Google Ads Data Retention Cut to 37 Months: What to Archive Before It Disappears

Google Ads Data Retention Cut to 37 Months: What to Archive Before It Disappears

Starting June 1, 2026, Google Ads applies a rolling 37-month limit to granular reporting data: hourly, daily, and weekly campaign performance data older than 37 months is permanently deleted from the Google Ads interface, the API, and BigQuery export pipelines. Monthly, quarterly, and annual aggregates still retain for 11 years. If you have not exported your pre-2023 granular data, the window is closing now.

Key takeaways

1. Google Ads deletes hourly, daily, and weekly data older than 37 months starting June 1, 2026. Monthly and annual data is still retained for 11 years.
2. The 37-month window is rolling: every day the clock advances, another day of historical granular data becomes permanently inaccessible with no recovery path.
3. BigQuery Data Transfer Service is the highest-risk integration: backfill runs for dates older than 37 months will overwrite existing BigQuery rows with empty values, not skipping them.
4. Data that survives the cut: monthly, quarterly, annual aggregates; conversion action configurations; audience definitions; Smart Bidding model history.
5. Priority export order: accounts with multi-year benchmarking workflows, seasonal trend analysis, or long-cycle B2B attribution models should archive first.

What exactly changes on June 1, 2026

Google's official data retention policy draws a clear line between granular and aggregate data.

Affected (37-month rolling limit): hourly, daily, and weekly reporting data accessible via the Google Ads UI, Google Ads API, Google Ads Scripts, and BigQuery Data Transfer Service.

Not affected (11-year retention): monthly, quarterly, and annual data. Conversion action setup, audience lists, and Smart Bidding model health are not covered by this policy.

The phrase "rolling window" is important. This is not a one-time deletion of data before a fixed date. Every day after June 1, the oldest day of granular data falls outside the 37-month window and is purged. An account that runs the same comparison query in July 2026 and July 2027 will get different results from the same date range.

For teams that use month-over-month or year-over-year benchmarks, especially for seasonal campaigns, this means pre-June analysis must be exported before the window closes, not after.

01. The BigQuery risk most teams miss

The ICP problem this creates for Google Ads teams using BigQuery: the Google Ads Data Transfer Service for BigQuery does not simply stop populating old data. If a transfer is manually triggered for a report date older than 37 months, Google's developer documentation confirms that the existing data in the BigQuery table for that date will be overwritten with an empty value.

This means a well-intentioned analyst running a backfill job after June 1, 2026 can silently destroy historical data that still exists in the table, replacing it with nulls. The destruction is permanent.

If your Google Ads BigQuery export pipeline runs automated backfills or scheduled historical re-pulls, disable the backfill logic for dates older than 36 months before June 1. Then export those older ranges manually first.

02. What data to prioritize

Not all historical Google Ads data has equal operational value. Prioritize export in this order:

High priority: archive before anything else:

• Daily campaign performance (impressions, clicks, cost, conversions) for accounts with seasonal patterns that span more than 3 years
• Weekly keyword performance for accounts that use year-over-year search trend analysis
• Daily ad group and ad creative performance for teams that track creative lifecycle across multiple years

Medium priority:

• Geographic performance data by day or week if you run regional campaigns
• Device segmentation data for accounts where mobile/desktop split has shifted significantly over time
• Audience segment performance if you have multi-year remarketing data

Low priority (11-year retention, no urgency):

• Monthly and annual performance summaries
• Conversion action configuration history
• Budget change log (this lives in change history, not reporting data)

Prooflytics

Stop stitching platform exports together

Every channel in one brief — plus the memory of what each one actually drove.

Start free trialSee pricing

14 days free · no credit card

03. Four ways to export before the deadline

Option 1: Google Ads UI bulk export For smaller accounts or one-time archival: Reports in the Google Ads UI, set date range to the maximum available window, download as CSV. Repeat for each campaign, ad group, and ad entity level you need. This is slow for large accounts but requires no technical setup.

Option 2: Google Ads API For accounts with API access: query historical data via the Google Ads Query Language (GAQL), specifying date ranges that extend to the beginning of your account history. Use batch queries with pagination to handle large result sets. Target the campaign, ad_group, ad, and keyword resources with all performance metrics.

Option 3: Google Ads Scripts For accounts with Scripts access but no API infrastructure: write a script that iterates through monthly date ranges from account creation to today, exporting daily breakdowns to Google Sheets or Drive. The limitation is Sheets row limits. Use multiple sheets or Drive exports for large accounts.

Option 4: BigQuery scheduled export (new, not backfill) For accounts already on BigQuery: ensure your Data Transfer job is running without gaps and that the export schedule covers daily granularity. Do NOT run manual backfills after June 1. Instead, verify the existing export history covers the full period you want to retain. If gaps exist, fill them before June 1 using the BigQuery scheduled export with correct date parameters.

What the data shows: the Google Ads retention framework

The ICP problem this creates for performance marketers: most teams build benchmarks from rolling 12-month windows and rarely look further back. The 37-month cut feels distant until a quarterly business review requires a 3-year seasonal comparison: the data simply is not there.

By the 'Google Ads granular data retention cut' changelog, Google's change aligns with industry norms: Facebook Ads Manager retains granular data for approximately 37 months, and most third-party analytics platforms use similar rolling windows to control storage costs. Google's unlimited retention policy was an outlier, not a standard.

The operational implication is straightforward: any team that has historically used Google Ads data for multi-year benchmarking, whether for seasonal campaigns, long-cycle B2B attribution, or year-over-year ROAS trend analysis, needs to treat the historical export as a one-time migration task, not an ongoing workflow. Once the data is in a permanent store (a data warehouse, a Sheets archive, or a third-party analytics platform), it is safe from the rolling window.

Prooflytics surfaces Google Ads performance data in the daily briefing with a rolling 90-day operational window. For accounts that need multi-year benchmarking, connecting Google Ads to Prooflytics alongside a permanent archive export covers both the daily intelligence need and the historical audit trail.

04. Five things that do NOT disappear

Before treating this as a full account audit, clarify what is safe:

1. Monthly, quarterly, and annual aggregate data: Google's 11-year retention policy covers these. If your benchmarking relies on monthly summaries rather than daily breakdowns, no action is required.
2. Conversion action configurations: the structure of your conversion actions (what you track, how you count) is account configuration, not reporting data, and is not covered by this policy.
3. Smart Bidding model signals: Google's internal bidding models retain relevant signals as part of account history, separate from the reporting data API.
4. Audience definitions and membership: audience lists, remarketing configurations, and Customer Match lists are account assets, not reporting data.
5. Change history: budget changes, bid adjustments, and structural account changes are logged in change history, which has separate retention governed by account audit logs.

Bottom line

• Google Ads deletes granular daily and weekly data older than 37 months on a rolling basis starting June 1, 2026. Monthly aggregates survive for 11 years.
• The BigQuery risk is the most dangerous: manual backfill triggers after June 1 overwrite existing data with empty values. Disable historical backfill jobs before June 1.
• Export priority: daily campaign, ad group, and keyword data for accounts with multi-year seasonal or trend benchmarks.
• Data that does not require action: monthly summaries, conversion action configuration, audience lists, change history.
• For teams connecting Google Ads to Prooflytics: see Google Ads marketing analytics for the full integration setup and metrics available in the daily briefing.
• You can read independent reviews of Prooflytics on G2 and compare it to alternatives in the marketing analytics category.