GA4 data retention: set 14 months today, and what the setting never protects

GA4 data retention is a single dropdown in Admin that decides how far back your explorations and funnel reports can look. A standard property can pick 2 or 14 months, and Google's help page says the choice affects only those reports. Set it to 14 months now, then export what you want to keep longer.

Everything below was read on Google's pages on October 25, 2026. Google changes these documents, so check the Data Retention screen in your own property before you rely on a number. This is a reference guide, not legal or compliance advice, and a retention dropdown is not a privacy policy.

GA4 data retention only touches explorations and funnel reports

The most useful sentence on Google's page is a short one: "The data retention setting only affects explorations and funnel reports." Your standard reports, the ones under Reports in the left menu, keep working over longer date ranges whatever you choose, because they run on aggregated data rather than on user-level and event-level records.

In our reading, that is why the complaint "I only see 2 months of data" is usually about the Explore section. A year-over-year comparison in a standard report is fine. A funnel exploration over the same year is not, once the setting has been left at 2 months.

Google's page does not say which value a new property starts on. The one third-party guide we read, from Analytics Mania in March 2025, says explorations show the last 2 months by default. We could not confirm that on Google's side, so look at your own screen rather than trusting either of us.

How to change GA4 data retention to 14 months

You need the Editor role on the property, according to Google. The steps on Google's page are short.

1. Open Admin and confirm the account and property in the top bar.
2. In the Property column, choose Data Settings, then Data Retention.
3. Under Event data retention, pick 14 months.
4. Decide on Reset user data on new activity (explained below).
5. Click Save.

Two details are easy to miss. Google says "Analytics waits 24 hours before implementing the change", and you can revert during that window. And the user-level setting sits next to the event one, so check both before you leave the screen.

If you run many properties, the Admin API exposes the same controls as three fields: eventDataRetention, userDataRetention and resetUserDataOnNewActivity. Its allowed values for standard properties are TWO_MONTHS and FOURTEEN_MONTHS.

What reset user data on new activity does in GA4

The API describes the option as: "If true, reset the retention period for the user identifier with every event from that user." Google's help page spells out both states. With it on, each new event from a user pushes that identifier's expiry out to now plus the retention period. With it off, data is deleted after the retention period whatever the user did since.

It applies only to user-level data. In our reading, leaving it on suits long-lived returning users, and turning it off gives you a hard cut-off that is easier to explain to a colleague who asks when data disappears. That is a judgment call, not a Google recommendation.

Does changing GA4 data retention restore old data?

Two sources we read disagree, and the disagreement is worth a section. Analytics Mania writes that the change "is not retroactive" and that you cannot select a longer period until enough new data accumulates. Google's own page says: "When you increase the retention period it is applied to data that you have already collected."

Our reading is that both can be true in practice. Google's sentence is about data that still exists. Anything already past the old 2-month window and deleted is not something a dropdown can bring back, and Google's page does not say otherwise. The same page says that when you reduce the period, "any affected data is deleted during the next monthly process", so deletion runs on a schedule rather than the instant a record expires.

We have not tested this on a property, so treat it as a reading of the documents. The practical advice is the same either way: change the setting before the data you care about is 2 months old, not after.

Who cannot get 14 months, and who can get more

Standard properties get the two options. Analytics 360 properties can also choose 26, 38 or 50 months, but per the Admin API reference those longer values are "Available to 360 properties only" and "Available for event data only".

There is also a downgrade path that catches busy sites. Google says "Large and XL properties are limited to 2 months", and when a standard property becomes Large or a 360 property becomes XL, the event-level setting "is automatically reduced to 2 months". Older event-level data then becomes inaccessible and is permanently deleted. Analytics warns property administrators when a property is on pace to hit the limit, and again when it does. If your property is close, export first.

How to keep GA4 data longer than 14 months with BigQuery

The way past the 14-month ceiling on a free property is to copy the raw events somewhere you control. Google's BigQuery Export does that, and it is the only route we found on Google's pages that keeps event-level history after Analytics deletes it.

According to Google's setup page, the link needs you to be an Editor on the Analytics property and to use an email address with OWNER access to the BigQuery project. You also need a valid form of payment on file in Google Cloud for the export to proceed, or you can use the BigQuery sandbox. The link itself is created from Admin, under Product Links, then BigQuery Links.

Three choices in that flow matter later:

• Data location. Google says it cannot be changed if the property already has a dataset, so pick it deliberately.
• Streams and excluded events. You choose which streams to include and which events to leave out.
• Daily, streaming or both. Daily writes one file per day for the previous day. Streaming makes current-day data available within a few minutes.

Three limits on the BigQuery route

The export is a copy, not a time machine, and Google's pages are plain about where it stops.

• It starts when you link. Google says data should start flowing "within 24 hours" of linking. Neither page we read describes a backfill of earlier history, and the BigQuery Export page says "Once you export data from Analytics to BigQuery, you cannot re-export it." We found no published backfill, so assume history begins on link day.
• Standard properties have a daily cap. "Standard properties have a daily BigQuery Export limit of 1 million events" for daily exports. If a daily export is paused because of the limit, Google says previous days' exports will not be reprocessed. Streaming has no event limit, but it is described as best effort, may contain data gaps, and costs $0.05 per gigabyte.
• Storage and queries are billed by BigQuery. The export is not free to keep. Google's page describes a free sandbox, with exports that exceed its limits incurring charges.

For a site under the daily cap, we would call that a small cost for keeping raw events past month 14. For a site over it, you are choosing between streaming's cost and gaps, or accepting that daily exports pause.

When owning the data beats extending GA4 data retention

Some teams would rather not depend on a vendor's deletion schedule at all. Upfynd lists 228 products in Analytics as of October 2026, and two of them speak to this problem directly.

Matomo describes itself as open source and offers hosting on-premise or in the cloud, with "100% data ownership" on its homepage. It also says you can "import your Google Analytics history and continue where you left off". We did not verify how that import behaves with GA4 data, so test it before you plan around it.

Jitsu is an open-source event collector, MIT-licensed and self-hostable according to its site. It sends events to a warehouse, and the destinations it lists include Snowflake, Big Query, Redshift, Postgres, MySQL and Clickhouse. If you already run a warehouse, a collector like this keeps retention in your hands, though you take on the pipeline. More warehouses sit in Databases & Infrastructure, and measurement tools aimed at marketers are in SEO & Growth.

We would not move off GA4 only because of the 14-month ceiling. The BigQuery export covers that for a sub-million-event site. A switch makes sense if the ceiling is one of several reasons you are already unhappy.

What to do this week

Open Data Retention on every property you own and set event data to 14 months, then check user data on the same screen. Link BigQuery on the same day for any property whose history matters, because the export only records from link day. If you build analytics software yourself, submit it to Upfynd so people comparing tools can find it.