When the Host Drops. How a Backup Presenter Keeps the Room Alive
The host will drop. Plan the second laptop.
A backup presenter is a named person with access to the same live session assets, sitting on a second machine, ready when Wi-Fi dies or the primary laptop freezes. The join code on the wall should not change. The poll mid-vote should not reset. If our plan is “we will email the file,” we do not have a plan.
AV crews already think this way. Two playback machines. A switcher. Reddit threads on ProPresenter redundancy are about A/B outputs, not hope. Corporate town halls pretend the CEO’s MacBook is eternal. It is not.
Quizzes, polls, and wheels should not live only on the CEO’s MacBook. If the engagement laptop is still up when the slide laptop dies, the vote can continue. If both things lived on one machine, we are rebuilding in public.
What actually fails in the rooms we work
The failure is rarely the platform in the abstract. It is a dongle. A sleep setting. A forced OS update. A presenter who closes the lid. Remote CEO on hotel Wi-Fi. The poll is open. Two hundred phones are waiting. The screen goes black.
If the only copy of the session lives on that laptop, we rebuild. People leave. HBR has already told you meetings are expensive. Restarting a live vote in front of a packed room is how we light that money on fire.
Pre-assign, then prove it in rehearsal
- Name the backup in the run of show. Not “someone from comms.” A person.
- Invite them to the smart folder by email before rehearsal. They see the quiz, the poll, the wheel. They do not get the primary’s password in a Slack DM.
- On the backup laptop, open the host view during rehearsal. Confirm they can advance. Confirm they can close a poll.
- Put the join code on a printed card and on the confidence monitor. If we must rebuild, at least the code is not only in the dead laptop’s RAM.
- If the primary drops, the backup speaks one sentence: “We are still live. Keep your phones on this code.” Then they drive.
We do not claim a secret one-click “transfer control” product switch. That is not on our first-party list. What we do ship is folders we share, and live-host sessions a second operator can run when they have been invited. Rehearse the handoff. Do not invent a feature in the blog.
Keep the audience path stable
Phones should not download a new app because the host changed. QR and PIN stay. If we reset the session, we reset trust. People think they did it wrong.
Live poll and live quiz should already be in that folder so the backup is not hunting Kahoot for the quiz and Slido for the Q&A while the CEO is frozen. One folder. One host UI.
Rehearsal can run small. The real room might not. Size it before show day. A lockout during failover is a second outage.
What the backup is allowed to skip
They are not there to ad-lib the strategy story. They keep the vote open or closed as planned. They pin the next question. They do not start a new icebreaker to fill panic. Panic is visible. Competence is a closed poll and a next slide.
If the primary returns, we do not bounce control back and forth. Finish the segment. Debrief later. Two cooks will double-close the poll.
AV can still fail the output. A backup presenter cannot fix a dead LED processor. They can keep the data on phones honest while video recovers. Say that out loud so the room is not staring at a black wall thinking the vote died.
Aftercare is part of failover
Archive the session. Notes: what dropped, which machine, how long. Put a spare dongle in the kit. Turn off sleep. Disable auto-update on show machines. This is unglamorous. It is why the second town hall does not repeat the first.
NIST’s guidance on readiness is for systems. Treat the live-host laptop as a system. Two people. Two machines. Shared folder. One join code. That is the whole pitch, and it is operational, not a cartoon of a magic failover button.
When the host drops, the backup already has the folder. The room never learns the password. The poll stays the poll.
Hotel Wi-Fi is a special case. The primary is on a keynote from a guest network. The backup should be on a tether or a wired port the venue gave us in writing. Two people on the same congested SSID is not redundancy. It is two ways to fail together. Test the backup path at the same hour as the show, not at 7 a.m. when the hall is empty.
If the live quiz is mid-timer, we do not restart it to “be fair.” The people who already answered are the record. Restarting punishes them and teaches the room that a glitch erases work. Close the question. Show what we have. Move. Fairness at that point is honesty about the glitch, not a redo.
Write the fail line in the script so nobody improvises comedy. Comedy during an outage is how we look unprepared. “Keep your phones on this code. We are switching laptops.” That is the whole line. Then switch.
Folders are the access model, not a chat thread of screenshots. Screenshots of a PIN on a wall go into Slack and stay there for a year. Invite the backup. Uninvite contractors after the event. The next town hall should not inherit last quarter’s agency login.
If we have a spinning wheel segment after the poll, the backup should have spun it once in rehearsal. A wheel that only lives in the primary’s cache is a prize we cannot award. Same folder. Same rule as the quiz.
Redundancy is unglamorous until the lid closes. Then it is the only thing the room will remember, and the only proof the backup was real.