The morning after an event, someone asks "how did it go?" and the answer is usually a number chosen because it's the biggest. Registrations. "Views". Total minutes. None of those tells you whether the event did its job. These are the metrics we report, what they mean, and the traps in each.
Registered, attended, no-show rate
Attended divided by registered. Free virtual events typically see 35–55% attendance; paid and internal mandatory events much higher. A low rate points at reminders and timing, not content. Report the rate, not just the registrations.
Peak concurrent viewers
The largest number watching at one moment. It's the honest measure of live reach and the number to size infrastructure on next time. Different platforms and simulcast destinations count it differently — agree a definition before the show and note it in the report.
Average concurrent viewers and the attendance curve
More useful than the peak: the shape of the audience over time. Plot concurrent viewers against the run order. The curve shows which sessions held people and which lost them, and where the break went on too long. This graph is the most valuable thing in a post-event report and the one most platforms hide.
Average watch time and completion
How long the typical attendee stayed, and what share saw the end. For a 90-minute town hall, 60+ minutes average is strong. Watch time is the metric most sensitive to production quality — bad audio shows up here first.
Drop-off points
Where people left. Cross-reference with the run order: a cliff after a 25-minute unscripted speech is a content note; a cliff at the same moment on every device is a technical fault. Both are worth knowing.
Interaction rates
Questions per 100 attendees, poll response rate, reactions per minute, chat participants. Reported as rates, not totals, so events of different sizes can be compared.
On-demand views in the following 30 days
For many events the on-demand audience exceeds the live one. Report it, and report watch time on it too — an on-demand "view" is often 20 seconds.
The vanity numbers
"Views" on social platforms (counted after 3 seconds). Total minutes watched without the number of people (a huge number that means nothing). Impressions. Registrations without attendance. Use them for reach context, never as headline outcomes.
Setting the baseline
Numbers mean more against a comparison: last year's event, the industry norm, or the target you set in planning. Agree the targets before the show so the post-event conversation is about results, not definitions. The report template has a place for them.
| Metric | Good sign | Warning sign |
|---|---|---|
| Attendance rate | >45% (free) | <30% |
| Avg watch time (90-min event) | >55 min | <30 min |
| Completion | >40% | <20% |
| Questions per 100 attendees | >5 | <1 |
| Poll response rate | >40% | <15% |
| On-demand views / live attendees | >0.5 | <0.1 |