Post-event teardown

The show is over, the wall comes down, the team is exhausted. The instinct is to pack up and worry about cleanup later. The right move is to spend 30 minutes doing teardown deliberately while the context is fresh — both because cleanup gets harder a week later, and because lingering access from contractors becomes a security flag the longer it persists.

This page is an opinionated teardown checklist.

Order of operations

Rough order, with reasoning:

1. Take final photos — handoff documentation and your own retrospective material
2. Unlink signs from the event — releases tier slots and prepares fleet for the next deployment
3. Archive the event — locks editing, frees concurrent-event slots
4. Revoke contractor access — removes the external-tech invitations and event-team memberships
5. Rotate Tailscale auth keys — if any contractor had the key, replace it
6. Power down hardware — last, deliberately
7. Run a retrospective — whatever-was-painful gets fixed before the next event

1. Final photo handoff

Walk the floor one more time and photograph every wall as it stands at end of show. This is for:

• Customer retrospective — show what their walls looked like during the event.
• Your own retrospective — what worked visually, what didn't.
• Insurance / disputes — proof the signage was up at the time of teardown if there's any later question.
• Marketing material — if the customer is OK with it, photos of branded walls are useful for proposals to similar customers.

A 5-minute walkthrough with a phone camera is enough.

1.5 — Working with the customer during teardown

Teardown is the customer's last impression of the show. The decisions here shape next-year's contract.

Final walkthrough with the customer

• Walk the floor together; confirm content was correct end-to-end
• Surface any issues you've already fixed — transparency reads as competence
• Note any feedback for the retrospective; don't address it in the moment

Equipment handoff

• If the customer owns any kiosk hardware (rare but possible), hand it back inventoried
• Document what's in your kit vs theirs for next time

Contractor wind-down conversation

• Tailscale access — when it'll be revoked. Post-teardown, not during. (Revoking mid-teardown leaves the customer locked out if anything needs adjusting during the breakdown itself.)
• DisplaySync access — same timing. Org membership stays until teardown is fully complete.
• Any ongoing post-event work the customer needs (data export, analytics handoff) — agree on owner and timeline now.

Post-event communication

• Confirm when they'll receive the retrospective doc (typically T+3 to T+7).
• Open the door for next event: when do they want to plan the next conversation?

2. Unlink signs

From the dashboard, open each sign's detail page in the event and choose Unregister (per-sign action).

What this does:

• Wipes the device-link fields from the Sign DB record (so the kiosk can be re-claimed elsewhere)
• Emits sign:unregistered to the connected kiosk
• The kiosk clears its local org/event/URL state and returns to the QR screen
• The dashboard transitions the sign to Unlinked state

The Sign records remain in the org with their names and locations — convenient for the next event in the same venue, where you can Flow B link the same hardware to fresh records.

If you want a deeper cleanup, Delete each Sign record after unlinking — frees the slot toward your tier's maxTotalSigns cap.

3. Archive the event

Dashboard → Event detail → Archive event.

Archiving:

• Transitions the event from completed → archived
• Locks all writes — no more sign claims, content changes, team edits
• Frees the concurrent-event slot toward your tier's cap
• Keeps everything readable forever for retrospectives

Archived events are reversible — you can unarchive if needed. The point is to make the active event count drop so your next show doesn't trip the concurrent events limit.

4. Revoke contractor access

Cleanup of any external people you brought in for this specific event.

Event team

Dashboard → Event detail → Team → remove anyone who shouldn't have continuing access:

• External AV techs brought in for this event — usually remove
• Stakeholders added for read-only — keep if they're long-term, remove if event-specific

Org membership

If you added someone at the org level for this event (not just event-team-level), drop them from the org as well:

• Dashboard → Organization → Members → Remove

This is the most-frequently-skipped teardown step. Stale org members accumulate and become a security flag — every additional member is potential access if their account is compromised.

Pending invitations

Dashboard → Organization → Invitations → revoke any unaccepted invites for this event's contractors. They can be re-invited next time without leaving stale invite emails working in their inbox.

5. Rotate Tailscale auth keys

If any contractor had access to your Tailscale tailnet — even read-only — rotate the auth key they could have onboarded devices with.

From the Tailscale admin console:

1. Mint a new auth key (same flags as the old one — reusable, pre-authorized, tagged tag:displaysync-sign).
2. Update your image / USB recovery kit / provisioner config to use the new key.
3. Revoke the old key. Existing kiosks already on the tailnet keep working (their node identity is cached, doesn't depend on the original auth key).
4. New kiosks (clones, fresh provisions) onboard with the new key.

Rotation cadence:

• Always after a high-security or contractor-heavy event — before the contractor's exposure window grows.
• Quarterly minimum even without specific incidents.
• Immediately on any suspected leak — published auth keys are credential leaks.

6. Power down deliberately

Don't yank the power on every kiosk at once — that fires a flood of offline notifications and looks like a real outage.

Suggested order:

1. Push a "thanks for coming" splash to every wall before doors close (optional, looks polished).
2. Walk the venue and power off in route order, same as load-in.
3. Pack hardware in labeled cases. The kiosk that goes back into the case marked "LOBBY-WEST" goes back to that location next time.

If you've already unlinked signs (step 2), powering them off later doesn't matter much — the kiosks just go offline and stay offline until next reclaim.

7. Retrospective

Once you're back at the office, run a 30-minute retro with the team that ran the event. Capture:

• What broke — the actual incidents, with timestamps.
• What didn't break that you expected to — useful for confidence-building.
• What you'd change in the image / process / venue communication for next time.

Anything in the retro that should change a default in your image or your process gets done before the next event, not "later." See Live event checklist → What to fix between events.

Special cases

Recurring annual events

If you'll do this same event next year:

• Don't archive immediately — keep the event in completed state for ~2 weeks while customer questions and post-event reports happen.
• Do delete contractor access immediately — those are still rotating-population accounts.
• Save the run-of-show document with annotations for next year.
• Save the captured Windows image version that ran on the production fleet — useful as a baseline next year, even if you re-image with newer software.

Failed events

If something went badly enough that you're going to revisit it:

• Don't archive yet — archiving locks the audit log scrolling, but the data is preserved either way.
• Pull all logs off every kiosk while they're still online. See Log locations → Quick reference.
• Capture timestamps in your retro — Sentry events and dashboard audit log are your timeline.
• Then proceed with normal teardown.

Customer hands you the fleet

If the customer keeps the kiosks (you handed them off rather than rented them):

• Don't unlink and don't archive — they'll continue using the same setup.
• Transfer org ownership if needed (Profile → Org → Transfer ownership).
• Document credentials — Tailscale auth key, kiosk user passwords, content URL ownership.
• Brief the new owners on the Operations section.

See also

• Live event checklist — the pre-event checklist that pairs with this teardown checklist
• Claiming signs → Replacement hardware — the unlink-and-re-link pattern
• Roles & permissions — what to revoke at which level
• Tailscale integration — auth key mechanics