How to export GA4 data: back up your history before you lose access

To export GA4 data, you have three options: download reports as CSV, PDF or Google Sheets from the Share this report menu, pull report data with the Google Analytics Data API, or link the property to BigQuery to get raw, unsampled events. Report downloads are the quickest and need no code. BigQuery keeps the most detail, but it only records data from the day you link it. Before any of these, raise your data retention to 14 months, because that one setting protects more history than any export.

GA4 data export is any copy of Google Analytics 4 data stored outside the GA4 interface, either as aggregated report tables (CSV, Sheets, API) or as raw event rows (BigQuery), so that you can still read it if you lose access to the property.

Why should you export GA4 data now?

Because GA4 history disappears on a schedule, and some of it cannot be rebuilt. Event-level data that powers explorations and funnel reports is kept for only 2 or 14 months on standard properties. Age, gender and interest data is always kept for two months, whatever you set. If your team reaches GA4 through a VPN or the property sits in one person’s Google account, losing access is a second, less predictable deadline. The guide to Google Analytics in Iran and its risks explains that second deadline in detail.

An export also helps when nothing goes wrong. A monthly CSV of channel performance lets you compare this year with last year after the exploration window has closed, and it gives you a baseline when you run a second tool next to GA4.

Which GA4 data can you export, and how?

Each method keeps a different layer of your data. Pick by what you will need to answer later, not by what is easiest today.

Most small and mid-sized sites need only the first row. The other two are for teams that already work in SQL or scripts, or that need user-level detail.

How do you download a GA4 report as CSV or Google Sheets?

Open the report, click Share this report in the top right, choose Download File, then pick Download PDF, Download CSV or Export to Google Sheets. You need the Viewer role at the property level. A CSV or Sheets export holds up to 100,000 rows, and you cannot export while a report is open in Customize report mode.

A routine that works for most teams:

1. Raise retention first. In Admin, go to Data settings → Data retention and set event data to 14 months. Google applies a longer setting to data you have already collected, so this extends what you can still explore. It needs the Editor role.
2. List the reports your team actually opens. Usually traffic acquisition, landing pages, conversions or key events, and e-commerce purchases.
3. Set the widest date range you have and add the secondary dimensions you use, such as session source / medium. Exports keep only what the table shows.
4. Download CSV, not PDF, for anything you may want to analyse later. PDF is for reading, CSV is for comparing.
5. Name files consistently, for example ga4_traffic-acquisition_2025-11_2026-10.csv, and store them in a shared folder your company owns.
6. Repeat on the first working day of each month while GA4 is still running.

A worked example (hypothetical). Say a store has 400 landing pages and you export the landing-page report by session source / medium, with about 35 source / medium pairs. The worst case is 400 × 35 = 14,000 rows, far under the 100,000-row cap. Add a date dimension by day for a year and the same table could reach 14,000 × 365 ≈ 5.1 million rows, which no single download can hold. The fix is to export one month at a time with a month dimension, or to keep daily detail only for the top pages.

When should you use the GA4 Data API instead?

Use the Data API when you need the same report every week, across many properties, or with more combinations of dimensions than you want to click through. A script can request dimensions and metrics for a date range and write them to a spreadsheet or database, which removes the manual monthly routine.

Two things to plan for. First, the API returns report-style aggregated data, not raw events, so it does not replace BigQuery for user-level analysis. Second, Google meters usage in quota tokens per property, per hour and per day, and bigger queries cost more tokens. The current numbers are on Google’s quota page; check them before you schedule a large backfill, and split long date ranges into monthly requests.

How do you set up the GA4 BigQuery export?

Create or choose a Google Cloud project, enable the BigQuery API, then in GA4 go to Admin → Product links → BigQuery links → Link. Choose the project and a data location, pick which streams and events to include, and select Daily export, Streaming export or both. You need Editor access on the GA4 property and the Owner role on the BigQuery project.

What the export does, per Google’s documentation:

• Daily export writes all raw, unsampled events from the previous day into one table per day (events_YYYYMMDD). Standard properties can export up to 1 million events a day.
• Streaming export sends events within minutes into an intraday table. Google calls it best effort, with no completeness guarantee, and notes that it lacks traffic-source attribution for new users. Use daily export as the record of truth.
• Data starts flowing within 24 hours of linking. Google’s setup page does not describe any backfill, so treat the export as starting on the day you link.
• Over the limit, it pauses. If a property keeps exceeding 1 million events a day, the daily export is paused and earlier days are not reprocessed. Event filters on the link help you stay under it.

On cost: linking is free and small properties can start in a BigQuery sandbox, but sandbox limits apply, the setup page names a 10 GB limit for free projects, and exports beyond the sandbox are billed. Streaming export costs $0.05 per GB, and Google says 1 GB is roughly 600,000 events.

A worked example (hypothetical). A site that logs 30,000 events a day generates about 900,000 events a month. Using Google’s rule of thumb of 600,000 events per GB, that is about 1.5 GB a month, so streaming it would cost about 1.5 × $0.05 ≈ $0.08 a month. At the same ratio, a 10 GB sandbox would fill after about 10 × 600,000 ÷ 900,000 ≈ 6.7 months. The real constraint for most teams is not the bill. It is having someone who can write SQL against the event tables.

What GA4 data can you not get back?

Some losses are permanent, and knowing them tells you what to export first.

The practical rule: aggregated reports survive longest, so CSV exports of them are your safety net. Anything event-level has a clock on it from the day it was collected.

How do you keep measuring after you export?

An export freezes the past. It does not record next month. If the reason you are exporting is a risk to your GA4 access, the second half of the plan is to start collecting data in a tool you control, before you need it.

AdPix is a web analytics platform that replaces Google Analytics 4: realtime, acquisition, engagement, conversion and e-commerce reports, plus attribution, lift testing and marketing mix modelling, with no data sampling and a free plan. It runs alongside Google Analytics without interfering, so you can install it next to GA4 today and compare both for a few weeks. Its data is hosted in Iran and can be exported at any time, which matters if the export problem is what brought you here. The free plan includes 100,000 events a month, 30 days of raw data and 13 months of aggregate data; see AdPix pricing for larger plans. This article does not assume any tool can import your GA4 history, so keep your CSV archive as the record of the years before the switch.

For the wider choice of tool, the Google Analytics alternative guide compares GA4, self-hosted Matomo and AdPix on sampling, retention, hosting and cost. To make sure the new tool records the same traffic sources and conversions as GA4, use the five layers in the website tracking guide and keep your UTM parameters identical in both tools.

Start with the two-minute step: open Admin → Data settings → Data retention, set it to 14 months, then download your traffic acquisition report as CSV for the full range you have.

Measure your site with AdPix today

The free plan covers 100,000 events a month, no credit card. Setup takes under five minutes.

Start freeSee pricing