GA4 data retention decides how long Google keeps user-level and event-level data for explorations and funnel reports. Standard properties can choose 2 or 14 months. Standard reports are not affected, so most businesses should pick 14 months and move on.
By Kristijan GjorgjioskiIndependent SEO consultantPublished 18 October 2026 · 6 min read
An exploration that stops at two months back is one of the most common GA4 surprises I see.
Someone builds a year-on-year comparison, sets the date range and gets eight weeks of data.
Nothing is broken. That is the GA4 data retention setting doing exactly what it was told.
This guide explains what the setting controls, what it leaves alone and which option to pick.
GA4 data retention is a property setting that controls how long Google keeps user-level and event-level data tied to cookies, user IDs and advertising identifiers.
After that period, the data is deleted automatically.
It is not a setting for how long your reports exist, and it has nothing to do with how GA4 counts sessions or engagement rate.
GA4 keeps aggregated data for its standard reports separately. Retention governs the detailed, row-level data that explorations and funnels are built from.
Google's data retention help page states it plainly: the setting only affects explorations and funnel reports.
This is where most of the confusion comes from, so here it is in one table:
| Area | Limited by retention? | Notes |
|---|---|---|
| Standard reports (Acquisition, Engagement, Monetization) | No | Primary and secondary dimensions and comparisons keep history |
| Explorations | Yes | Date ranges longer than retention show only data inside it |
| Funnel reports | Yes | Same rule as explorations |
| Key events data | Yes | Follows the same retention period |
| Age, gender and interests | Always 2 months | Fixed, whatever you choose |
| Google signals data | Max 26 months | Shorter retention wins if you set it |
| Google Ads data in GA4 | Rolling 36 months | Separate window, set by Google |
So a year-on-year traffic comparison in the standard Traffic acquisition report works fine on a 2-month property.
A funnel exploration comparing this spring with last spring does not.
In my audits, the teams hit hardest are the ones who rely on explorations for anything seasonal: retailers comparing peak periods, B2B firms looking at long sales cycles, and anyone building path analysis.
The options depend on your property type.
| Property | User-level data | Other event data |
|---|---|---|
| Standard GA4 | 2 or 14 months | 2 or 14 months |
| Analytics 360 | 2 or 14 months | 2, 14, 26, 38 or 50 months |
There is one exception worth knowing. Google limits Large standard properties, and XL 360 properties, to 2 months.
If a property crosses that threshold, event-level retention drops to 2 months and older event data is permanently deleted.
Google emails admins when a property is on pace to hit the limit and again when retention has been reduced.
That is rare for small and mid-sized sites. It is worth checking if your site sends very high event volumes.
For almost every business I work with, the answer is 14 months.
The reasoning is simple:
When might 2 months be the right call?
Even then, I would ask what decisions might need explorations next year. Retention cannot be applied backwards, so the cost of choosing short is paid later.
Retention is one of the first settings I check in a GA4 audit, because it takes a minute to change and the cost of leaving it grows every month.
Changing retention is safer than most people think, with one exception.
Increasing retention (2 to 14 months) applies to data that has already been collected and not yet deleted. So switching today keeps the last two months for the full 14, not just data from today onwards.
Reducing retention (14 to 2 months) deletes the affected data in Google's next monthly deletion run. That is the exception: deleted data cannot be recovered.
A few mechanics from Google's documentation:
That last step matters. Anyone reading an exploration later needs to know when the longer history started.
Below the retention period sits a switch that changes how user-level data expires.
It only applies to user-level data, not to event data.
I leave it on for most properties.
Switching it off can break cohort and lifetime explorations for your most loyal visitors, who are often the ones you most want to understand.
On a standard property, 14 months is the ceiling for explorations. If you need more, you have three options.
1. Rely on standard reports for long comparisons. Year-on-year traffic, key events and revenue sit in standard reports, which retention does not touch. Many questions never needed an exploration in the first place.
2. Link BigQuery. Standard properties can export raw events to BigQuery, with a daily export limit of 1 million events according to Google's BigQuery export guide. The tables live in your own Google Cloud project, so retention in GA4 no longer decides how long you keep them. The export only starts from the day you link it, so link early. Parameters you never registered as custom dimensions in GA4 are still in the export, which is handy for high-cardinality values like IDs.
3. Export key explorations on a schedule. A monthly export of your most important funnel or path report to a spreadsheet is crude, but it keeps a record.
For sites where the numbers feed budget decisions, I usually recommend BigQuery as part of a GA4 setup.
It is the only option that keeps event-level detail you can query later.
Retention only decides how long data is kept. It says nothing about whether the data was right in the first place.
If explorations look short, fix retention.
If they look wrong, the cause is usually elsewhere: duplicate tags, internal traffic, broken cross-domain journeys or key events firing on clicks.
My GA4 audit checklist covers those checks in order, and why GA4 data does not match explains the gaps you should expect against other tools.
Open Admin, find Data retention and look at the number.
If it says 2 months and nobody chose that on purpose, change it to 14 months now.
Every week you wait is a week of exploration history you will not get back.
If you would like someone to go through retention alongside the rest of your settings, events and attribution, my Google Analytics audit checks each one against how your business actually sells and ranks the fixes by impact.
No. According to Google Analytics Help, the data retention setting only affects explorations and funnel reports. Standard aggregated reports, including their primary and secondary dimensions, keep their history. If an exploration stops at two months back, the retention setting is the likely cause, but the same figures should still appear in the standard reports.
New properties I review almost always sit at 2 months, which is the shorter of the two options on standard properties. Google's help page lists the options but does not state the default, so check Admin rather than assuming. If you see 2 months and nobody chose it deliberately, change it to 14 months.
No. Once Google's monthly deletion process removes expired user-level and event-level data, it is gone. Increasing retention only extends data that still exists. That is why the setting is worth changing on day one, and why a BigQuery export is the only way to hold event-level history longer than the property allows.
Leave it on unless your privacy adviser asks otherwise. With it on, each new event from a returning user pushes back that user's expiry date, so explorations about loyal customers keep working. With it off, user-level data expires at the end of the retention period even if the person keeps visiting. It only applies to user-level data.
About the author
Kristijan Gjorgjioski
Independent SEO consultant · 6+ years in SEO · 60+ client websites worked on
I'm an independent SEO consultant with 6+ years in SEO. I run audits, migrations and content programmes for businesses in Europe, the UK and the USA, and I build the automation behind my own research.
Book a free 30-minute call. I look at your site first and tell you plainly what I would fix, and in what order.