An In-Depth Look at Google Analytics 4: New Capabilities, Benefits, and Disadvantages

About a year ago, the Google team presented App + Web functionality, which allows you to combine data from websites and mobile applications in one Google Analytics resource. Since then, Google has tested this new type of resource, made changes to it, finalized it, and brought it out of beta under a different name. Meet Google Analytics 4.

In this article, we talk about what Google Analytics 4 is, how it differs from Universal Analytics, what value it gives to businesses, and what problems you can solve with it. We also analyze which companies should start using Google Analytics 4 right away.

Google Analytics 4 is a new resource type in Google Analytics. It looks a little different than a Universal Analytics (UA) resource, and it’s easier and faster to configure. In presentations, the Google team has repeatedly called this new type of resource the future of analytics, citing the following:

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

Let’s take a closer look at these benefits.

An event-based approach allows you to collect reliable and consistent data across multiple devices and platforms.

In the standard web version of Google Analytics, everything is built around user sessions, and in Firebase, everything is built around events. Therefore, it’s difficult to analyze a user’s transition between platforms, since there’s no universal measure of user behavior. Even with raw data, you need to make great effort to build a high-quality user flow.

Google Analytics 4 combines all analytics around events. This allows you to collect the same standardized data for all devices and platforms, improving the quality of your data and providing you with a single report across the user path.

One of the main advantages of Google Analytics 4 is its machine learning and natural language processing (NLP) functions, which you can use to:

• predict the probability of a conversion and create audience forecasts for Google Ads based on that probability
• warn you about important trends in your data (i. e. products that are in demand due to changing user needs)
• find anomalies in reports
• predict the likelihood of customer outflows so you can effectively invest in retaining customers

The Google team plans to continue developing in this direction and adding new forecasts such as ARPU so all Google Analytics 4 users can adjust their marketing strategies and increase their ROI using machine learning insights.

1. Google Analytics 4 is privacy-focused and uses the gtag. js library, which works without cookies. Accordingly, we can expect that in the near future, Google will abandon Client ID and rely only on internal device and browser identifiers as well as a cross-platform User ID identifier generated in the CRM.
2. IP anonymization in Google Analytics 4 is configured by default and cannot be changed.

So far, the most advanced integration is with YouTube. Google is actively working to improve the quality of evaluation for YouTube campaigns (for example, to allow you to track view-through conversions). This will allow you to find answers to these sorts of questions:

• How does my YouTube ad campaign affect specific audience involvement indicators?
• How does my YouTube campaign affect the bounce rate, events on my website (not necessarily conversions), etc?

With deeper Google Ads integration, you can create audiences and run campaigns that attract new customers with relevant and useful offerings no matter what device they use.

In addition, in Universal Analytics, the BigQuery Export feature is available only to users of the paid version, whereas in Google Analytics 4, this feature is free for everyone. You can activate data collection in BigQuery cloud storage in the Google Analytics 4 resource settings.

Google Analytics 4 considers individual users who interact with your company, not the devices and browsers they use.

It does this using three levels of identification:

• User_id
• Google Signals
• Device_id

By implementing event-based analytics, Google Analytics 4 enables you to better track a user’s path from first touch to conversion and reorder. Moreover, if a user completes the same event more than once using different devices, the data for this event will be merged into a single touchpoint. For example, if a customer puts an item in the shopping cart on a smartphone and then on a laptop, the “Add to shopping cart” event will only be counted once.

Let’s compare the key tracking concepts in Universal Analytics and Google Analytics 4:

Google Analytics 4

• Analytics is built around events, not sessions. Since sessions are an artificial concept, Google proposes abandoning them. If you need session data, you can build it yourself by working with raw data in Google BigQuery.
• There are advanced data collection settings for the entire website and settings that change with each event.
• With built-in end-to-end user_id reports, you don’t need to create a separate view to use user_id.

In Google Analytics 4, there are three types of events and their parameters, as there are in Firebase.

• Collected automatically — example: page_view, session_start, view_search_results, scroll, file_download (See the documentation for a complete list of events.)
• Recommended events are grouped into business areas: retail and e-commerce, travel, games (See the full list here.)
• Custom — all other events you would like to implement and monitor (Limited by Google Analytics 4.)

Recommended and custom events are implemented independently.

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

Google Analytics 4 doesn’t have such concepts as category, action, and event shortcut.

For existing settings and collected data, these properties are mapped to event settings. If you want to see properties in Google Analytics 4 reports, you need to register them.

These events are automatically collected if you have a “config” gtag.js fragment implemented.

The page_view event has these preset parameters:

• page_location
• page_path
• page_title
• page_referrer

Google Analytics 4 reports have sessions, but they’re considered differently than in Universal Analytics:

1. A session is triggered by the automatically collected session_start event.
2. The session duration is the interval between the first and last events.
3. Interactions are automatically recognized (no interaction event needs to be dispatched).
4. The late case processing timeout is 72 hours (versus 4 hours in UA Properties). If you compare the number of sessions in a Google Analytics 4 and Universal Analytics report, you may encounter fewer sessions in the former because hits sent after a session is complete can be assigned to the correct session within 72 hours. Accordingly, session reports are issued for a longer period of time.
5. The session duration cannot be configured in Google Analytics 4 at this time.

In order for custom dimensions and metrics to be included in Google Analytics 4 reports, they must be transferred to a new resource according to Google’s rules. Whereas hit-level and user-level parameters have analogs in Google Analytics 4, there are no equivalents for session-level parameters. Alternatively, you can define them at the hit level.

To use custom product-level definitions, you must add them separately. It isn’t yet clear how this will work, because the feature is still in development and there are no reports on Ecommerce that contain custom product-level definitions.

Google Analytics 4 has introduced a new User Properties feature.

User Properties are definitions that correspond to a specific audience/user: gender, city, new or returned customer, permanent customer, etc.

Properties that affect specific users extend to all their behavior. Based on User Properties, Google Analytics 4 forms audiences for personalizing ads.

You should already implement Google Analytics 4 if:

• you collect data on your website through Data Layer and Google Tag Manager
• you use few tags (meaning minimal adjustments)
• you’re actively using YouTube Ads and User ID-based remarketing
• you’re actively using 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

The faster you move to Google Analytics 4, the faster you can start to collect historical data, the more information you’ll have for decision-making, and the faster you’ll get value from machine learning insights. As we’ve already seen, Google Analytics 4 and Universal Analytics have significantly different data structures and data collection logic. Therefore, combining data from these two resources will be problematic.

Currently, there is only a manual way to import ad cost data into Google Analytics 4. You can automate this process and save valuable time, using the solution from OWOX BI.

Important! If you plan to import advertising costs to Google Analytics 4, then you need to add the required parameter utm_id (campaign identifier) to the links of your advertisement campaigns.

You may encounter problems implementing Google Analytics 4 if:

• code is the main tracing method on your website
• you use Google Tags Manager as your main tracking method and have many tags in the container (especially if tags are bound to auto events)
• you have a large website with many subdomains, each of which you track separately
• you don’t have a 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 isn’t clear what events are more important to add to the GA interface first and what it’s better to postpone.)
• you have a website and applications without a common event hierarchy
• your team hasn’t yet worked with raw data in BigQuery and isn’t familiar with the principles of Firebase Analytics / App + Web and with the upload scheme

If these statements apply to you, we recommend first building general data collection logic and only then implementing Google Analytics 4. Otherwise, you can quickly find yourself without free slots for user parameters.

If you don’t have a built-in data collection scheme, you can collect useless events in BigQuery and face export restrictions in addition to paying to store garbage data that won’t be useful to anyone (for example, data on scrolling events and banner views).

The main drawback of Google Analytics 4, in our opinion, is the scheme for exporting data to Google BigQuery, in which the key parameters of events and users are stored in nested fields. This means that in order to obtain the necessary information from Google Analytics 4 tables, you’ll need to process more data compared with OWOX BI data streaming or BigQuery export in standard Google Analytics 360.

Google and OWOX analysts so far recommend using both versions of Google Analytics resources. For this, you’ll need to:

1. create and configure a new Google Analytics 4 resource
2. add a tracking code manually or via GTM (We recommend using Tag Manager because it’s faster and more convenient.)
3. Consider what events and settings you want to collect into the new resource type
4. Use both resource types at the same time to compare how data is collected
5. Please note:
   1. You can add only one Firebase project to a single Google Analytics 4 resource
   2. However, you can configure multiple data streams from different applications into a single Google Analytics 4 resource

1. Google Analytics 4 is the most profound update ever to the logic of Google Analytics. Now everything is built around events, event parameters, and users — not around sessions as it was before.
2. Cross-platform analytics between your website and applications out of the box is one of the key features and drivers of Google Analytics 4.
3. You can use an already configured Google Analytics resource via gtag. js or GTM to configure a new Google Analytics 4 resource.
4. When you set up Google Analytics 4, it automatically creates a new WP resource, and only from the time you set it up will data be collected. No data is migrated from older WPs.
5. The Google team doesn’t urge everyone to abandon the old Google Analytics and switch to the new one. They recommend running the new Google Analytics 4 in parallel with Google Analytics and starting to collect data into it. The source of historical data remains the standard Google Analytics.
6. While the new Google Analytics 4 has flaws and not all features are available yet, the developers are rolling them out gradually.
7. You can set up free uploading of data from Google Analytics 4 to Google BigQuery. The export scheme is the same as for Firebase.
8. You can already configure Google Analytics 4 resources and start collecting data. The sooner you configure resources, the more historical data will be collected.