The spreadsheet you rebuild every month, already built

Segments that survive.

Define brand, sections, languages and intents once as rules, regex included, and every report reads by them. Publish a page tomorrow and it lands in the right segment without you touching anything.

The whole long tail.

Past the 1,000 rows Search Console's own table shows: every query row Google's API returns, pulled day by day, searchable, filterable by intent and brand, and exported to CSV with no row cap or to Excel up to 200,000 rows.

Pivots that hold up.

Any dimension by any dimension with multi-level drill-down and subtotals, computed on the server so site size does not slow the interface. Ratio metrics are recomputed at every level rather than averaged, so a subtotal is never quietly wrong. Every figure also carries the window it covers and whether it is complete or sampled, which is what lets you defend it in a client meeting.

Cannibalization is a stated rule, run over the whole archive.

A query is listed once per language, and only when between 2 and 10 of your pages in that language appeared for it, those pages drew at least 50 impressions between them, no single page held 70% or more of those impressions, and the top two held at least half. The same term served by /de/ and by /fr/ is Google choosing a page per locale, and a page holding 95% is a clear winner, so the list keeps neither.

The rule runs over any range you pick, and the range reaches a long way back: the first sync copies in the ~16 months Google still holds, and from there the archive keeps a rolling 24 months, past the day Search Console deletes its own copy.