GA4's Data Retention Setting Is Probably Deleting Your Data at 2 Months
GA4's event-level data retention defaults to 2 months, and most properties never get it changed to the 14-month maximum. Here is exactly what that setting controls, what it does not affect, and why the gap goes unnoticed until someone needs a year-over-year Exploration report.
GA4's Data Retention Setting Is Probably Deleting Your Data at 2 Months
GA4's data retention setting controls how long user-level and event-level data is stored before automatic deletion, with two options: 2 months or 14 months. Unless a property owner has explicitly changed it, most GA4 properties are still running on the shorter default - meaning any Exploration or funnel report reaching back further than roughly 60 days is working with data that no longer exists, silently.
Key takeaways
- GA4's retention setting offers two choices - 2 months or 14 months - and defaults to the shorter option unless someone has explicitly changed it in Admin settings.
- The setting only affects Explorations and funnel reports; GA4's standard aggregated reports (including primary and secondary dimensions) are unaffected and retain data regardless of this setting.
- Age, gender, and interest data is capped at 2 months of retention regardless of the general setting - even on a property set to 14 months.
- Extending the retention period applies retroactively to already-collected data that hasn't yet been deleted, so changing the setting later still helps, but any data already deleted under the shorter window is gone permanently.
- Because standard reports keep working normally, the gap is invisible in day-to-day use - it only surfaces the moment someone tries to build a custom Exploration comparing further back than the retention window allows.
Teams that rely primarily on GA4's standard reports for daily and weekly monitoring can go a long time without noticing this setting exists, because nothing about their regular workflow is affected by it - the gap only becomes visible during exactly the kind of ad hoc, deeper analysis (a year-over-year cohort comparison, an unusual funnel investigation) that Explorations are built for.
Data retention setting: the GA4 Admin configuration controlling how long user-level and event-level data persists before automatic deletion - distinct from, and much shorter than, how long GA4's standard aggregated reports remain queryable.
Exploration: GA4's ad hoc, custom-report-building interface (funnel analysis, path exploration, cohort analysis) that draws on the underlying event-level data directly affected by the retention setting - as opposed to the pre-aggregated standard reports, which are not affected.
BigQuery export solves GA4's own retention ceiling with raw, unsampled data that persists as long as the dataset does -- GA4 to BigQuery Export covers what to verify without writing SQL yourself.
Why the gap between "reports work fine" and "Explorations are missing data" goes unnoticed
The operational pain this creates for anyone who eventually needs to run a genuinely custom analysis: standard GA4 reports (the default dashboards most teams check daily) pull from pre-aggregated data untouched by the retention setting, so a property sitting on the 2-month default shows no visible symptom in normal day-to-day use - conversion trends, channel reports, and acquisition dashboards all populate correctly regardless of the retention setting.
The first moment the gap becomes visible is almost always a specific, ad hoc analytical need: building a cohort comparison against data from eight months ago, running a funnel Exploration that needs to reach back further than the retention window, or investigating an anomaly that requires user-level event detail beyond the retained period. At that point, the missing data isn't flagged with an obvious warning - the Exploration interface simply can't return results for the period beyond retention, which can read as a bug or a misconfigured report rather than what it actually is: data that was already deleted before anyone went looking for it.
An uploaded contact list's row count and its usable matched audience size are two different numbers once LinkedIn's match rate is applied -- LinkedIn Matched Audiences covers what most B2B accounts half-configure.
Turn scattered analytics into one clear picture
Every source in one brief. The whole picture. Your decision.
14 days free · no credit card
Why the age/gender/interest exception matters for audience analysis
The ICP problem this creates for teams doing longer-term audience composition analysis: even a property correctly configured to the 14-month maximum still has age, gender, and interest data capped at 2 months specifically - a narrower exception that applies regardless of the general setting, and one that's easy to miss since it behaves differently from every other data type on the same property.
The practical implication: any analysis that depends on demographic or interest-category breakdowns over a longer historical window (a year-over-year audience composition shift, a longer seasonal demographic trend) cannot be built from GA4's own retained data alone, even on a maximally-configured property - that specific data category needs an external export or a different data source if it's needed beyond a 2-month lookback, since GA4 itself will not retain it longer no matter how the general setting is configured.
What to change now, and what changing it does not fix retroactively
The ICP problem this creates for a team discovering this gap for the first time: the instinct is to immediately change the setting to 14 months and assume the problem is solved, but that only protects data collected from this point forward - it does not restore anything already deleted under the shorter window.
The practical sequence: change the retention setting to 14 months in GA4 Admin now, regardless of whether an immediate need exists, since there's no meaningful downside to the longer window and the cost of discovering this gap during an actual analysis need (rather than proactively) is real lost historical context. GA4's modeled conversions have their own separate, unrelated gap in what gets estimated versus counted - worth checking alongside retention as part of the same broader GA4 configuration review. For any analysis that genuinely requires retention beyond GA4's own 14-month ceiling, export the relevant event-level data to BigQuery on an ongoing basis - BigQuery export is unaffected by GA4's own retention setting and is the standard practice for teams that need longer-term event-level access than GA4's own interface provides.
Prooflytics ingests GA4 channel-level data into the daily briefing from GA4's standard reporting layer, which is unaffected by this retention setting - a property's retention configuration does not change what shows up in the daily briefing's own channel trend view, since that pulls from the same unaffected aggregated layer GA4's own standard reports use.
Bottom line
- Check your GA4 property's data retention setting now - most properties are still on the 2-month default, and it only takes a few clicks to change to the 14-month maximum.
- Standard reports are unaffected by this setting; the gap only shows up in Explorations and funnel reports reaching beyond the retention window.
- Age, gender, and interest data is capped at 2 months regardless of the general setting - a separate exception worth knowing about for longer-term audience analysis.
- Changing the setting protects data going forward; it does not restore anything already deleted under a shorter window.
- Book a walkthrough to see how Prooflytics pulls GA4 channel trends from the same aggregated reporting layer that's unaffected by this retention setting.
Frequently asked questions
How do I check my property's current retention setting?+
In GA4, go to Admin > Data Settings > Data Retention, under the property's data collection settings - the current setting (2 or 14 months) is shown directly there, along with the option to change it.
If I change the setting to 14 months today, does old deleted data come back?+
No - data already deleted under the shorter setting is permanently gone. Changing the setting protects data collected from the change forward, and also extends retention for any already-collected data that hasn't yet reached its deletion point under the old setting, but it cannot restore anything already removed.
Does this affect my standard acquisition and conversion reports?+
No - standard aggregated reports, including primary and secondary dimension reports, are explicitly unaffected by this setting according to Google's own documentation. The setting only affects Explorations and funnel reports specifically, which draw on the underlying event-level data directly.
Is 14 months long enough, or should I always set up BigQuery export?+
For most standard year-over-year analysis needs, 14 months (GA4's own maximum outside GA4 360) is sufficient. Teams that regularly need multi-year event-level comparisons, or that need the 2-month-capped demographic data retained longer, should set up BigQuery export regardless of the standard retention setting, since it has no retention ceiling of its own.
You can read independent reviews of Prooflytics on G2 and compare it to other marketing intelligence platforms in the category.
Turn scattered analytics into one clear picture
Every source in one brief. The whole picture. Your decision.
14 days free · no credit card
Continue reading
Google Ads Data Retention Cut to 37 Months: What to Archive Before It Disappears
Starting June 1, 2026, Google Ads deletes granular reporting data older than 37 months on a rolling basis. Monthly and annual data stays for 11 years. Here is what you lose, what survives, and the exact steps to archive before the cutoff.
GA4 to BigQuery Export: What a Marketer Actually Needs to Know (Without Writing SQL)
BigQuery export solves GA4's two-month retention problem and gives raw, unsampled event data - but marketers usually delegate the technical setup and then can't evaluate whether it's actually working. Here is what to check without needing to write a query yourself.
How to Set Up GA4 Correctly in 2026: The Task Assistant Checklist
Google Analytics 4 now includes a Task Assistant that guides property setup through six structured categories, launched in May 2026. For marketing teams configuring a new property or auditing an existing one, this guide covers what each category contains, why it matters, and what the most commonly skipped steps cost you in data quality.
Marketing Data Latency: Why Yesterday's Dashboard Number Isn't Final Yet
Most marketing platforms backfill and revise conversion data for days after it first appears. Here is why the number you see this morning is a preliminary estimate, and how to avoid making a budget call on data that hasn't settled yet.