The Future is Here: A Detailed Overview of Google Analytics 4

The year 2020, among other things, has brought to the world the long-awaited Google Analytics 4, which is now officially released and available worldwide. What are the key changes and how Google Analytics 4 differs from Universal Analytics? Who will benefit from the transition to the new data model and how to get started with Google Analytics? The new property type is explained by Kate Guzevataya, Head of Web Analytics at Promodo.  

What is Google Analytics 4

The new Google Analytics property was originally presented as App + Web, which presupposed combining data from web and mobile apps to ensure a more unified view of customer data. Let’s see why the new property type is so promising and what new features it brings to the users.

1. Scalable cross-platform analytics built around events
2. Machine learning (ML) and natural language processing (NLP) functions are available to all Google Analytics users
3. Maintaining privacy and avoiding the need to set cookies is a priority
4. Seamless integration with all Google products
5. Cross-platform user identification so you can see the entire path of a user across devices and platforms

We’ll elaborate on each point separately.

Scalable cross-platform analytics

In the standard web version of Google Analytics, everything was built around user sessions, that are an artificially brought indicator. The new Google Analytic is all built around events. This allows you to collect the same data across all devices and platforms, thus improving the reporting quality and providing you with a unified view of the user path.

Firebase Analytics is an analytics platform for accounting of mobile and web applications. Google Analytics 4 is essentially an analogue of Firebase with its interface and data accounting logic, as well as some additional functionality that Firebase may lack.

Machine learning

Some Machine Learning and NLP features were represented in the previous version of Google Analytics but were used very little, which is understandable as more of ML potential was to be added (and actually was included) in Google Analytics:

• predict the probability of conversions and create audience forecasts for Google Ads
• warn about the trends discovered in your data (i. e. products that are in demand due to changing user needs)
• find anomalies in reports, which is about the notifications that were present in Universal Analytics. The information displayed in Google Analytics 4 tells you about the errors occurred, any issues with tracking conversions, incorrect calculation of the Conversion Rate, or problems with UTM parameters tracking;
• predict the likelihood of customer outflows so you can effectively invest in retention. The Google team plans to further develop this area and add new forecast indicators. Such as, ARPU, which allows users of a new resource to adjust their marketing strategy and increase the ROI using ML insights.

Maintaining privacy

As part of the GDPR, preserving the privacy of user data remains the main priority. This means that if the user does not consent to sharing personal data, nobody has the right to collect and analyse their statistics. Previously user data was obtained through cookies. Now Google is considering completely rejecting the use of cookies and switching to tracking by browser & mobile device IDs.

So, what does privacy mean in the context of Google Analytics 4:

• Google Analytics 4 uses the gtag.js library that works without cookie files.

The latest version of Universal Analytics already uses this library, but few people switched to it, as this library needs to be installed in the site code. Here at Promodo we see no need to rush with the gtag.js, as the standard Universal Analytics is installed via Google Tag Manager and doesn’t require any additional actions.

We would assume that in the nearest future Google will abandon Client ID to fully rely on internal device & browser identifiers, as well as a cross-platform identifier that is generated in CRM, i.e. User ID.

• IP anonymisation in Google Analytics 4 is configured by default and cannot be changed.

Actually, IP anonymisation was not tracked in the previous versions of Google Analytics either. The only difference is that in the new GA, there is no way to customise the analytics for tracking.

Seamless integration with all Google products

Integration with YouTube is considered the most advanced so far. Google is actively working to improve the quality of evaluations for YouTube campaigns, for example, making it possible to track view-through conversions. This will help to understand the following:

• how my YouTube ad campaign affects specific audience involvement indicators;
• how my YouTube campaign affects the bounce rate, сonversions and other events on my website.

In the new version of Google Analytics, measurement of YouTube campaigns represents a more advanced system that allows to analyse statistics on YouTube within Google Analytics interface.

A deeper Google Ads integration allows creating audiences and running campaigns that attract new customers with more relevant and useful offers. And the most valuable is that the device the customers are using doesn’t matter.

Google Analytics 4 has direct free data export to Google BigQuery (a cloud-based database for storing, processing data and building various reports). Previously, the export feature was paid and provided only in GA 360, while in GA 4 it’s now free. Google BigQuery is good because it provides constant access to raw, non-aggregated data, which means you’ll be able to build various reports at any stage of project maturity.

Cross-platform user identification

One of the key points of the GA update is the ability to see the entire user journey that can start on the website and continue in the mobile application. To count real users (not the devices and browsers they used) who interact with your company, the new resource type uses 3 levels of identification:

User_id
Google Signals
Device_id

User ID is the main authentication layer and one of the key figures in this update. It is the User ID that allows to combine information from devices and websites inside Google Analytics. The user ID, however, will have to be customised as it is not set up by default.

Implementing event-driven analytics means you can more accurately track the user’s journey from first touch to conversion and repeat orders. Moreover, if a user performs the same event on various devices, this data will merge into a single touchpoint. For example, if a customer adds an item to the shopping cart twice (on a smartphone and then on a laptop), the “Add to shopping cart” event will be counted once.

Previously, the results for the application and for the website were presented as separate users, which could be difficult to combine. Google Analytics 4 makes it easier and faster.

‍

Google Analytics 4 vs Universal Analytics

Google Analytics 4 has three key differences over the previous version:

Analytics is no longer built around sessions, but around events. Google suggests abandoning the artificially created concept of sessions. If you need session data, you can build it manually using raw data from Google BigQuery.

There are advanced data collection settings for the entire website and settings that change with each event.

GA 4 has built-in end-to-end user id reporting.

Key metrics

Universal Analytics contained numerous metrics, such as Page/Screen View, Events, Sessions, social, various Hit types, User ID, Client ID and the others. In the new resource type Google reduces all the metrics to the four key ones:

• Event
• Event Parameters
• User Properties
• User ID

Three types of events and settings in Google Analytics 4

Google Analytics 4, in line with Firebase, provides three types of events and their parameters:

1. Collected automatically. For example: page_view, session_start, view_search_results, scroll, file_download. A complete list of events is available here.
2. Recommended events, which are grouped into business areas: retail and e-commerce, travel, games (See the full list here.)
3. Custom events, that are basically all other events you would like to implement and monitor (These are limited by Google Analytics 4.)

Recommended and custom events are implemented independently..

Each event can have additional definitions

Custom Definitions are end-to-end dimensions and metrics that help you stay within Google Analytics 4 limits.

Previously, the “Add to cart” event had a separate script and an individual parameter for each value, while now they are united by a single event structure (product name, product cost, size, color, etc. – all of them are passed to parameters, as well as user properties).

No categories, actions and event shortcuts

In Google Analytics 4 there does not exist such concepts as category, action and event shortcut.

For existing customisations and collected data, these properties are mapped to custom event settings. You need to register the properties in GA 4 reports in order to see them. In order to see some additional properties transmitted with each event, you can select them in the interface.

Page views become page_view events

Along with that, the page_view event in GA4 has preset parameters:

page_location
page_path
page_title
page_referrer

The page_view event is fired automatically if the “config” snippet of gtag.js is implemented.

Sessions and session counting in Google Analytics 4

Sessions are present in Google Analytics 4 reports, but the way they are counted differs from the one of Universal Analytics:

• The session is initiated by the automatically collected session_start event. This is a preset event that starts collecting data after setting the counter
• Session duration is the interval between the first and last events
• Interactions are recognised automatically (no interaction event dispatch required). Common user actions like the interactions with the screen, elements on the site are tracked; the number of interactions is displayed
• The late request processing timeout is 72 hours (versus 4 hours for UA Properties). If you compare the number of sessions Google Analytics with Universal Analytics reports, you will find that sessions are lower in GA 4. This is because the hits that were sent after the end of the session can be assigned to the relevant session within 72 hours. Accordingly, generating session reports will take longer

Currently, session duration in GA 4 can’t be configured.

Custom Dimensions & Metrics

In order to include custom dimensions and metrics in Google Analytics 4 reports, the latter must be transferred to a new resource type according to the rules of Google. In Google Analytics 4, there are analogues for hit-level and user-level parameters, but no equivalents exist for session-level parameters. Although, you can define them at the hit level.

To use custom product-level definitions, you must add them separately. This feature is still in development, and there are no eCommerce reports that would contain custom product-level definitions, thus it’s not yet clear how this will work.

Whereas previously in Google Analytics there were 4 levels of indicators focused on hits, users, sessions and products, now their equivalents are events and event parameters.

User Properties

In Google Analytics 4 there appears a new feature called “User Properties”. These are definitions that correspond to a specific audience/user. For example, gender, city, new or returning customer etc.

Properties that concern specific users are applicable to their behaviour. Based on User Properties, audiences for personalising ads are formed.

New approach to data presentation

Let’s see how the Traffic Sources report looks like in Google Analytics 4:

The “Events” report is visually represented in a different way:

Scatter chart shows which event is most popular when added to the cart (no such visualisations were available in the previous version)

A new way of building Custom Reports

In Google Analytics 4, there is no longer the “My Reports” menu. Instead, a promising “Analysis” appears, with its two tabs: “Analysis Center” and “Template Gallery”. In addition, the report types have been replaced by “Methods”.

Funnel analysis

If you have a set of events using which you can form a funnel, you can do this manually.

Note: first you need to set up the correct tracking, and only then generate reports, as shown in the pictures.

In the same report, you can create segments for any step in the funnel.

Create conversions in 1 click

Another difference between Google Analytics and the previous version is that you no longer need to create goals manually, you just need to use the slider for the event that is being tracked to become a conversion.

New editor for creating segments and audiences

In Google Analytics 4, you can add a new segment from the funnel steps.

To do this, select “Build an audience” and create the necessary condition (or do not change anything).

Who should consider transition to Google Analytics now

It’s worth to start implementing Google Analytics 4 right now provided:

• you use Data Layer and Google Tag Manager to collect data on your website;
• you are an active user of YouTube Ads and User ID-based remarketing;
• you are an active user of Firebase and your team is familiar with Firebase data collection logic as well as with the App + Web (Firebase) data schema for export tables in BigQuery;
• you use a few tags (then you will only need to make minimal adjustments).

Our opinion: the new resource type is definitely worth implementing for everyone. Moreover, the sooner you switch to GA 4, the faster you will begin to collect historical data, which means there will be more information that can be used for making decisions. Plus, the faster you get value from ML insights. New analytics will not interfere with the old one, as long as it can be used separately. But the fact that it will develop rapidly and eventually everyone will switch to it (as it was once with Universal Analytics) is a fact. One of the difficulties is that the data structure and logic of their collection are significantly different for GA 4 and Universal Analytics. Therefore, it will be rather problematic to combine the data of the two resource types.

Why wait and not rush with the transition to Google Analytics 4

You may face some problems implementing Google Analytics 4 in case:

• code is the main tracing method on your website;
• you use GTM as the main tracking method and have many tags in the container, especially if tags are bound to auto events (in this case, you can start a new counter and set up traffic collection, but do not reconfigure the tags yet);
• your website is large and has many subdomains, each of which you track separately;
• you have no system of metrics (unified names and values) ​for events and their parameters as well as a unified approach to the hierarchy of events (in this case, it’s just unclear what events should be added to GA interface first and which ones are better to postpone);
• you have a website and an app with no common hierarchy of events, and you plan to implement end-to-end analytics on them;
• your team is not familiar either with raw data from BigQuery or with the principles of Firebase Analytics/App+Web and the upload scheme.

And more cases to show that Google Analytics 4 is not a fit for you (at least yet):

• if you expect several teams to be using the new resource type simultaneously – there is currently no access control implemented in GA 4;
• if you want to analyse expenses and ROAS for non-Google campaigns – there is no data import in the new resource yet;
• if you need to export conversions to Search Ads 360 and Display & Video 360 – the integration with other Google products is not yet working at full capacity.

How to upgrade to Google Analytics 4

Since we recommend using Universal Analytics in parallel with Google Analytics 4, we will explain how to switch to the new analytics without abandoning the old one.

1. Create and configure your GA 4 property by following the instructions on Google Support page.
2. Add a tracking code manually or via GTM. We recommend using the Tag Manager to make it more convenient and faster.
3. Define what events and parameters you want to collect into a new resource type.
4. Use two resource types in conjunction to compare how data is collected.

Important notes:

• you cannot combine the old data with the new counter, since the resources have different logic, i.e. the whole concept is built around events but not around sessions as it was before;
• only one Firebase project can be added to one GA 4 resource;
• At the same time, you can configure multiple data streams from different apps into a single Google Analytics 4 resource.

Wrapping up

The main update of the whole Google Analytics concept: in GA 4 everything is built around events, parameters of events and users, and not around sessions, as it was before.

The key feature of the App + Web functionality is cross-platform analytics between the website and the applications.

To configure a new GA4 resource, you can use the previously configured GA via gtag.js or GTM.

When setting up GA 4, a new WP resource is automatically created. And this is the moment that launches data collection. It’s impossible to migrate data from old WPs.

Google does not urge you to abandon the old Google Analytics and switch to the new one. They (and we at Promodo) recommend launching the new Google Analytics 4 in parallel and start collecting data into it. The source of historical data is still the standard GA.

There are flaws in the new GA 4. Moreover, not all features are available to users yet – the developers are gradually rolling these out.

There is no way to import costs from non-Google sources into Google Analytics 4 yet.

Data upload from GA 4 to Google Big Query can be configured for free. The export scheme is the same as in Firebase.

You can start customising GA 4 assets and collecting data. The earlier you set it up, the more historical data you will be able to collect.

‍