Raw Data Exports

Full GSC Data

Full GSC Data is the widest single export we produce: every query × landing page pair Search Console attributes to your site, rolled up over a 90-day window and joined to that day's site-level totals. It is intended for analysts who want the query-to-page relationship in one table rather than pivoting two separate reports together.

What it contains

  • Date — the snapshot date the 90-day rollup was taken on, not a traffic date.
  • Query and Page — the pair. One row per combination.
  • Impressions (90d), Clicks (90d), CTR (90d), Avg Position (90d) — that pair's totals across the trailing 90 days.
  • Site Clicks / Impressions / CTR / Avg Position (day) — the whole site's numbers on the snapshot date, repeated on each row so you can compute share without a second query.

Note what is not here: there is no device or country dimension, and the query × page figures are a 90-day rollup rather than one row per day. For those, use the dedicated reports linked at the foot of this page.

How it's produced

Search Console will not return query and page together at daily grain for a whole site, so we build a rolling 90-day query × page snapshot in the warehouse and stamp it with the date it was taken. This report reads the snapshot whose date falls in your selected window; if none does, it falls back to the most recent snapshot on or before your end date, and failing that to the latest one we hold — so the report is never simply empty.

Each row is then joined to the site-level totals for that same snapshot date. The dataset is denormalised on purpose: it is optimised for analytical pivots, not for production transactional use.

When to use it

  • Finding which queries a given page actually earns — and which pages compete for one query.
  • Building a custom dashboard in Looker Studio, Power BI, or your own BI environment.
  • Reconciling and validating any of our derived reports.

Caveats & limits

  • This is not every row Search Console holds. The export is capped at 5,000 query × page rows, ordered by impressions, so a large site sees its head and not its tail. Google also anonymises rare queries and applies its own row limits to the API, so neither we nor the export can claim completeness.
  • Query × page totals will not sum to the site-level columns on the same row. The site figures are one day's totals; the pair figures cover 90 days, and anonymised queries are missing from the pair rows but present in the site totals.
  • Position is per-row impression-weighted; for multi-row aggregations always re-weight by impressions, never simple-average.
  • Wide date ranges generate large exports; pull what you need.

Related reports