Search Console insights

Moz/Semrush-style keyword intelligence, computed entirely on the Search Console data you already own — no third-party keyword tool, no per-query API bill. Five reports turn your synced history and one live pull into the questions those tools charge for: what's about to rank, what's under-clicked, what's competing with itself, what each page ranks for, and what moved.

It builds on the read-only Search Console integration and its history sync. If that page's seo-pro:gsc-sync has been running, these insights work with zero extra API cost for three of the five surfaces.

Prerequisite

The three snapshot surfaces read the persisted seo_gsc_metrics history, so schedule seo-pro:gsc-sync (see Search Console → history) first. The more days you have synced, the deeper the trend comparison goes.

The five surfaces

1. Striking-distance keywords

Queries whose impression-weighted average position sits in 5–20 — page-one adjacent — ranked by impressions. These are the "one push to page one" wins: a little more relevance or internal linking often moves them into the clicks.

2. CTR opportunities

Queries that rank well but are clicked below the expected rate for their position. Each query's actual CTR is compared to a blended industry CTR-by-position curve; the ones far below their expected rate — with real impressions — are title/description rewrite candidates, sorted by the estimated missed clicks they're leaving on the table. This list is the natural input to the AI meta suggester: it hands you the exact queries worth rewriting for.

3. Cannibalization

Queries where two or more of your URLs compete for the same term. Splitting one query across several pages divides your authority and confuses Google about which page to rank — this surfaces the overlaps so you can consolidate or differentiate them.

4. Query clusters

The queries each page actually ranks for, grouped by page — the page's real topic footprint in Google's eyes. Useful for spotting a page that's drifting off its intended topic, or one that's quietly ranking for a valuable term you never targeted.

5. Trend vs previous period

The biggest movers — clicks, impressions, position and CTR — for the current window versus the equal span immediately before it. Position is only compared when a query had traffic in both periods (there's nothing to compare a brand-new or fully-dropped query against).

Where the numbers come from: live vs. snapshot

Each surface reads whichever source answers it correctly at the lowest cost. The persisted history can't reconstruct which query paired with which page (it stores each dimension separately), so the two surfaces that need that pair are the only ones that go live — and they share a single cached request.

Surface	Source	Why
Striking-distance	Local snapshot	needs per-query position + impressions, already in your synced history — no API cost
CTR opportunities	Local snapshot	same owned data; the expected-CTR curve is a static benchmark, not a lookup
Trend deltas	Local snapshot	needs genuine day-by-day history, which is exactly what the sync stores
Cannibalization	Live (query × page)	the query→page pairing isn't stored, and persisting every pair would multiply your storage
Query clusters	Live — shares surface 3's fetch	same pair data, grouped by page instead of by query

So a visit to the Insights page costs at most one Search Console request, cached for search_console.cache_ttl seconds. The pair surfaces are live on purpose: cannibalization and clustering are point-in-time questions where you want the current picture, and Search Console's quota is generous for one cached pull per view. The snapshot surfaces never touch the network.

In the dashboard

With the Filament plugin installed, Search Console Insights appears under the SEO nav group (only when the integration is enabled). It's strictly read-only: every surface renders as a section, the snapshot surfaces show a "sync your history" hint when empty, and a failed live pull for the pair surfaces renders a sanitized inline notice — never a blocked page.

Configuration

Everything lives under search_console.insights in config/seo-pro.php. The defaults are sensible; tune the thresholds to your site's scale.

'search_console' => [
    // ...
    'insights' => [
        // Rolling window (days) the snapshot surfaces aggregate over,
        // anchored to the latest synced day.
        'window_days' => 28,

        // Max rows a surface returns to its panel section.
        'max_rows' => 50,

        // (1) Striking distance: impression-weighted position in [min,max]
        // with at least this many impressions.
        'striking_distance' => [
            'min_position' => 5.0,
            'max_position' => 20.0,
            'min_impressions' => 30,
        ],

        // (2) CTR opportunity: queries ranking at/above max_position, with at
        // least min_impressions, whose CTR is at least min_gap_ratio below the
        // expected curve.
        'ctr_opportunity' => [
            'max_position' => 10.0,
            'min_impressions' => 50,
            'min_gap_ratio' => 0.30,
        ],

        // Optional override of the expected CTR-by-position curve
        // (position => percent). null uses the built-in blended curve.
        'ctr_curve' => null,

        // (3) Cannibalization: a query with this many URLs each drawing at
        // least min_impressions.
        'cannibalization' => [
            'min_urls' => 2,
            'min_impressions' => 10,
        ],

        // (4) Query clustering: queries per page above min_impressions.
        'clustering' => [
            'min_impressions' => 10,
            'max_queries_per_page' => 15,
        ],

        // (5) Trend deltas: surface queries with at least this many
        // impressions in either period.
        'trend' => [
            'min_impressions' => 20,
        ],

        // The shared live (query,page) fetch for surfaces 3 + 4.
        'pair_days' => 28,
        'pair_row_limit' => 5000,
    ],
],

Expected-CTR curve

The CTR-opportunity curve is a heuristic blended from published organic CTR-by-position averages — it's a yardstick, not a claim about your specific site. A query flagged here is a candidate to review, not a proven defect. If you have your own measured curve, drop it into insights.ctr_curve as a position => percent map.

See also

• Search Console — the read-only integration + history sync these insights read
• White-label reports — period-over-period movers in the branded PDF
• AI assist — rewrite the titles/descriptions the CTR surface flags