Beginning June 1, 2026, Google Ads applies a 37-month retention period to granular performance statistics, while higher-level data can remain available for up to 11 years. Advertisers and agencies that need long-term daily, weekly, segmented, or historical analysis should export and store the required data on a recurring schedule rather than assuming the platform will preserve full detail indefinitely.
What data is affected by the 37-month retention rule?
Granular performance statistics are subject to the 37-month retention window. High-level data can remain available for 11 years. Examples of granular analysis that may be affected:
- –Daily performance
- –Weekly performance
- –Device segments
- –Location segments
- –Network segments
- –Detailed campaign/ad-group history
- –Asset or product-level trends
- –Long-term query or keyword analysis
What is the difference between granular and high-level data?
| Data type | Example | Reported retention |
|---|---|---|
| Granular statistics | Daily/weekly and detailed segmented performance | 37 months |
| High-level data | Broader aggregated historical totals | Up to 11 years |
The operational problem is that a high-level total cannot recreate lost daily or segmented detail.
Who is most affected?
- –Agencies with long-term clients
- –Seasonal ecommerce brands
- –Businesses using multi-year forecasting
- –Teams building marketing mix models
- –Advertisers migrating accounts or platforms
- –Companies with audit or compliance requirements
- –Organizations using the Google Ads API
- –Businesses evaluating long-term brand and non-brand trends
What should advertisers export?
Core account performance
- –Date, account, campaign, ad group, campaign type, channel/network
- –Impressions, clicks, cost, conversions, conversion value
- –All-conversion metrics where relevant
- –Bid strategy, budget
Search data
- –Keyword, match type, search term where available
- –Search category/insight exports
- –Negatives, impression share metrics
Conversion data
- –Conversion action, primary/secondary status, counting method
- –Value, attribution setting, conversion lag data
Ecommerce data
- –Product ID, product title, category/product type
- –Impressions, clicks/interactions, cost, conversions, revenue/value
- –Merchant Center diagnostics snapshot where useful
What export cadence should be used?
| Dataset | Recommended minimum cadence | Reason |
|---|---|---|
| Core daily performance | Monthly | Maintains continuous daily history |
| Conversion action settings | Monthly and after changes | Preserves configuration context |
| Search terms | Monthly | Visibility and report definitions can change |
| Product performance | Monthly | Supports seasonal and SKU analysis |
| Asset performance | Monthly/quarterly | Preserves creative history |
| Full archive snapshot | Quarterly | Recovery and audit reference |
What should API users change?
Review any query that requests data older than 37 months with granular segments. PPC News Feed reported that invalid older granular ranges can trigger DateRangeError.INVALID_DATE.
- –Restrict granular queries to supported dates
- –Use stored historical data for older periods
- –Handle API errors explicitly
- –Backfill available history before it expires
- –Monitor for missing dates
- –Document API version changes
Data-retention readiness checklist
- –Long-term reporting use cases are documented
- –Required granular fields are identified
- –Automated export cadence is set
- –Stable IDs are stored (never only names)
- –Currency and time zone are stored
- –Conversion settings are archived
- –API date windows are updated
- –Missing-date alerts are active
- –Access and retention permissions are defined
- –Client handoff process is documented
- –Backups are tested
Related guides
Source notes
- PPC News Feed, "Data Retention Update In Google Ads," May 4, 2026.
- Google Ads official announcement and Google Ads API documentation for the data retention policy effective June 1, 2026.