You open an exploration to compare last year’s social campaign with this year’s. The date range reaches back past a year, and the older stretch comes back blank. No error, no warning, just empty rows where last year’s sessions should be. Then you open a standard report for the same dates and the numbers are there. Both views read the same property, so one of them is being limited by something the other is not.
The comparison that comes back empty
The symptom is specific. An exploration with a date range longer than the property’s retention period shows nothing for the extra time. It does not fail or flag the gap. It simply returns no data for the part of the range that has aged out, which is easy to read as “the campaign did not run” rather than “the data is gone from this surface.”
The cause, on Google’s Data retention help page, comes down to one sentence: the retention setting does not affect standard aggregated reports, even if you create comparisons in them, and only affects explorations and funnel reports. That is why the standard report still answers and the exploration does not.
This post explains a mechanism. It ranks no tool and recommends no vendor. Everything below comes from Google’s own page, fetched on 1 October 2026.
What the setting deletes and what it leaves alone
Google defines the controls plainly: they give you the ability to set the amount of time before user-level and event-level data stored by Google Analytics is automatically deleted from Analytics servers. The retention period applies to data associated with cookies, user identifiers such as User-ID, and advertising identifiers.
| Item | What Google’s page says |
|---|---|
| Event data retention options | 2 months or 14 months for a Google Analytics property. The same period applies to key events data. |
| Standard aggregated reports | Not affected, including primary and secondary dimensions, even with comparisons. |
| Explorations and funnel reports | These are the surfaces the setting affects. |
| Age, gender and interest data | The two-month period is always applied, regardless of your settings. |
| Google-signals data | Maximum of 26 months, regardless of your settings. |
| Large and XL properties | Limited to 2 months. |
The page also lists longer options for Google Analytics 360 properties. This post does not cover them. If your property is a standard one, 14 months is the ceiling.
How deletion and changes actually work
Data that reaches the end of the retention period is deleted automatically on a monthly basis. Lowering the period is not instant either: any affected data is removed in the next monthly process. Google’s example is a change from 26 months to 14 months, after which anything older than 14 months is deleted in that next run.
Whenever the period is modified, Analytics waits 24 hours before implementing the change. During that window you can revert and the data is unaffected. That is a useful safety net if someone changes the setting by mistake.
The direction matters. Raising the period applies to data you have already collected and have not already deleted. It does not bring back anything that is gone. That is the whole reason to check the setting before a campaign starts, not after the comparison comes back empty.
A separate switch, Reset user data on new activity, controls what happens to user identifiers. Turned on, the retention period of the identifier resets with each new event from that user, so expiry becomes the current time plus the retention period. Google’s example: with retention at 14 months and a user who starts a new session every month, that identifier is refreshed monthly and never reaches the 14-month expiry. Turned off, data tied to the identifier is deleted after the retention period. Google states that this reset applies to user-level data only.
Google also gives the date range example directly: with retention set to 14 months, a date range of 14 months plus 1 day leaves that extra day unavailable in your reports.
A worked example on a campaign calendar
This is a hypothetical, an illustration of the rule and not a case study. No company is being described and no figures are invented.
Picture a social campaign that ran through March 2025. The property uses the 14 month event data setting. A marketer builds a year-on-year exploration on 1 October 2026.
- Fourteen months back from 1 October 2026 lands around 1 August 2025.
- March 2025 sits roughly five months before that line, so it is outside the window. Because deletion runs monthly, the exact moment those days vanish can lag, but the exploration is not guaranteed to show them.
- An exploration range from 1 March 2025 onward would therefore show nothing, or only partial data, for the early part of the range.
Now open a standard report and set the same March 2025 dates. Per Google, the retention setting does not affect standard aggregated reports, so that view is not limited by it. Same property, same dates, two different answers, and the difference is the surface, not the campaign.
What to export or archive before the window closes
- Check the setting. In Google Analytics, click Admin, make sure you are in the right account and property, then in the property column click Data Settings, then Data Retention. Google notes you need the Editor role to change it.
- Set event data retention to the longest option available to the property. On a standard property that is 14 months. Do it before the campaign, since raising it later cannot recover what is deleted.
- Save the standard-report figures a later comparison will need, in a document or spreadsheet, at the end of the campaign. A wrap report should store its headline numbers then, because an exploration rebuilt a year later may not have the underlying data.
- Record the UTM and campaign naming used at the time beside those numbers. A figure without its tagging convention is hard to reuse, and a UTM convention that survives a handoff covers how to write one down.
Keep one distinction in view. Saving a result, a number or a table, is different from saving the ability to re-slice the data later. Only the first is guaranteed to survive. Whether any export route preserves the second is not something Google’s retention page addresses, and this post makes no claim about export formats or BigQuery behaviour. The same habit of saving results before access closes applies elsewhere, as in what you can export before a social tool shuts down.
Where this sits next to attribution questions
Retention decides whether data can still be queried. Tagging and privacy tooling decide whether it was attributed correctly in the first place. A blank exploration is a retention problem. A campaign whose traffic landed in the wrong channel is a different problem. For that side, see why privacy-first analytics reports social traffic as direct.
Sources, date checked, and what this is not
Prices verified 1 October 2026. This post contains no prices. Every figure in it is Google’s, from its Data retention help page at https://support.google.com/analytics/answer/7667196, checked on 1 October 2026.
This is not Watchdog data and is not on the re-verification schedule, so Google may have changed the page since the date shown. Followedapp is published by the team behind RecurPost, which is covered vendor 34 under the same rules as every other vendor.
Check the setting this week
Open Admin, then Data Settings, then Data Retention, and note what your property is set to. If a campaign is about to start, set the longest option first, and put the end-of-campaign figures in a spreadsheet the day the campaign closes.
FAQ
Does GA4 data retention delete my standard reports?
No. Per Google’s Data retention page, checked 1 October 2026, the setting does not affect standard aggregated reports, even with comparisons. It only affects explorations and funnel reports.
If I raise the retention period, do I get deleted data back?
No. Google says an increase is applied to data already collected and not already deleted. Check the setting before the campaign, not after.
Sources
- Google Analytics Help, Data retention, fetched 1 October 2026 (HTTP 200)
- A UTM convention that survives a handoff, fetched 1 October 2026 (HTTP 200)
- Why privacy-first analytics reports social traffic as direct, fetched 1 October 2026 (HTTP 200)
- What you can export before a social tool shuts down, fetched 1 October 2026 (HTTP 200)