Session Facts

Session facts hold the observation-level record for a measured property. An Analyst can record a session date, observation date, source reference, outcome and the values used to discuss conversion. The parent property gives the fact a responsible owner and a definition. Without that link, an otherwise plausible conversion count cannot be placed in the property review.

Collection is page driven or manual. It deduplicates the consented facts for one property, records an operation run and numbered steps, and writes a state transition. A review decision is recorded when the collection needs approval. This trail matters when an analyst returns to a source after the campaign has changed.

• Observation. Each fact carries a sequence, description, date and source reference for investigation.
• Outcome. Pending, Accepted and Rejected distinguish an unresolved fact from a reviewed one.
• Property link. The parent source supplies the context for session, visitor, goal and conversion values.

Agree which external event source supplies session facts and how its source reference is preserved. Document the consent basis the team will use before importing or entering observations. Decide whether a session and a conversion arrive as separate facts or in a combined source record; the decision should be consistent so comparisons use comparable windows. The team should also agree who reviews Pending facts and which reasons justify a Rejected outcome.

A collection retry needs to refer to the same property and input. The run and step records let the Analyst see whether work already succeeded, failed or requires reconciliation. Do not substitute a newly typed count for a source fact simply because the dashboard has an empty value.

For a product landing page, the Analyst collects consented sessions after confirming the property and its consent basis. A fact marked Pending is reviewed against its source reference and observation date. If the record belongs to the property, the reviewer can move it to Accepted. If it belongs elsewhere, mark the fact Rejected and use its source reference to explain the decision during review. The collection run should show the same source revision as the property handoff.

Next, compare the accepted facts with the property’s session and conversion counts. A cohort comparison should follow review, with the run marked succeeded only after the values are read back. If that read-back fails, the run exception remains visible for investigation. This gives the next Analyst a concrete place to resume rather than an unexplained change in the plotted trend.

• Session Facts Review — Inspect outcomes and measured values beside their property.
• Pending Session Facts — Work the observations awaiting a decision.
• Collection Schedule — Check captured windows and freshness before comparing periods.

See collect sessions for the source handoff and compare cohorts for the verified outcome.