get_growth_evidence
Read a reproducible paid, SEO, GA4 and GEO evidence bundle before making broad growth recommendations.
Parameters#
start_date,end_date: inclusive ISO dates, up to 400 days; required unlessevidence_sectionis given.site_source_id: selected site; defaults to the primary site.account_ids: optional paid account IDs, which must belong to the organization and be mapped to the selected site.page_url: optional exact normalized landing-page URL.evidence_section: optional<evidence_id>:<section>(for examplegev_…:paid) to read a stored section of earlier evidence instead of building new evidence.offset: character offset into that section; pass the previous page’snext_offset.
Result#
Returns the settled comparison window, source freshness, exact normalized query and landing-page matches, separate source metrics, acquisition coverage, sampled API citations, recent changes, linked work and bounded investigation leads. evidence_id and section evidence_refs come first and identify the persisted evidence used by recommendations; pass the refs verbatim as growth_context.evidence_refs.
The result is bounded to about 24,000 characters. Long lists keep their leading rows (campaigns and search terms by cost). truncated_lists gives the totals behind what was cut, and omitted_sections names any section left out entirely. Read the full stored section with evidence_section, which returns content (a page of the section’s JSON text), total_chars and next_offset (null on the last page).
No matching query row means unknown coverage. Search terms are incomplete. Current creative destinations establish observed exposure, not historical destination identity. Ad-platform conversions and GA4 session-acquisition key events must not be added together. Mention/citation samples, AI referrals and revenue remain distinct. Topic clusters and incrementality hypotheses require investigation; observed overlap never proves savings.
Organization dashboard permission is required. Unknown or foreign site/account IDs return validation errors.