GA4 Data Retention: Why 2 Months Isn't Enough (and How to Fix It)

Admin
Admin
May 22, 2026 · Reference
GA4 Data Retention: Why 2 Months Isn't Enough (and How to Fix It)

Quick answer: GA4's default user-level data retention is 2 months. That's almost certainly not what you want — any year-over-year comparison, multi-month cohort analysis, or extended attribution window is impossible at the default. Change it to 14 months in Admin → Data Settings → Data Retention. The change takes 30 seconds and applies going forward (it doesn't retroactively recover lost data).

This is the single most common configuration finding in GA4 audits — and the most frustrating, because once data has aged past the retention window, it's gone. There's no recovery. Below is what the setting actually controls, what breaks at the default, and how to plan around the 14-month ceiling.

In this guide
  1. What "data retention" actually controls
  2. What breaks at 2 months
  3. How to change it (30-second fix)
  4. Living with the 14-month ceiling
  5. Going beyond 14 months with BigQuery
  6. FAQ

What "data retention" actually controls

GA4's data retention setting governs user-level and event-level data — the granular detail behind your reports. It does not control aggregated standard reports, which GA4 retains indefinitely.

The practical line is:

  • Within the retention window — you can use Explorations, build custom funnels, segment by any dimension, and query custom event parameters. The full data is available.
  • Past the retention window — standard reports (Acquisition, Engagement, etc.) still show aggregated numbers, but Explorations return empty results, custom segments don't work, and parameter-level breakdowns are gone.

The default is 2 months. Maximum is 14 months. Both are user-level retention; event-level retention is automatically tied to it.

What breaks at 2 months

Year-over-year comparisons

This is the obvious one. If you want to compare May 2026 to May 2025, you can't do it in Explorations at 2-month retention — the May 2025 data has been purged. You'll see aggregated numbers in standard reports, but you can't segment by source/medium, audience, or any custom dimension.

Multi-month cohort analysis

Cohorts that track users acquired in January through their behavior in March, April, and May require all that data to be in-window. At 2 months, January users are gone by March 1.

Extended attribution and conversion paths

GA4's attribution lookback for acquisition is up to 90 days for paid channels and 30 days for organic. If your retention is 2 months (~60 days), you've already lost some of the upstream touchpoints before the conversion happens.

Audience definitions that use a longer lookback

If you build an audience like "users who viewed a product in the last 6 months but didn't purchase," that's impossible at 2-month retention. The audience will be empty or radically undersized.

Backfilling investigations

A finance team asks, "Why was our paid search conversion rate 20% lower in February?" If it's now May and you're at 2-month retention, you can't investigate — February's user-level data is purged.

How to change it (30-second fix)

  1. Open GA4. Click Admin (gear icon, bottom-left).
  2. Under the Property column, click Data Settings → Data Retention.
  3. Change User and event data retention from 2 months to 14 months.
  4. Leave Reset user data on new activity set to On (the default). This means the retention clock restarts when an existing user returns — usually what you want.
  5. Click Save.

Important: the change applies going forward. Data that has already been purged at 2 months is not recoverable. The sooner you make the change, the less historical pain you'll have later.

Living with the 14-month ceiling

14 months is the maximum GA4 will retain user-level data. That covers year-over-year comparisons (with one month of buffer) and most attribution windows. It does not cover:

  • True 24-month cohort analysis.
  • Long-form retention or LTV studies past 14 months.
  • Multi-year trend explorations at the user level.

If you need any of those, you have two options: aggregate reporting (using standard reports, which retain indefinitely but lose granularity) or BigQuery export.

One caveat about "Reset user data on new activity"

With this toggle On, a returning user's retention clock resets — so a user who visits monthly stays in-window indefinitely. With it Off, the 14-month clock is absolute from first interaction. Most businesses want it On; turn it Off only for strict compliance reasons (the shorter the user-data lifetime, the lower the GDPR exposure).

Going beyond 14 months with BigQuery

If you need longer-term granular data, enable the BigQuery export in Admin → BigQuery Links. Once enabled, every GA4 event is streamed to a BigQuery dataset, where it stays as long as you want (you control retention at the BigQuery level, not GA4's).

Key things to know:

  • BigQuery export is free up to 1 million events per day, then standard BigQuery pricing applies. For most mid-sized properties, the daily quota covers everything.
  • Export starts the day you enable it — there's no backfill. Enable it as early as possible.
  • BigQuery data isn't subject to GA4's thresholding, so for sensitive-data segments (small audiences, demographic breakdowns) it's the only way to see complete numbers.
  • If you change reporting identity in GA4, your BigQuery data isn't affected — you have the raw events to recompute whatever you need. See our reporting identity guide for what's affected by the choice.

Frequently asked questions

Does changing data retention affect historical data?

No. Changing the setting applies only to data going forward. If you've been at 2-month retention for a year and switch to 14, you don't suddenly get a year of data back — anything already purged is gone. Make the change as early as possible.

Why is 2 months the default?

Google's stated reason is privacy — minimizing the retention of user-level data is consistent with data-minimization principles in GDPR. In practice, the default is also a frequent audit finding, suggesting the privacy-vs-utility trade-off here isn't obvious to most users.

Is 14 months really the maximum?

Yes, for GA4's reporting interface. The only way to retain longer is BigQuery export — see the section above.

Will changing to 14 months violate GDPR?

Not on its own. GDPR requires that retention be tied to a documented purpose, not that it be as short as possible. 14 months for analytics and conversion modeling is a defensible business purpose, particularly when combined with Consent Mode v2 and a documented privacy policy. Consult counsel for your specific situation.

Does the 14-month setting affect Google Ads attribution?

Indirectly — Google Ads has its own conversion window (up to 90 days for view-through and click-through), independent of GA4's retention. But if you're using GA4-imported conversions for bidding, the upstream user data driving those conversions needs to be in-window for the analytics workflow to be debuggable.

Find data-retention issues automatically

Data retention is one of the first things a Snifflytics audit checks — along with reporting identity, time zone, currency, and the rest of the property-level configuration that quietly distorts everything downstream. Run a free audit to see your current settings, the issues they create, and the fix priorities for each.

Share this post: Twitter LinkedIn