Comparing BigQuery and GA4 user interfaces

👋🏻 This document presents the differences between GA4 and BigQuery in terms of data display in the user interface (UI) and the limitations of each solution.

Data scope

GA4

Aggregated event and user data, i.e. data grouped by date, time, dimension, etc.

BigQuery

The data is at event level, so it is not aggregated.

Data access method

GA4

Data accessible via GA4 interface.

BigQuery

GCP console or any reporting application that can query BigQuery data

High cardinality

⚠️ Dimensions with high cardinality are dimensions with over 500 unique values in one day. If this limit is exceeded, the data in the (other)More information :

GA4

One limitation of GA4 is its cardinality limit for dimensions. GA4 can only generate reports on a limited number of unique values for a dimension, before grouping the less frequent values in a row. (other). The cardinality limit for GA4 dimensions is 500 unique values per day.

BigQuery

BigQuery is better suited to handling high-cardinality data than GA4. There is no cardinality limit for dimensions. So you can generate reports on any number of unique values for a dimension, without having to worry about encountering (others).

Sampling

ℹ️ Imagine you want to estimate how many people live in a city of 100,000. You could count how many people there are on a street 100 meters long and multiply by 1000, or count how many people there are on a street 50 meters long and multiply by 2000 to get an accurate estimate of the total number of people in the city.

GA4

Google Analytics uses data sampling when the number of events exceeds your property type limit (500 per user). This allows you to explore your data in detail using a representative sample, which is then extrapolated to provide accurate information.

• No sampling for standard reports except for tunnel reports
• Sampling possible in exploration reports when quota exceeded

BigQuery

BigQuery gives you access to 100% of your data.

Tresholding

ℹ️ Data thresholds are applied to prevent anyone viewing a report or exploration from determining the identity of individual users based on :

👉🏻 Demographics
👉🏻 Interests
👉🏻 Other signals in the data.

Google signals are data from users logged into their Google Account who have activated ad personalization.

GA4

BigQuery

Thresholding doesn't really apply to BigQuery data. There is no demographic data (age, gender, etc.) from the Google signals that are sent to BigQuery.

Data-based allocation

GA4

In GA4, there are three types of allocation models:

1. Rules-based models for paid and organic sources: these models assign a fixed weight to each channel according to its position in the conversion path.
2. The rules-based model for Google's paid channels: this model assigns a fixed weight to each Google Ads channel according to its position in the conversion path.
3. Data-driven attribution: this model assigns a weighting to each channel according to its actual impact on conversion.

BigQuery

GA4 uses its own session-level attribution model. As a result, this information is not directly accessible in the data exported to BigQuery, and cannot be calculated with the utmost precision.

Conversion and behavioral modeling

☝🏻 Google uses modeling to estimate online conversions that cannot be observed directly. This makes it possible to provide more accurate reports, optimize advertising campaigns and improve automated bidding.

GA4

GA4 uses modeled data in its key reports to attribute conversion events across channels.

⚠️ Behavioral modeling in the GA4 explorations section only concerns path, funnel and free-form tables.

BigQuery

Modeling is not included. BigQuery data contains cookie-free signals collected by Google Analytics when consent mode is enabled, and each session has a user_pseudo_id different.

Modeling can lead to differences between standard reports and granular data in BigQuery :

👉🏻 Fewer active users may be seen in GA4 reports than in BigQuery, as the modeling attempts to predict multiple sessions from the same user who has refused cookies.

Limitations

GA4

• Reports and analyses: up to 150 customized reports per property.
• Explorations: up to 200 individual explorations per user and per property, and up to 500 shared explorations per property can be created. Up to 10 segments can be imported per exploration.

BigQuery

Standard properties have a daily export limit of 1 million events per day. Analytics 360 properties have almost unlimited export .