GA4 Data Retention Settings

9 min read

Learn

Quick Reading

Estimated 9 mins

Prereq

Intermediate

Basic ML concepts helpful

Interactive

Static Playbook

Static guide & reference tables

Retention Only Affects User-Level and Event-Level Exploration

GA4's Admin > Data Settings > Data Retention control (configurable to 2 months or 14 months) only governs how long event-level and user-level data remains available for ad-hoc analysis in Explorations — it does not affect standard aggregated reports, which retain data indefinitely, and it does not affect the BigQuery export, which is entirely independent and governed by your own BigQuery dataset's retention policy.

Why This Trips People Up

A common point of confusion: someone notices their standard GA4 reports (Reports > Life cycle, Engagement, etc.) still show data from 18 months ago, and assumes the retention setting doesn't work. In reality, those are pre-aggregated reports unaffected by the setting. Try to build a Funnel Exploration or Free Form exploration querying raw event data from 15 months back on a property set to 14-month retention, and it silently returns nothing for that excluded range — this is the setting actually taking effect, just invisibly to anyone not building custom explorations.

Data Surface	Affected by Retention Setting?	Actual Retention
Standard Reports (Life cycle, Acquisition, etc.)	No	Indefinite, pre-aggregated
Explorations (Free form, Funnel, Path, Segment Overlap)	Yes	2 or 14 months, per your setting
BigQuery Export	No — separate system entirely	Governed by your BigQuery dataset's own expiration policy
User-ID / cross-device stitching lookback	Yes	Bounded by retention setting

bigquery_retention_policy.sql

sql

-- Set your own long-term retention policy on the BigQuery dataset
-- that receives the GA4 export — independent of the GA4 UI setting
ALTER TABLE `project.analytics_123456789.events_*`
SET OPTIONS (
  partition_expiration_days = NULL -- NULL means never expire; set a number
                                     -- of days if you want automatic cleanup
);

-- Confirm current dataset-level default expiration
SELECT default_table_expiration_ms
FROM `project.analytics_123456789.INFORMATION_SCHEMA.SCHEMATA`
WHERE schema_name = 'analytics_123456789'

14 Months Is Usually the Right Default, With BigQuery as the Real Archive

There's rarely a good reason to choose the 2-month option over 14 months — the only tradeoff is privacy posture, not cost. Most properties should set retention to 14 months for maximum analytical flexibility in Explorations, and treat BigQuery export (which you control indefinitely) as the true long-term data archive for anything beyond that window, including year-over-year comparisons that exceed 14 months.

Interaction With Reset on New Activity

GA4 offers a toggle to reset the retention countdown on each new user activity (vs a fixed countdown from first collection). This specifically affects user-level data used for things like remarketing audience membership windows that reference user-level history — leaving it off means a user's data ages out on a fixed schedule from their first-seen event regardless of ongoing activity, which can be the correct choice for privacy-conscious configurations but surprises teams expecting continuously 'fresh' long-term user profiles.

Retention Setting Changes Are Not Retroactive Fixes

Changing the retention setting from 2 to 14 months does not restore data already purged under the shorter setting — it only extends retention going forward. If you need historical depth beyond what's currently retained, the BigQuery export (if it was enabled from early on) is your only recovery path; there is no way to recover already-expired Exploration-level data from Google's side.

What's Next?

Data retention and raw data access both tie into how much you can realistically do with the GA4 Data API outside the UI — including the quotas and limits that govern how much you can pull programmatically, which is worth understanding before building any custom reporting on top of GA4.

I build these systems professionally.

Whether it's a RAG pipeline, analytics migration, or AI workflow — let's talk.

Work With Me →