Twelve people in a room, all being polite about it.
Whenever a group experiences something together — a concept, a cut, a workshop, a space — the only record of what happened is what people say afterwards. We measure what their bodies did while it was happening.
The gap
Self-report is the instrument the whole field runs on, and everyone using it knows its faults. People rationalise. They recall the ending and forget the middle. They are agreeable in a room with a moderator, and they describe what they think rather than what they felt.
Physiological measurement has existed for decades, but it has always meant one participant at a time, wired up, in a lab. What has not existed is a way to measure everyone in a real room, simultaneously, while the thing is actually happening. That is what we built.
Participants wear a light wristband. Nothing to download, no accounts, and no names or personal data collected at any point. We capture heart rate variability from everyone at once and compute the group's response in real time, then produce a timeline of how the room moved against what was in front of it.
Where it applies
A moderated group of eight to twelve, seated, indoors, still — the best measurement conditions we work in, and a group size our system covers completely rather than by sampling.
You get a second channel running alongside the transcript: where the room's response diverged from what the room said, and which moment produced it.
Audiences watching the same thing at the same moment is the cleanest case there is — the stimulus is shared and precisely timed, so the response maps directly onto the cut.
We have measured high-arousal screen content and captured sustained sympathetic activation, which is how we know the instrument reports engagement and not only calm.
Wellbeing and leadership programmes are delivered to rooms of people and justified with attendance figures and satisfaction surveys.
We report at programme level — by format, session type and time of day. Never on individual facilitators, and never on identifiable employees.
Designers build for a felt response and then measure dwell time and ticket sales, because nothing else was available.
We give you the room's physiological arc against the run of show — which passage held people, where the room came together, where it scattered.
What the group measure tells you
Averaging a room tells you very little. Ten people at wildly different states and ten people all in the same state can produce the same average, and those are completely different rooms. The useful question is not where the room is on average, but how far apart the people in it are.
That is what our Convergence Score measures: the spread of participant states at a given moment, computed continuously through the session. It moves when the room converges and moves again when it scatters, and it does that in real time rather than after the fact.
How tightly clustered the room is at a moment in time. A rising score means people are arriving at a similar state; a falling one means the room is spreading out.
Because it is cross-sectional, it responds immediately — you see the room tighten on a particular line, a particular cut, a particular silence.
Whether participants are moving together over a window of time, rather than simply sitting at similar levels. An established method in the neuroscience literature, computed here in real time at room scale.
A room can converge without synchronising, and synchronise without converging. They are different quantities and we report both.
Where this earns its place in applied work: a transcript records what twelve people said, in sequence, after the fact. The group measure records what all twelve did, simultaneously, while it was happening. When the two disagree — a room that said it was interested but scattered, or said little and converged hard — that disagreement is the finding, and it is not visible any other way.
And because the measure is continuous, it points at a timestamp rather than a conclusion. You are not told what the moment meant. You are told which moment to go back to.
What you actually receive
A live view during the session. The room's response as it unfolds, which matters most where someone is running the room and can still change what they do.
A group report afterwards. The room's arc mapped against your own timeline — stimulus by stimulus, segment by segment, or moment by moment, depending on what you are running.
An individual report for every participant. Anonymous, reached by a session code, with no account and no email. In research settings this is usually what participants find most interesting, and it costs you nothing to offer.
De-identified session data under a written agreement, where your own analysts want to work with it directly.
What we will not claim
Read this before the first conversation
- We measure autonomic response, not emotion. Arousal and vagal state are physiological quantities. Calling them joy, trust or intent would be an interpretation we cannot support, and we will not put it in a report.
- Physiology does not tell you what someone was thinking. It tells you when something changed, and therefore who to ask and when. The interview is still the interview — it is just aimed better.
- We do not predict behaviour. Nothing here forecasts purchase, recall or persuasion.
- A response is not a verdict. A strong reaction is not automatically a good one, and a settled room is not automatically an engaged one.
- Our group metrics are early-stage. The signal is validated against a clinical ECG reference. Our synchrony measure is an established method from the neuroscience literature. Our convergence metric is our own, patent-pending, and not yet independently validated. We say so rather than waiting to be asked.
Practical detail
Start with one session
The useful first step is almost always to instrument something you are running anyway, and see what the output looks like on your own material. Tell us what you are testing and we will tell you plainly whether this is the right instrument for it — including when it isn't.
Talk to us about your session