Roster coverage unavailable until its source loads.

Lead coverage unavailable until its source loads.

Joins use startDate; departures use explicit cancelDate, never lastUpdated. Missing dates suppress affected totals and retention. Deleted records and earlier unrecorded clients cannot be reconstructed. Current active counts come from client status, never Cronometer seats. The first homepage visit each Melbourne day saves an immutable aggregate observation in existing app storage; no names or raw IDs are stored. Private server storage also retains salted pseudonymous enrolled membership, allowing observed cohort comparisons; API responses expose aggregate counts only. Observations begin when this version is used, with no backfill. Historical aggregate-only observations cannot establish retention. Observed retention compares the same privately recorded starting cohort at two dated endpoints. Paused clients remain enrolled. Membership changes between visits may be missed; gaps and actual dates are shown.

Consults count next-week calendar events explicitly named consult or consultation. Other calls are excluded; unnamed consults cannot be inferred. Calendar truncation suppresses the total. A scheduled date is not the date the booking was made.

Lead-to-client conversion is unavailable: there is no durable lead-to-client identity link. Signed lead stages alone do not prove an enrolled client. Advertising cost, CAC and ROAS are unavailable without attributable advertising data.