> ## Documentation Index > Fetch the complete documentation index at: https://greatquestion.co/support/llms.txt > Use this file to discover all available pages before exploring further. # Panel & Participation Analytics > How the Analytics page counts panel size, net change, movement, health, contactable rate, utilization, and the invitation funnel. The **Analytics** page includes a set of reports about your research panel and the participation funnel. They answer questions like how big your panel is, whether it's growing, how much of it you can actually reach, how much of it you're using, and what happens to the invitations you send. Admins can access them from the **Analytics** item in the sidebar. ## How the panel numbers are counted The panel reports are built from a snapshot Great Question takes of your panel every night. A few conventions apply across all of them: * **The panel** is every candidate in your account that hasn't been deleted and didn't come from an external provider. Externally recruited participants are never counted. * **Monthly values are closing snapshots.** A month's value is the last snapshot taken in that month, not a sum or an average of the days. * **Gaps stay gaps.** A month with no snapshot shows as empty rather than zero, and no value carries forward across it. * **History starts at the first snapshot.** The series begins the first night a snapshot was taken; panel history before that can't be recovered. ## Panel reports ### Panel size How many candidates were in the panel at the end of each month. This is the whole pool, whatever state each candidate is in, not the part of it you can still contact. ### Panel net change How much the panel grew or shrank over each month: the panel at the end of the month minus the panel at the end of the month before. Negative means the panel lost more people than it gained. It's a net figure, so it doesn't say how many joined and how many left, only where the two landed. The first month in range has no month before it to compare against, so this series starts a month later than the others. ### Panel movement Candidates added to and removed from the panel in each month, counted on the day it happened. **Added** counts everyone who joined that month, including anyone since removed, so added minus removed matches the net change report. One caveat: movement counts on calendar months while net change measures between closing snapshots, so the two agree exactly only when a snapshot ran on the last night of both months. **Removed** means deleted. Unsubscribing doesn't remove anyone from the panel; it makes them unreachable, which shows up in the contactable panel report instead. ### Panel health The panel split by why someone can or can't be approached, counted as it stands right now. **Contactable** is the one to read first: opted in and not excluded by any [eligibility filter](/support/support/candidates/candidate-eligibility-settings). **Eligible** counts only the health filters, **Opted in** counts only consent, and the remaining buckets (over-invited, over-participated, unresponsive, no-showed, opted out, unsubscribed) name the filter that excluded someone. A person can be in more than one bucket at once, over-invited *and* unresponsive, for example, so the buckets overlap and don't add up to the panel. Read each one against the panel size, not against the others. As bars, this report shows the panel today; as a table, it shows the same buckets month by month from the nightly snapshots. ### Contactable panel The share of the panel that could actually be approached at the end of each month: opted in (subscribed, consented, and not set to never contact) and not excluded by any health filter. This is the number that matters for "can we run the study," where panel size only says how many names are on file. It falls when people unsubscribe and when health filters trip, so a panel that's growing and a panel that's reachable can move in opposite directions. ### Panel utilization The share of the panel invited to a study in each month: the candidates invited at least once that month, over the panel at the end of it. A few definitional details: * Counted from when the invitation was sent, or when the person asked to take part, not from when they were added to the study. * Someone invited to three studies in a month counts once. * Studies the product creates to carry a standalone incentive are left out, as are deleted studies and deleted participations. Because invitations are counted across the month while panel size comes from the snapshot that closed it, a month that gained and invited candidates after its last snapshot can read just over 100%. ## Invitation funnel The invitation funnel shows what became of the study invitations sent in the selected period, across four stages: **invited**, **applied**, **booked**, and **completed**. It's a cohort funnel: every stage measures the same invitations, the ones sent in the range, and counts how many of them ever reached that stage, whenever that happened. An invitation sent in March that leads to a completed session in April still counts in March's funnel. That's why a stage can never exceed the one above it, even when a study straddles the end of the range. **Invited** counts a participation the moment it's sent or requested. **Applied**, **booked**, and **completed** each count it once, however many times it was rebooked. Standalone-incentive studies are never counted, and deleted studies and participations are left out unless you include them. **Still need help?** Contact us at [support@greatquestion.co](mailto:support@greatquestion.co) — median response time is 19 minutes during [support hours](https://greatquestion.co/features/about-our-support-team).