What does Google Analytics 4 really cost? Free tier, GA4 360 pricing and the limits of the free version

"But GA4 is free" - true, and yet companies spend four figures on implementation, five figures on consent and data infrastructure, and six figures on GA4 360. This post adds up where the costs really sit. And it answers the question almost no pricing comparison asks: what do you pay for the free version in data instead of dollars?

Is Google Analytics 4 free?

Yes. The standard version costs nothing - no collection fee, no expiry date, no ceiling above which Google holds out its hand. The "10 million hits limit" from Universal Analytics days does not exist in GA4; plenty of older guides still drag it along anyway.

What exists instead: limits. GA4 Standard caps retention, query depth, export and cardinality - and that is where the real price of the free version sits. More on that below.

What does GA4 360 cost?

Google publishes no list price. The pricing model is event-based and sold through Google Sales and certified resellers. The most reliable figures (as of 17 Aug 2026):

The often-quoted "$150,000 a year" comes from the old Universal Analytics 360 flat-rate model - for GA4 360 the entry is considerably lower, but the bill grows with event volume. Roll-up and subproperties count additionally at half event weight. Sales runs through certified resellers, who publish no numbers and price per event volume.

What the money concretely buys is above all four things: unsampled explorations (not available at all in standard properties), the fresh daily BigQuery export with updates throughout the day, expanded data sets against the (other) row, and an SLA - which only applies with a 360 contract. On top, almost every property limit grows with it, such as the Data API quota from 200,000 to 2 million tokens per day and data import storage from 10 GB to 1 TB.

When is it worth it? As a rule of thumb: persistently more than 25 million events per month, teams that need unsampled explorations without SQL, or compliance requirements with an SLA. For most other cases the BigQuery export is the more economical route - it lifts the same limits for a fraction of the cost. So whether you need 360 is decided by the limits below, not by the marketing.

The hidden costs of the free version

Even without 360, a clean GA4 setup is not free:

The setup range matches the public figures of analytics agencies in the German-speaking market (e.g. traffic3, Vienna, retrieved Aug 2026: from about €1,000 for small websites, €2,000-5,000 for mid-sized projects). It is open-ended upwards - an enterprise setup with server-side tagging plays in a different league.

The hard limits of the free version

This is where the real price sits. All values come from the official comparison table and the linked documentation pages (as of 17 Aug 2026):

Four of these limits hurt most in day-to-day work:

1. Data retention: 14 months pass quickly

Standard properties keep event and user data for 2 or 14 months - and the factory default is 2 months. Aggregated standard reports are untouched by this, but every exploration and every funnel ends at that edge. A real year-over-year comparison at event level is no longer possible in year two. Demographic data (age, gender, interests) is always deleted after 2 months, regardless of the setting. Check the setting right after setup under Admin → Data retention and switch to 14 months - it costs nothing, yet it rarely happens on its own.

2. Data thresholds: rows that simply go missing

As soon as Google Signals or demographic data is involved, GA4 hides rows with few users - system-defined and not adjustable. Small segments, B2B traffic, long-tail landing pages: exactly there, data goes missing without any hint in the report. Google's own ways out: expand the date range - or the BigQuery export, though Google Signals data never arrives there.

3. Sampling: above 10 million events, GA4 estimates

Exploration queries in standard properties switch to samples above 10 million events per query. Standard reports do not sample - but as soon as you segment freely, larger properties leave the ground of measured numbers.

4. Cardinality: the (other) row eats your detail

Dimensions with more than 500 unique values per day count as high-cardinality; beyond the cardinality limit of 50,000 values, GA4 condenses the rest into the notorious (other) row. If you track many URLs, search terms or product IDs, you lose exactly the rows that would be interesting.

All four limits have the same way out - and it is cheaper than 360: the BigQuery export.

What does the BigQuery export cost?

The export itself is free for standard properties, with a daily limit of 1 million events per day. Two details almost nobody mentions: if your property exceeds the limit persistently, Google pauses the daily export and does not backfill missed days - you only get an email warning. And the streaming export has no volume limit but costs $0.05 per GB (roughly 600,000 events per GB).

Beyond that you only pay for BigQuery itself (pricing, EU, as of 17 Aug 2026):

• Free tier: 10 GiB of storage and 1 TiB of query processing per month - permanent, not a trial.
• Storage: $0.02 per GiB/month (active), $0.01 long-term; Frankfurt (europe-west3): $0.023 / $0.016.
• Queries: $6.25 per TiB scanned (EU multi-region), Frankfurt $8.125 - the first TiB per month is free.

Worked example: a property with 500,000 events per month produces around 0.8 GB of export data monthly. Storage and typical audit queries stay entirely within the free tier for a good year, and cost cents after that - while every limit from the previous section (retention, sampling, thresholds, cardinality) is lifted in the export, because you work on raw data.

A larger property stays manageable too: 20 million events per month are roughly 33 GB of export data and still fit under the daily limit. After a year, about 400 GB sit in storage - 4 to 8 dollars a month, depending on the age of the partitions. BigQuery only gets expensive through frequent full scans over long time ranges, not through collecting.

Estimate your own costs in five minutes

1. Read off events per month - in GA4 under Reports → Engagement → Events, sum the event count of the last 30 days.
2. Convert to data volume - roughly 600,000 events equal 1 GB of export data.
3. Hold it against the free tier - the first 10 GiB of stored data per month are free; every further GiB costs $0.01 to $0.02 monthly. Even ten times the free tier stays under $2 a month.
4. The 360 question only arises above 25 million events per month - below that, the export answers it in practically every case. How to set the export up and which audits it enables is covered in the BigQuery export guide.

The invisible cost of bad data

The most expensive line item appears on no invoice: decisions made on broken data. Three classics with a direct money angle:

• Duplicate transactions inflate measured revenue - and every ROAS calculation on top of it is wrong. How the auditor finds them via second-buckets is covered in duplicate purchases in GA4.
• Unassigned and Direct shares shift budget into the wrong channels because attribution runs into a void - see unassigned traffic.
• AI traffic inside the Direct channel: of all cohorts, the currently best-converting visitors hide as Direct - why, is covered in the post on the new "AI Assistants" channel.

If you invest €2,000 in a setup but steer five figures of monthly ad spend on a broken channel distribution, you saved at the wrong end.

Is one of these limits already hitting you?

Three checks you can do yourself in five minutes - or have checked automatically:

1. Data retention: is your property still on the 2-month default? The GA4 Auditor reads the data_retention_settings via the Admin API and flags the setting - the quiet classic among audit findings.
2. BigQuery link: is the export active at all? Without it there is no raw-data fallback line when sampling and thresholds strike.
3. Channel distribution: how large are your Direct and Unassigned shares? From 20 % Direct it turns amber, from 35 % red - the thresholds are documented in the help page on attribution & acquisition quality.

The GA4 Auditor checks all three automatically - as part of its 24 checks across Admin API, Data API, BigQuery export and on-page analysis. Check your property now.

Related topics

• BigQuery export for GA4 - the auditor's standpoint
• Duplicate purchases in GA4 - why they happen
• AI traffic in GA4: the new "AI Assistants" channel
• GA4 audit checklist
• Help page Traffic-light checks from BigQuery

#pricing#ga4-360#bigquery#ga4