BigQuery & Basics

What is the difference between UA & GA4?

Reminiscing about the good days of UA? Well, we’ve moved on.

Let’s jump straight into it.

This post isn’t going to list UI differences, those are a given.

We’re talking about the underlying data differences.

Essentially the data models are different

  • The core of Universal Analytics is sessions - everything else is centred around sessions.
  • Google Analytics 4 stores events - these are the building blocks for everything else.

Sessions are the core to the Universal Analytics data model

us-data-model-centred-on-sessions

The screenshot shows the data structure for UA which is centred on a session.

Each row is 1 session.

ua-interface-screenshot

Within each session, we have supporting information around the session such as:

  • The Time/ Date
  • The Visit ID
  • The Device
  • The “Hits”
  • Location
  • Source

But it’s all about the session - that was the core of UA.

Events are the core of GA4’s data model.

An Event Based method to collecting data means that every single thing counts as an event.

ga4-data-model-centred-on-events

Then when we look at GA4 data, each row is based on an event.

The other columns provide additional information about that event.

starting-point-event-name

The types of things that can be an event includes:

  • Starting a session
  • Making a purchase
  • Doing a site search
  • Scrolling down a page

Those events contain all the information to roll up to everything else into sessions and users.

So what has changed between UA & GA4?

The primary change is the measurement model and the move to an Event & Parameter model over the previous session/ pageview model.

This has meant that the way we’ve changed the way we look at a number of:

  • Metrics
  • Dimensions
  • Conversions

Let’s get into some of those changes.

Sessions have changed slightly between UA & GA4

GA4 Sessions

A GA4 session starts when a session_start event is fired.

This creates a unique session_id which is maintained with other events in the session (i.e other pageviews or key events).

GA4 session ends when there has been a 30 minute period of inactivity (though you can adjust the settings this is a longer timeout period).

UA Sessions

A session in Universal Analytics lasts as long as the user was active, but could come to an end if:

  1. There has been a 30 minute period of inactivity.
  2. The session went over midnight, meaning that a new session would start after midnight.
  3. If a user picks up new campaign parameters whilst on the site.

GA4 introduced Engaged Sessions

One challenge that GA4 looked to address was how to quantify a valuable session.

Universal Analytics has attempted to do this with bounces.

Sessions are only counted as Engaged sessions if either:

  • The user has been on the site for 10 seconds.
  • There were multiple pageviews
  • There has been a conversion.

This is, in many ways, a far more effective way of measuring a “valuable” session.

The introduction of Active Users in GA4

A new addition with GA4 is the concept of an “Active User.”

A user is considered an active user if either:

  1. They’ve had an engaged session.
  2. Analytics detects a first_visit/ first_open
  3. Analytics detects a engagement_time_msec / user_engagement

The advantage of this is that you can see which users have not only come to the site, but those who have come to the site and done some form of action.

Other Engagement Metrics which stem from Engaged Sessions

  • Engaged sessions per active user - The average number of engaged sessions per pseudo_user_id or user_id.
  • Average Engagement time - Engagement time is the total length of time the website was in focus or your app was in the foreground.
  • Engagement Rate - The percentage of engaged sessions on your website (compared to regular sessions).

Ecommerce Metrics

The setup of e-commerce events is actually significantly easier on GA4 compared to UA which used to involve a very specific configuration for Enhanced e-commerce.

In GA4 e-commerce events are collected in the same way as every other event.

A number of e-commerce events are also automatically tagged as Key Events such as:

  • View Item List
  • Select Item
  • Add to cart
  • Add to Wishlist
  • Purchase

The actual numbers being collected across GA4 & UA, though, should stay fairly similar.

GA4 Key Events replaces conversions

What’s the difference between Events vs Key Events?

As we’ve said before, everything is an event in GA4.

But clearly some events are more important than others; A purchase is more valuable than a pageview, for example. The priority conversion events on GA4 are called Key Events. This is what we previously called Conversions in GA4.

Examples of Key Events are:

  • form_submit
  • purchase
  • generate_lead

So how do you convert an event to a Key Event?

Go to the GA4 interface then go to Admin > Events.

This will list the full list of events that you have on site, and you can toggle these to be Key Events.

ga4-existing-events

What does the GA4 Export Schema look like?

If you want to read a whole blog on this then head over this way.

The event is core to BigQuery GA4 export and comes with supporting information.

Each event has:

  • event_name - Every event needs a name.
  • event_timestamp - Every event arrives at a timestamp.

There are also other pieces of information which help us tie everything together such as:

  • session_id - These events were all part of the same session.
  • user_id - These events all belonged to the same user.

Then there are pieces of information about the user:

  • is_active_user - Is the user currently considered active?
  • user_id - Is the user a logged in user?

Then there is information related to the event such as:

  • The device the event took place on
  • The location of the event
  • The traffic sources for the event.
  • e-commerce information.

In BigQuery, the schema looks as follows:

these-are-nested-fields

The real clever part of the GA4 data schema is event_params (Event Parameters).

This is what allows it to be so flexible.

Each event can have 25 totally customisable properties.

These parameters let you send whatever you want to GA4 as long as you give it a consistent name.

See the example below:

example-event-params

To wrap it all up.

Events are the core of GA4. Everything is built around them, providing a more flexible and scalable way to track user interactions.

For UA, Sessions were the core. Everything was built around them.

GA4’s event-based model offers far more flexibility than was possible with UA - especially when accessing the raw data via BigQuery.

Onto the mini sales pitch.

Pipeline - GA4 takes that raw data and transforms it into simple to use tables, doing all the complex transformations for you and letting you configure it all through a simple UI.

If this is something you are interested in then head over to the GA4 Pipeline product page.