Google Search Console's 16-month limit: the official rule, and what it deletes

Google Search Console keeps 16 months of performance data. Google's documentation puts it in one sentence: "Search Console keeps data for the last 16 months." The line is on the official Analytics Help page about connecting Search Console, which adds that "reports in Analytics also include a maximum of 16 months of data." Linking GSC to GA4 doesn't extend the limit.

You can see the limit in the product. Set the Performance report to "Last 16 months" and look at the left edge of the chart. Your site's history goes back further than that edge, and Google has deleted everything past it.

The limit covers all Performance report data (clicks, impressions, CTR and position) in the UI and the API alike. Older rows are deleted on a rolling basis, with no warning and no final export. Whatever your site did this month will be gone 16 months from now. The Search Console data retention page lists where else the data can live and how long each option keeps it.

Quick answers about the 16-month limit

Does the Search Console API keep more than 16 months?

No. The API reads from the same store as the UI. It returns more rows per request, but no older dates.

When did Search Console switch to 16 months?

In 2018, when the new Search Console launched. The old Search Analytics report showed 90 days. Google hasn't moved the limit since, and has never said why it stops at 16 months.

Does Looker Studio keep Search Console data longer?

No. The Search Console connector queries the live API on every refresh, so dashboards stop at the same 16-month edge. The Looker Studio retention page explains why its cache doesn't count as storage.

Can you get deleted Search Console data back?

No. Once a date passes the 16-month mark, Google keeps no copy, and no tool can restore it. The only data that survives is data someone archived before it expired.

Does the BigQuery bulk export include older data?

No. The export starts collecting on the day you enable it, with no backfill of the months Search Console already holds.

What the limit breaks

Year-over-year comparisons go first. Comparing full calendar years, say 2025 against 2024, needs 24 months of data. You can compare a recent month against the same month last year, but push the question back more than a quarter and the baseline has already expired. The four-month ceiling on year-over-year comparisons works through the arithmetic.

Seasonality is next. One Black Friday is an anecdote. To separate a seasonal dip from an actual problem you want two or three complete cycles, and GSC hands you one, sometimes with a bite taken out of the far end. One cycle is a pattern of one, and it can't tell a seasonal dip from a decline.

Then there's core update forensics. When traffic drops and you suspect an algorithm update, the first thing you need is clean data from before it rolled out. If that update is more than 16 months old, the "before" no longer exists, and diagnosis becomes guesswork. Measuring core update impact takes that baseline plus earlier years of the same season.

The quietest loss is content decay. Pages rarely die overnight. They leak clicks over two or three years, and inside a 16-month window a page sitting at 60% of its former traffic can look flat and healthy. Tracking content decay means comparing each page with its own peak, and the peak is usually older than 16 months.

Four ways to keep your history

All of them share one hard constraint: nothing is retroactive. Each option preserves data from the day you switch it on, while Google keeps deleting from the far end in the meantime.

Manual exports

The UI exports what a report shows, capped at 1,000 rows per table. You can pull more through the API with a script or a Sheets add-on. This works for a single small site, if you remember to run it every month. In practice nobody remembers by month three.

Looker Studio

The official GSC connector gets mistaken for an archive a lot. It isn't one. It reads the live API on every refresh, so your dashboards show the same 16-month window and lose the same data. Fine for reporting, useless for retention.

BigQuery bulk export

The serious first-party option. Search Console can stream daily performance tables into a BigQuery dataset you own, including long-tail queries the API samples away. The catch: it needs a Google Cloud project, billing enabled, and SQL to get anything back out. It also starts collecting from the day you enable it. There is no backfill of the 16 months you already have.

An API archival service

Tools that call the Search Analytics API on a schedule and store the rows in their own database. On connect they backfill everything GSC still has, then sync daily from there. This is the category Month17 sits in: it archives your properties and serves the whole history over MCP, so you can query it from Claude or Cursor instead of a SQL console.

Whichever option you choose matters less than when you switch it on. Every week of waiting is another week gone from the far end of your window.