Metric definitions
Ruphish separates recorded observations from summary progression. That distinction explains why a summary can show an opened recipient even when their timeline has no separate open event.
| Observation | What was recorded |
|---|---|
| Accepted/sent | SMTP acceptance |
| Opened | A tracking-image request |
| Clicked | A campaign landing-page visit |
| Submitted | A supported presence-only form event |
| Reported | A report notification |
Counts and progression
Section titled “Counts and progression”Raw event totals include repeated observations. Unique counts include each attributable recipient once for the selected event type. Progression counts infer that a clicked or submitted recipient has reached the opened stage, and a submitted recipient has reached the clicked stage. Acceptance and reports are never inferred.
Campaign analysis summary cards, the dashboard, and the campaign execution overview count each recipient at most once per category within a campaign. Recipient analysis cards and event history count all recorded observations, including repeats. A retry of the same submission delivery is deduplicated. Campaign analysis retains separately labelled event totals and activity charts.
Those inferences do not add timeline events or invent timestamps. Charts, event tables, and event exports retain recorded observations. Password presence requires an explicit positive boolean; submission alone does not mean credentials were entered.
Dashboard scope
Section titled “Dashboard scope”The dashboard selects recipients by their first recorded SMTP acceptance in the chosen period. Subsequent observations count through the server’s snapshot time. A repeated send does not move a recipient into a later cohort. The same address in different campaigns counts separately.
Opened, clicked, reported, and credentials-entered percentages divide by the selected sent-recipient count. A zero denominator displays a dash. The 7/30/365-day views compare absolute interaction counts with the preceding equal-length cohort, whose observations stop before the current period begins. All time has no preceding-period comparison.
Campaign-analysis scope
Section titled “Campaign-analysis scope”Analysis applies its recipient, event, source, and time filters before aggregation and pagination. Its lower time bound is inclusive; its upper bound is exclusive. These filtered observations can differ from the dashboard’s acceptance cohort. Don’t compare the two counts without matching their scopes.
Historical and automated activity
Section titled “Historical and automated activity”An imported reported flag without a corresponding report event is shown separately; no report timestamp is invented. Ambiguous legacy records cannot be assigned to a recipient by guesswork, so unique-recipient statistics exclude unresolved attribution while event exports can retain those observations.
Scanners, proxies, remote-image settings, and caching can create or suppress requests. Treat rates as observations under these constraints, not as a measurement of a person’s intentions or password validity.