ARTICLE
Organising hybrid participation with dembrane
Most hybrid participation sessions aren’t really hybrid. There is the room, where the energy is: flip charts, coffee, people leaning into conversations at tables. And then there is the grid, a wall of faces on a screen, politely watching. The online participants were promised participation. What they got was a livestream with a chat box.
This isn’t anyone’s fault. Facilitators are stretched thin as it is, and giving the online group a real conversation of its own means breakout rooms, which means more parallel conversations than any team can follow. So the online half gets simplified: watch the plenary, type your thoughts in the chat, thank you for joining. The people who made the least effort to be there are heard the least. Usually they’re also the parents, the caregivers, the people who live an hour away. Exactly the voices the session was supposed to include.
We think that if you make listening easier (in any context), it will happen more.
It doesn’t matter how people connect
dembrane does one thing that turns out to matter a lot for hybrid: it is agnostic to how people are connecting. A conversation at a table in the room, a breakout room on Zoom or Google Meet or Teams, an interview, even a street conversation recorded on the way in: if it’s recorded into your dembrane project, it’s part of the session. Every conversation lands in the same place, gets the same features, and flows into the same report.
That changes the design question. You stop asking “how do we include the online people in the room’s conversation?” and start asking “how do we make sure everyone has the opportunity to join a great conversation, and how do we capture all that richness?” The first question has no good answer. For the second we made this hybrid assembly guide.
One hybrid evening, (assembly required)
Maybe you’ve run some variant of this: a hall with a livestream pointed at the stage, an online plenary with slides, a Mentimeter word cloud if you’re lucky. Breakout conversations where the best things of the night were said, and got distilled into 3 word post-its.
Here is the same evening with dembrane underneath it. Say 120 people signed up for an evening town hall about the new neighbourhood plan: 80 in the hall at ten tables, 40 at home on Zoom.
What you need
One dembrane project with the participant portal set up: one QR code, which also works as a link
A hall with tables, one phone per table (participants’ own phones are fine, data isn’t saved to their device)
Your usual video call software with breakout rooms enabled, and a laptop streaming the plenary into it
One co-host who owns the online room
No extra hardware, no AV crew, no note-takers
What you do
The week before, thirty minutes. Create the dembrane project, write the conversation starter participants will see, print the QR code for the tables, and paste the portal link into the calendar invite for the online group.
19:00, plenary. The welcome, the presentation, the questions from the floor. The hall listens, Zoom watches the stream. Nothing changes here.
19:20, break out. In the hall: each table scans the QR code, taps record, puts the phone in the middle. Online: your co-host opens eight breakout rooms and drops the portal link in the chat. Everyone opens it on their own device, next to the call. Within a couple of minutes, all eighteen conversations are being captured.
While people talk. Every conversation, in the room and online, appears in your live overview. Participants get socratic follow-up questions that pull their conversation deeper. You walk the floor. Your co-host can hop between rooms based on what they read in the live conversation overviews.
20:30, plenary close. When the chairs scrape back into a circle, ask dembrane what themes are emerging across all eighteen conversations, and read the first patterns back to everyone, the hall and the grid at once. The online rooms are in there, with the same weight.
The next morning. One report. Table four sits next to breakout room six. Send it to everyone who came, whether they were in the hall or at home.
That’s the whole assembly.
The only genuinely new step is number three, and it’s a QR code.
Choices and gotchas
How can we have online and offline participants interacting? In theory, people in the room can also join the video calls. Not hard, might take some getting used to.
Everyone scans, or one recorder per room? Online, the most robust option is everyone opening the portal alongside the call. Fifty participants in ten rooms means fifty recordings, and dembrane deduplicates and merges them in the backend into one stream per real conversation (you won’t see the merging in the interface yet, that view is on its way). The simpler alternative is one designated recorder per room, as long as they know they’re it.
The headphones gotcha. If the person recording wears headphones, the portal hears only their voice. For now, the recorder should play the call out loud: video call on their laptop, speaker on, portal on their phone recording the laptop next to it.
Upload-after as a safety net. Zoom and Meet can record breakout rooms, and you can upload the audio to dembrane afterwards. Note that breakout rooms are not recorded automatically: someone in the room, host or participant, has to start the recording manually. It works great as a fallback, and it’s better than losing the conversations. But you give up the live layer: no socratic questions during the conversation, no live overview, and Verify works best when each conversation has one main recording from the start.
Explore in parallel is fine. If four people in one room use the Explore feature at once, they get four different threads. That’s not necessarily a problem, it often makes the conversation richer. Just know it’s happening and decide whether one person drives it per table, or let everyone run it.
What you get on the other side
The street interview from the morning sits next to breakout room six sits next to table three. The socratic questions worked the online conversations just as hard as the in-person ones. And when you send participants the results, the people who joined from their kitchen table can find their own thinking in it, because they were actually in the data, not in an appendix summarising the sparse chat.
Communities are already running sessions this way. RadicalxChange runs hybrid events with dembrane in the mix, and within the EU AI4Deliberation consortium we’re piloting large-scale online deliberation with partners like ActionAid, where breakout audio flows straight into a dembrane project next to everything else.
What we’re building next
The setup above treats your video platform as a given. We’re also working on the other end of the problem. dembrane is the official host of the EU deployment of Frankly, an open source video platform for small-group dialogue and deliberation, developed by the Berkman Klein Center for Internet & Society at Harvard University. On Frankly, events can run without a host: participants see the framing content and flow straight into breakout rooms, thousands at a time, and every room’s audio lands in dembrane automatically. No separate QR codes, no designated recorders, no headphone gotchas. It’s in closed beta now, and it’s coming soon.
If you run large sessions and want to be one of the first to try it, get in touch! And if what you actually want is not a new video platform but dembrane plugged into the video tool you already use, we genuinely want to hear that too. That opinion is shaping what we build next.
