Google Analytics 4: An Introduction to Predictive Metrics
If you’re a marketer, you want to be able to anticipate your customers’ needs and give them what they want before they even know they want it. That’s where predictive metrics come in. Predictive metrics are a new feature in Google Analytics 4 (GA4) that use machine learning to help you do just that. In this article, we’ll give you a brief overview of predictive metrics and explain how and when they can benefit your business.
Table of Contents:
- What are Predictive Metrics?
- Types of Predictive Metrics in GA4
- Prerequisites
- Are Predictive Metrics Available for All Websites?
- How to Use Predictive Metrics?
- Conclusion
What are Predictive Metrics?
Predictive metrics are a new feature in GA4 that uses machine learning to predict future customer behavior. Based on historical data, predictive metrics can tell you which customers are most likely to convert, which channels are most likely to result in a conversion, and which devices your customers are most likely to use.
Predictive metrics are important because they allow you to anticipate your customers’ needs and take action accordingly. With predictive metrics, you can proactively reach out to customers who are likely to convert and provide them with the information they need to purchase. You can also adjust your marketing campaigns based on predicted customer behavior to reach the right people at the right time with the right message.
Types of Predictive Metrics in GA4
Currently, there’re three types of predictive metrics in GA4. They include:
- Purchase probability . The probability that a user who was active in the last 28 days will log a specific conversion event within the next seven days.
- Churn probability . The probability that a user who was active on your app or site within the last seven days will not be active within the next seven days.
- Revenue prediction. The revenue expected from all purchase conversions within the next 28 days from an active user in the last 28 days.
Only purchase/ecommerce_purchase and in_app_purchase events are supported for the Purchase probability and Revenue prediction metrics.
Simply put, they’re based on your website visitors’ data, structured event data, and Google’s machine learning algorithms. They generate predictions about the future behavior of your users. For your convenience, we illustrated it with a matrix below.
As you can see from the table, predictions are made for the near future and based on the recent past.
Prerequisites
Since predictive metrics are derived from Google machine-learning algorithms, you must meet several criteria to train the predictive metrics. They’re the following:
- The least possible number of positive and negative samples is 1,000. The least number of non-operative or inactive users is 1,000 as well. By positive, we mean that you should have 1,000 purchasers or conversions, while negative means that the other 1,000 users haven’t purchased anything.
- As highlighted above, GA4 uses three types of predictive metrics. To trigger them, you need to set up the purchase event and send it to the GA4 property.
- The Model quality should be maintained over a particular duration, which is generally 28 days.
You’ll get predictive metric updates daily if you’ve met these prerequisites. Otherwise, you’ll stop receiving updates. To check the status of each prediction provided by GA4, go to the Predictive section within Suggested audiences templates in the Audience builder.
An audience will be marked as “Not eligible to use” if there’s insufficient data to use predictive metrics.
Are Predictive Metrics Available for All Websites?
To benefit from predictive metrics, you need to follow some basic steps. First of all, you need to have Google Analytics 4 installed. If you haven’t done it yet, it’s time to migrate because Google deprecates Universal Analytics on July 1, 2023.
When your GA4 property is set up, enable benchmarking data sharing. This will send anonymous data to the predictive learning models and improve the quality of predictions for your account. To do this, go to the Admin tab, click Account Settings, and choose Benchmarking.
Then click Save at the bottom of the page.
The second step is to track purchase events in GA4. Generally, predictive metrics are possible when you have configured one of the three events in your GA4 property:
- purchase: your property needs to send it, and you should collect the currency and value parameters to be eligible for both purchase and churn probability
- ecommerce_purchase: currently, GA4 is not a full-scaled eCommerce analytics tool yet; for now, you can best use the purchase event
- In_app_purchase: it’s collected automatically, but you need to link your account to your Firebase account for Android apps
To check if you have set up the required events, go to Configuration → Events and look at the list of events.
How to Use Predictive Metrics?
Predictive metrics are available in the audience builder and Explorations.
Audience Builder
GA4 predictive audiences are groups of users that share similar characteristics and behaviors. These groups are created by Google Analytics 4 using machine learning algorithms. By understanding the common characteristics of these groups, you can better target your marketing messages and campaigns. With predictive metrics, you can use the following types of custom Audiences:
-
- Likely 7-day purchasers. These are users that are likely to purchase within the next seven days. The predictive metric used in this audience is Purchase probability greater than the 90th percentile (more than 90%).
-
- Likely first-time 7-day purchasers. These users will likely make their first purchase within the next seven days. The predictive metric is Purchase probability greater than the 90th percentile (more than 90%)ANDLTV equals 0.
-
- Likely 7-day churning users. They’re active users who will not likely visit your website within the next seven days. The predictive metric used in this audience is Churn probability is greater than the 80th percentile (more than 80%).
-
- Likely 7-day churning purchasers. They’re active users who are unlikely to visit your website within the next seven days. The predictive metric used in this audience is Churn probability is greater than the 80th percentile (more than 80%)ANDLTV greater than 0 ORPurchase event
OR
Ecommerce_purchase .
- Likely 7-day churning purchasers. They’re active users who are unlikely to visit your website within the next seven days. The predictive metric used in this audience is Churn probability is greater than the 80th percentile (more than 80%)ANDLTV greater than 0 ORPurchase event
- Predicted 28-day top spenders. These are users that are likely to generate the most revenue in the next 28 days.
Also, you can use predictive metrics in the Analysis tab when you create the User Lifetime report. All you need to do is to follow a few steps:
-
- Navigate to your Google Analytics 4 property and select Explore from the menu
-
- There you’ll see different exploration templates. Click on the plus sign and select the blank template
-
- Click on the drop-down in Tab settings under Techniques and select User lifetime
-
- Click on the + sign menu under Metrics in the Variables tab
-
- An overlay will appear on the right side; click on Predictive metrics from the list
-
- All predictive metrics will be further divided into three types:
If you click on any of these metrics, you’ll also see a few submetrics. Submetrics show the probability of users being assigned to the main predictive metric. Let’s open Purchase probability, for example.
10th percentile means a 10% probability that a user who has been active on your site within the last 28 days will purchase within the next seven days. Average means the average probability that a user who has been active on your website within the last 28 days will purchase within the next seven days.
-
- Then select Churn probability and Purchase probability and click Apply.
- You’ll see the selected predictive metrics in the Metrics list under the Variable tab. Now you should double-click on any of the predictive metrics to add them to the report.
Once completed, a User Lifetime report will be created using predictive metrics. It will look similar to the one below:
Conclusion
Predictive metrics are a powerful new tool that can help you stay ahead of your competition by anticipating your customers’ needs. If you’re not using predictive metrics yet, now is the time to start! Get in touch with our experts to unlock the power of predictive metrics, or if you haven’t installed GA4 yet, contact us for a smooth transition.
Get in touch
Got a question? We'd love to hear from you. Send us a message and we'll respond as soon as possible.
By clicking submit, you agree to our Privacy Policy
Latest Insights
Get the tips from our experts to optimize and scale your campaigns