Analytics
Read event, audience, networking, and revenue metrics through the Core API.
All analytics resources are read-only. They return precomputed or request-time projections through the same JSON:API, scope, filter, and pagination contract as operational resources. Unless stated otherwise:
metric-versionidentifies the metric definition version;generated-atis the response generation time;time-zoneis the event’s IANA time-zone name;- ratios are returned as decimals (
0.72displays as72%); counts per 100 and averages use their stated units, and should not be converted again; and - a rate may be null when its denominator is zero.
Keep omitted fields separate from null values and measured zero. Preserve the resource’s metric definitions and denominators; do not assume similarly named event, segment and sponsor metrics can be combined. An unexpected rate should be investigated rather than clamped to 100%.
To turn these readings into a useful output, follow Create dashboards and reports. It includes reporting scopes, event use cases and copyable agent prompts.
Interpret metrics
Use the API’s field names in a source table alongside friendly labels. Define the denominator for every rate and the population behind each count.
| Metric or comparison | Interpretation to preserve |
|---|---|
| Joined registrants | Joining the event in Brella. It is not a count of physical arrivals. |
| Accepted meetings | Meetings in the accepted state. Completion or attendance needs separate evidence. |
| Session RSVPs, check-ins, tracked viewers and ratings | Separate signals. Display rating counts with averages; do not treat RSVPs or tracked views as verified room attendance. |
| Meeting-point utilization | Assigned meeting points divided by configured meeting points for a timeslot. It is not proof of occupied tables. |
| Meeting no-shows | Recorded skipped ratings on accepted meetings. The count is rating records, not necessarily distinct meetings or people. |
| Sponsor booth visits and leads | Recorded booth-page activity and captured leads. These do not establish physical footfall, qualified opportunities or sales revenue. |
| Event and segment engagement | Their definitions differ: event engagement uses accepted-meeting participation or chat notifications; segment engagement also includes other meeting participation and tracking. Do not substitute one rate for the other. |
| Date filters | Event dates select events; session dates select scheduled sessions. They do not automatically limit every metric to activity that happened during a reporting window. |
Keep null as “not available/not calculable,” and an observed 0 as zero. Display a ratio such as 0.72 as 72%; display accepted-meetings-per-100-joined-registrants as a count per 100, without multiplying again. Do not clamp unexpected rates to 100% or average percentages across unequal populations. Combine numerators and denominators only when their definitions and populations are compatible.
Custom audience lists can overlap; people can view several sessions, represent several sponsors or attend several event editions. Do not sum segment, session, sponsor or event counts and label them unique people. Sponsor meeting totals can also count representative participation rather than distinct event meetings. Keep the resource’s grain explicit.
Ticket amounts use currency-specific minor units. Keep currencies separate; for EUR, 1250 means EUR 12.50. Purchase creation filters and purchase completion filters affect different metric families. Email opens/clicks are recorded tracking signals, not proof a person read or acted on a message.
Event comparisons
Choose comparable event IDs deliberately. comparison_event_ids is required, and at least two IDs must remain inside the token’s event selection. Fewer than two returns an empty successful result; missing the parameter returns 400. Required filters are applied even when absent from the scope’s optional query-field selection. Inspect the returned comparison-event-ids to confirm the permitted cohort.
The benchmark compares the selected organization events, not the industry. previous-event-id means the most recent earlier start date within that cohort; it is not automatically the previous edition of the same event brand. Preserve null comparisons. A change from 65% to 72% is 7 percentage points, or approximately 10.8% relative growth; those are different statements.
Timeline charts
Validate timeline counts before using them for a reporting decision. A source review identified a possible same-day undercount in the current timeline projection; this has not been verified against a live deployment. Until verified for your data, use event summaries and session/timeslot comparisons for the default dashboard.
If using event timeline analytics, reconcile a known sample with multiple records created on the same day against permitted source records. The points are cumulative: do not sum them. A daily delta needs a verified preceding point; do not assume the first point starts at zero. Preserve returned dates and verify the day boundary rather than inferring it from time-zone. Omit a chart whose totals cannot be reconciled.