GA4 Explained: What It Is, How It’s Different & Everything You Need to Know

Posted on Categories Analytics

Table of contents

If you’re a digital marketer or operate a business online and use Google Analytics, you’ve probably seen or heard some chatter about GA4. It was officially released in October 2020, when it became the new default Analytics property type when setting up new accounts.

As easy as Google makes it to set up, there’s not a ton of great information out there about why you should set up a GA4 property. If you’re wondering what GA4 is, what it does, and why in the world you should care – this one’s for you!

An overview of Universal Analytics (UA)

Before we get into the latest and greatest of GA4, let’s talk about what it’s replacing. Universal Analytics was the latest data collection technology for web-based analytics before the release of GA4 in October 2020. It runs on Google’s analytics.js framework for web properties, an SDK for mobile apps, and the Measurement Protocol for other devices.

UA is being replaced by GA4, which is an all-encompassing analytics framework built to measure websites, apps, and other digital platforms.

But what is GA4, why did Google make it, and why should marketers or business owners care? Before we answer that, let’s define what Universal Analytics is, how it works, where it’s lacking, and why GA4 is the analytics platform of the future.

How Universal Analytics works

Universal Analytics (UA) code lives on your website and waits for an interaction that causes data to be sent to GA, known as a hit. A hit can take many forms (called ‘hit types’), but the most common are:

  • page tracking hits (pageview)
  • event tracking hits (i.e. a contact form submission)
  • ecommerce tracking hits (purchases)
  • social interaction hits (i.e. Facebook likes)

When a hit occurs, data is sent to GA. This data includes:

  • HTTP request information such as hostname, referrer, browser type, and language
  • Browser/system information such as Java and Flash support, and screen resolution
  • DoubleClick cookie data such as interests, demographics, etc…
  • Client ID (more on this later)
  • Google Analytics will also set and read first-party cookies to obtain user session data and ad campaign information associated with the page request

For example, this is the data sent in a pageview hit:

v=1              // Version.
&tid=UA-XXXXX-Y  // Tracking ID / Property ID.
&cid=555         // Anonymous Client ID.

&t=pageview      // Pageview hit type.
&   // Document hostname.
&dp=/home        // Page.
&dt=homepage     // Title.

This means that each report in UA contains data derived from hits. GA4 is very similar to UA in that user interactions prompt data to be sent to Analytics, with one big difference; there are not different hit types in GA4, there are only events.

In summary, Universal Analytics code sends hit data to Analytics whenever prompted to do so by user interaction. For example, when a pageview hit takes place, Analytics will receive the pageview and populate the following metrics and dimensions:

  • Device type
  • Browser
  • Users
  • Sessions
  • Time/day
  • Page URL
  • etc… 

At this point, you might be wondering, what’s wrong with Universal Analytics? That’s a fair question and one that most marketers are thinking about themselves. One of the most glaring issues with UA is how it measures and defines users.

How a user is defined in Universal Analytics

If you’ve looked into Google Analytics for your business before, you may have wondered what the difference is between users, new users, sessions, page views, and all of these other metrics.

Say that you needed to answer the simple question “how many people were on my website last month?” How would you go about doing this? You’d probably just check the users metric, since users are people, right?

Kind of. This is the first problem with UA.

Inside of all of the hit data sent from your website to Analytics is what is known as the Client ID. This randomly generated string gets stored in browser cookies after a new user has sent a hit to Analytics for the first time so that subsequent visits can be associated with the same user.

This means that the users metric in Universal Analytics isn’t really the number of people who have visited your website. Rather, it’s the number of Client IDs that have sent hits to your Analytics property.

Ok, but why would the number of users (client IDs) be different from the real number of people that have visited my site?

Great question!

We’ll use a common scenario from the vacation rental industry as an example. Imagine that someone named Chris is at work and thinks about planning a getaway. They find a website that they trust and some rental properties that they like. Later that night, they’re on their phone looking at those same properties. The next day, they book a property on their home computer.

How many users would Google Analytics track in this situation? The answer is three. One from the work computer, one from the cell phone, and one from the home computer. This is not exactly ideal.

image showing the same person visiting a website on a phone, laptop, and computer to illustrate how one person can access a website multiple ways, and be tracked as multiple users.

What about User ID?

User ID was created with the scenario above in mind, as a way to track users more accurately by leveraging an authentication system to keep track of who is who. This requires working with a developer to create an authentication system that users can log into, and must assign each user in the system with a unique and persistent ID that is sent to Analytics.

User ID is not turned on by default and requires creating a separate view from your main Analytics view, as standard views can not be converted to or integrated with a User ID view. 

Also, if a user doesn’t log in, then no data will be sent to GA since there is no User ID to match it to. So in many cases, all of that fancy developer work will be for nothing. 

Why did Google create GA4?

In short, Google created GA4 for three main reasons: more accurate user tracking across devices and platforms, integrate a forward-thinking approach to privacy, and better integrate AI and machine learning with web analytics.

The first reason that GA4 was created was to build a web analytics platform that could work across devices and platforms. As we’ve already discussed, UA doesn’t have many great options for tracking users on various platforms and devices, which tends to be how people engage with websites and apps nowadays.

Another big reason that Google created GA4 was to better adhere to the recent rise in privacy laws such as GDPR and CCPA. As people begin to become aware of their privacy online, more laws and requirements will be passed that websites must adhere to. GA4 helps by creating three different levels of user tracking (more on this later), as well as offers AI-powered machine learning that can give us insight into those users who refuse to be tracked at all.

How is GA4 different from Universal Analytics?

In short, GA4 completely overhauls the way that we collect data and gives us the ability to track a single user across multiple devices and platforms. While Universal Analytics was entirely dependant on browser cookies storing Client IDs, or websites with an authentication system sending User IDs, GA4 introduces the concept of reporting identities that fills in the gaps prevalent in UA.

Earlier we mentioned that in UA, a view with User ID set up must be separate from a view utilizing the standard Client ID. GA4 solves this limitation by introducing three different ways to identify visitors to your website, and all reporting identities can be included in the same reports. 

The three reporting identities offered in GA4 are:

  • Device ID (formerly Client ID) – This is what used to be known as Client ID in Universal Analytics, and is stored in the client browser as a first-party cookie.
  • User ID – This is similar to User ID from UA in that a website or app must be configured to support it with an authentication system that generates a unique and persistent ID for each user (typically a login system).
  • Google Signals – Google Signals is a way for GA4 to identify a user across devices based on whether or not they’re signed into other Google products and have opted-in to Ads Personalization.

GA4 properties have the option to change the default reporting identity. You may set the default reporting identity to “By device only” or “By User-ID, Google signals, then device.”

  • By device only will only utilize device ID. This is not ideal, as many users are opting out of using cookies.
  • By User-ID, Google Signals, then Device is the ideal setting for GA4. With this option we tell Google to utilize all methods of user identification, first looking for a user ID, then  Google Signals data, before finally settling on device ID. In short, this is the most accurate and comprehensive setting, as well as the one that we recommend using.

Say, for example, our friend Chris from earlier was signed in to the same Gmail account on their work computer, home computer, and phone. Through Google Signals, GA4 would be able to identify all three of Chris’ sessions and attribute them to the same user, as long as Chris has enabled Ads Personalization.

The best part about GA4 is that all three reporting identities can function in the same property. This means that you can count both users who are logged in and users who are not logged in under the same Google Analytics view.

GA4 will also allow anyone to use advanced analysis reports, which used to only be an option for Analytics 360 users 🎉.

User ID vs Client ID vs Google Signals

Client ID (Device ID)User IDGoogle Signals
Cross-device user tracking?NoYesYes
Developer implementation required?NoYesNo
Available in UA?YesYesNo
Available in GA4?YesYesYes

What can marketers do with GA4?

1) Predict the future

While you might not be able to know next week’s lottery numbers, GA4 enables you to predict several metrics that are important to marketing campaigns. This is made possible by the introduction of Google’s advanced machine-learning to your Google Analytics data. GA4 will look at the current and past behavior of your users and use that information to predict what the future might look like. 

So what kinds of things can the GA4 crystal ball tell us? Here are three major metrics that will change our approach to marketing:

Purchase probability – this metric defines the probability that a user who was active on your site in the last 28 days will log a specific conversion event within the next 7 days.

VR insight: This is particularly useful for the vacation rental industry because it helps identify which users might be the closest to booking a property in the near future.

Churn probability – have you ever wondered how many people on your website are just window shoppers? Churn probability will give us more insight into which users are engaged and which aren’t. Technically speaking, churn probability is the probability that a user who was active on your site in the last 7 days will not be active within the next 7 days.

VR insight: We can use this metric to exclude website users with a high churn probability from our remarketing lists, that way we can limit wasteful spending and get in front of an audience that’s more likely to book a property.

Revenue prediction – this is another conversion-based predictive metric that measures the revenue expected from all ecommerce purchase conversions within the next 28 days from a user who was active in the last 28 days.

VR insight: This metric can be used to identify which users might be the most valuable over the next 28 days. This information could then be used to create a predictive audience for targeting in Google Ads.

2) Measure engagement more intelligently

It’s finally time to kiss bounce rate goodbye!

As digital marketers and business owners who operate online, we’ve spent a great deal of time talking about bounce rates. Technically speaking, a “bounce” is when a user views one page on your website and leaves before taking any action or viewing another page. The bounce rate is, well, the rate at which users bounce.

This isn’t a particularly useful metric.

Think about how you interact with website content in your personal life. News articles, dinner recipes, blog posts, and so many other types of content can all be consumed by a user and still lead to a “bounce.” Website interactions are complex, and reducing their outcomes to a metric as simple as “bounce” leaves so much valuable information on the table.

Enter GA4, which replaces the idea of a “bounce” with what is now called an Engaged Session. For a session to qualify as an engaged session, the following must be true: the session ended with 10 seconds or more of engagement time, or had 1 or more conversion events, or had 2 or more page views.

Engagement time vs. session duration

In Universal Analytics, session duration was calculated by finding the difference between when the first hit of a session and the last hit of a session. This is not a very accurate metric, as users often have multiple tabs, windows, or apps running, and may minimize the window that your website is on. This leads to an inflated session duration metric.

GA4 is much smarter and does not count time during which the website is not running in the foreground towards the overall engagement time. As you can see, this is much more sophisticated than bounce rate, and as marketers in the 21st century, we are happy to leave bounce rate behind.

3) Market in the same world that we live in

Simply put, Universal Analytics was not built for a world in which one person uses multiple devices to access a website or app. And if you’re a business that engages users on both a website and mobile app, you can imagine how impactful it will be to have the ability to analyze all of your user metrics in one dashboard.

Even if you only operate a website, the addition of Google Signals as a means to track individual users across devices will yield much more accurate reporting. 

When should I switch to GA4?


That’s it, that’s the whole section. Do it now.

Done it yet? No? Go do it!

In all seriousness, the best time to get started with GA4 is today. While GA4 isn’t ready to fully replace Universal Analytics (yet), Google has announced that UA will be depreciated at some point, but we don’t know exactly when. Since GA4 isn’t fully ready yet, it’s not time to remove your UA code and switch over.

With that being said, you can (and should) install GA4 on your site and run it in parallel with the existing UA implementation. This way, you have plenty of historical data when you do decide to start using GA4 full-time.

If you wait until you’re ready to use GA4 to add the tracking code, you’ll be starting fresh with no data.

Be sure to check back next month for another GA4 article that dives deeper into the modeling & AI capabilities of GA4, and how it will fill the gaps in your data. We’ll also take a closer look at reporting in GA4, and give you some helpful tips and tricks to get off the ground quickly.

Better get trackin’! Not sure how? Hit up the marketing ninjas at ICND.

Author: Mike Doute

Mike is a full-stack marketer originally from Detroit, but currently living in coastal North Carolina. When he's not working with SEO/SEM clients, you can usually find him listening to classic rock, enjoying the local craft beer scene, or taking care of plants.