will remove the cap after the repair.” Maya can correct that immediately: Jules still owns the cap, and removal needs its own authorization and recovery check. The storage decision does not silently include a second change.

If the incoming lead still needs the outgoing person for basic context, improve the record and obtain backup help. Do not use the gap as a reason to keep an exhausted person in charge indefinitely. For a specialist handoff, use the same approach: ask the replacement to explain the current query or repair, its target and the next expected result.

A fresh lead may not be the only relief needed

Check the technical responders too. A new coordinator cannot compensate for a critical specialist who also needs a replacement. Identify who needs relief and whether the replacement understands the work already underway. Give those operational tasks owners.

For practice, let a partner take over using only the document and a short conversation, then write the next update. Compare it with the known state and fix any ambiguity. Working alone, take a break before simulating the read-back. Call that a desk review, not a test with a second person.

Leave Rowan ready for 15:55

Keep the accepted transfer and its corrections in the record. There will be time for a fuller retrospective. For now, Rowan knows which decision comes next, Jules still holds the cap and the rollback has not run. Maya can step away without taking those facts with her.

Sources

Google's incident-response guidance discusses incident roles and handoffs. The Dispatch staffing scenario is synthetic.