According to Relevant Audience’s compilation, on 1 June 2026 Google Ads reduced data retention: hourly, daily and weekly data are kept 37 months, monthly, quarterly and annual data 11 years, and reach and frequency data 3 years. For most accounts, 37 months is enough for three years of comparisons, but accounts with long history and fine granularity should export older data, especially before big changes such as AI Max, Display and DSA. Confirm the values in Google Ads Help.
What was reported
Relevant Audience’s compilation lists, as confirmed, that on 1 June 2026 Google Ads cut data retention: hourly, daily and weekly data kept 37 months; monthly, quarterly and annual 11 years; reach and frequency 3 years. The source is a third-party compilation.
Google usually publishes these limits in Help. Confirm the numbers before planning based on them.
Why history matters
To decide well, you need to compare seasons: Christmas with Christmas, summer with summer, before and after a change. With 37 months of daily data, you have three full seasons. For niche campaigns or long cycles (property, B2B, training), you may need more.
If you are going to make big changes (migration to AI Max, end of Display in Search, DSA, call-only ads), having the “before” saved is the only way to prove the effect.
Data retention in Google Ads
37 months
for hourly, daily and weekly data, since 1 June 2026
According to a third-party compilation: monthly, quarterly and annual 11 years; reach and frequency 3 years. Confirm in Google Ads Help.
What to export
Export campaign, ad group, keyword, search term, ad, location, device, hour of day and conversion reports, by day, for the longest available range. Save in CSV or spreadsheet, with clear names and dates.
Also save a log of important changes (budgets, strategies, campaigns) to give context to the numbers.
How to export without mistakes
Set the longest possible date range, apply day segmentation, choose relevant metrics (cost, clicks, impressions, conversions, conversion value, CPA, ROAS) and download. For large accounts, split by periods to avoid row limits. Check that totals match the dashboard.
Store files in a safe place with a backup, and record the export date.
Having the “before” saved is the only way to prove a change's effect.
Who should own the history
The company. If your agency manages the account, ask for periodic copies of reports and raw data. Include this in the contract: quarterly delivery of exports and administrator access. If you change agency, the history can’t stay with the previous one.
Having the data is also protection against platform surprises.
Merchant Center and other reports
The compilation records that Merchant Center redefined its performance report on 24 August 2026, with history restated from 1 July. If you use Shopping, also save Merchant Center performance data before and after the change, and don’t compare series without checking the definition.
Same logic for Google Analytics and your e-commerce platform: your own data repository avoids dependencies.
A simple routine for SMEs
Quarterly, export the main reports to a dated folder. Once a year, also save a full export. Keep a “change log” sheet and a “results by month” sheet. With this, anyone new understands the account’s history in half an hour.
The routine costs little time and protects years of learning.
Questions for your manager
Where are the account history exports kept? How often? In what format? Do we have the “before” of big changes? Who owns the files?
If you want to set up an export and data archive routine for your account, with a results-by-month sheet, I can help create it.
An illustrative example: why three years may not be enough
A training company with a long decision cycle and peaks in September wants to compare this autumn’s campaign with the last four. With 37 months of daily data, it only sees three autumns. The example is hypothetical, but shows the point: if your business has strong seasonality, regular export is what lets you compare seasons beyond the limit.
It also helps in audits and disputes: with exported data, you can prove what you spent and obtained in a specific period.
Common export mistakes
Exporting only the summary and not the daily; exporting without day segmentation; forgetting search terms and ads; not saving metric definitions; saving only in the personal account of someone who left the company; and not testing whether the file opens. Each of these mistakes makes the archive less useful than it seemed.
Run a test: ask a colleague to open the file and answer a question (for example, how much did we spend in March 2024 on campaign X). If they can’t, the archive needs improvement.
Summary in three lines
Google Ads keeps daily data for 37 months, according to the compilation. Export the main reports and a change log now. The history belongs to the company, not the agency or the platform.
Frequently asked questions
How long does Google Ads keep data?
According to the compilation, 37 months for hourly, daily and weekly data; 11 years for monthly, quarterly and annual; 3 years for reach and frequency.
Do I need to export everything?
At least campaign, keyword, search term and conversion reports by day, and a change log.
When should I export?
Now, and then quarterly. Before any big change, always.
Sources consulted
Keep reading
- Targeted ads that delivered spyware to 45 devices: what the Discover-Tech case teaches those who buy online advertising
- Chrome now asks for approval before connecting to HTTP sites: what to do on your landing pages and ads
- Safari 27 blocks firms that turn emails into identifiers: what changes for remarketing and measurement