Profile
Understand every project-level field and status on Profile, what it changes, and what remains historical.
What you see on Profile
Profile is the setup record for one monitoring project. The header shows the project name and description, while the page summarizes its status, creation date, website and monitoring configuration. The Run monitoring control starts a new project run when your role allows changes. A first-run notice appears when active prompts exist but no completed or partially completed monitoring run is available yet.
Profile is not an analytics dashboard. Status describes whether the project is active, paused, or archived; it is not a performance score. Created is the database creation time, not the first monitoring date. Any recalculation notice reports background maintenance of derived metrics and should be allowed to finish before comparing recently changed setup.
Fields and their effects
| Field | What it controls | Historical effect | ||||
|---|---|---|---|---|---|---|
| Name and description | How the project is identified to account members. | Existing evidence is unchanged. | ||||
| Website | The owned site used for project context | channel matching | audits | products | and source-gap interpretation. | Old answers and sources are not rewritten. |
| Status | Whether the project participates in normal ongoing work. | Existing runs and exports remain available. | ||||
| Schedule | When recurring monitoring becomes due. | It changes future runs only. | ||||
| Project language and market context | Defaults used when building and interpreting the monitored prompt set. | Historical prompt snapshots retain their collected context. |
Runs, status, and recalculation
Run monitoring creates one execution obligation for each enabled prompt-provider pair. The number of answers expected therefore depends on active prompts and the providers selected on each prompt; it is not a fixed project count. A partially failed run may still supply reportable completed answers, but every dashboard denominator is smaller where executions failed or never completed analysis.
Some setup corrections can start metric recalculation. The notice is operational state, not a new score. Wait for completion, then compare the same filters and inspect answer counts. Starting another run does not repair an incorrect alias, market, or topic on evidence already captured.
Keep the project boundary trustworthy
- Use one project when the brand, competitor set, audience, language, and reporting owners share one measurement design.
- Split projects when combining markets or brands would create an average that nobody can act on.
- Record the effective date of website, schedule, brand, market, or prompt-set changes.
- Review Brands, Markets, Topics, and Tags after changing the profile because they define the evidence users will compare.
- Export important baselines before archiving a project or making a major measurement redesign.