GA4 Data Retention: Change This Before You Lose Data
The default setting quietly deletes your user level data. Here is what to change and what it does not cover.
There is a setting in GA4 that defaults to two months and quietly deletes your user level data.
It affects Explorations, audiences built on historical behaviour and any analysis that needs individual user data. Standard reports are unaffected, which is why nobody notices until they try to do something interesting.
Changing it takes about thirty seconds. Doing it today matters, because it does not apply retroactively.
Where it is and what to set
Admin, then Data settings, then Data retention.
Set event data retention to fourteen months. That is the maximum on the free version of GA4.
There is a second toggle for resetting user data on new activity. Turn it on. It means a returning user’s retention clock restarts, so your active users do not age out.
Save. That is the whole job.
What the setting actually controls
This is the part that causes confusion, so let us be precise.
Aggregated data in standard reports is kept indefinitely. Sessions, users, page views, conversions by day. That data does not expire.
User and event level data expires. That is what powers Explorations with custom dimensions, path analysis, cohort analysis and segment overlap.
So the symptom is not that your reports go blank. It is that Explorations refuse to go back further than two months, and nobody connects that to a setting they never saw.
Why the default is two months
Google set a conservative default for privacy reasons. It is a defensible choice for them and a bad one for almost every business using the product.
There is no downside to fourteen months for most organisations. If your legal or privacy position requires a shorter retention, that is a real reason to leave it, and it should be a documented decision rather than an accident.
Fourteen months is still not enough
Here is the uncomfortable part. Fourteen months means you cannot compare this Black Friday to the one before last. You cannot look at three year trends. You cannot analyse a cohort over a long lifecycle.
For a lot of businesses that is a genuine limitation, and the only real answer is to keep the data somewhere else.
Which is the BigQuery export. It is free, it keeps whatever you choose to keep, and it is the actual solution to this problem: setting up the GA4 BigQuery export.
Not retroactive, which is the urgent part
Changing retention from two to fourteen months does not restore data already deleted. It applies from the change forward.
So if you set this today, in fourteen months you will have fourteen months of data. If you set it in six months, you will have waited six months for nothing.
This is the entire reason it is worth doing immediately rather than adding to a list.
Check every property
Retention is per property. If you have several, check all of them.
Agencies and consultants: this is worth adding to your onboarding checklist. It costs nothing, it takes thirty seconds, and in a year it is the difference between being able to answer a question and not.
Related settings worth reviewing while you are there
The data settings section has a few other things that matter and get ignored.
Data collection. Google signals, which affects modelling and demographics but also introduces thresholding in reports. Worth understanding before turning it on or off.
Data filters. Internal and developer traffic. Covered separately: filtering internal traffic from GA4
Reporting identity. Blended, observed or device based. This changes your user counts meaningfully, and blended is the default.
Thresholding, since it comes up here
If you have Google signals enabled and you look at a small segment, GA4 sometimes hides the data entirely to protect user privacy. You see a warning icon and no numbers.
That is thresholding, and it is one of the reasons people go to BigQuery. Longer retention does not fix it. Only raw data access does.
The thirty second checklist
Admin, Data settings, Data retention.
Event data retention to fourteen months.
Reset user data on new activity to on.
Save.
Repeat on every property you own.
Turn on the BigQuery export while you are in there, because it is the only real long term answer.
Need this built properly?
This is the cheapest thirty seconds of analytics work available, and it only works if you do it early. I set up and repair tracking for ecommerce stores and lead generation businesses, so the numbers in your ad accounts match the money in your bank account.
Have a look at what I do, or book a call and we can go through your setup together.
