Search Console Insights (Upload Import)

Search Console Insights analyzes your Google Search Console exports without requiring an API connection. You upload the ZIP, the module auto-detects the type (Performance or Index Coverage), computes prioritized insights deterministically, and lets the AI explain each recommendation in plain language with a concrete example.

rankion.aiWhat it does

• Two export types, auto-detected — Performance ("Performance on Search": queries, pages, countries, devices, search appearance, dates) and Index Coverage ("Coverage": indexed vs. not-indexed over time, critical issues). The type is detected from the file/column layout.
• Opportunity detection (Performance) — Striking Distance (positions 11–20 with upside), CTR outliers (high impressions, low CTR → title/meta opportunity), "seen a lot, barely clicked", branded vs. non-branded, position distribution, click trend.
• Error detection (Index Coverage) — indexation trend, 404s, redirects, "excluded by noindex", "crawled – currently not indexed", etc., prioritized by affected pages.
• Deterministic numbers + AI explanation — all metrics are computed in PHP (no hallucinated values); the AI explains each card and provides concrete actions with an example (e.g. a suggested title).
• Project assignment & filter — assign each upload to a project (on upload or later) and filter the dashboard per project.

rankion.aiWhen to use

• You don't have OAuth connected but still want to analyze your GSC data.
• Quick overview of quick wins (page-2 keywords, CTR opportunities) and indexation problems.
• Client/project-separated analysis across multiple properties.

rankion.aiUI workflow

1. Export — in Google Search Console, export the Performance or Index Coverage report as ZIP.
2. Upload — at /search-console, upload the ZIP (max 25 MB); optionally assign a project.
3. Analyze — KPIs, trend chart and insight cards appear; AI explanations are added in the background (ai queue).
4. Act — work the cards by impact; each card states a concrete next step with an example.

rankion.aiVia AI chat / API

The Rankion agent can use the module directly:

• GET /api/v1/search-console — lists all team uploads (optional ?project_id=). Returns each upload's type, property, date range, row counts and id.
• GET /api/v1/search-console/{id} — returns an upload's insights: deterministic metrics plus ai_summary and ai_actions (action + example) per card.

Typical chat request: "Look at the Search Console data for and help me improve it." → the agent lists uploads, picks the right one and fetches the insights.

rankion.aiImportant to know (data reality)

• The GSC UI export is capped at 1,000 rows per table and ~45% of queries are anonymized → totals are systematically understated. The module therefore shows relative opportunities, not exact totals.
• Cannibalization and query-page mismatch need query+page in one row (API/Bulk export only) and aren't reliably derivable from a plain upload.
• True index/coverage errors live fully only in the GSC index report; the upload surfaces the critical issue categories with page counts.