All Google Analytics accounts have now been migrated from Universal Analytics to GA4. This is a significant change, especially for businesses in the advertising industry, where performance is often evaluated on a session-by-session basis. GA4 measures performance on a user-by-user basis, which requires businesses to review their advertising targeting and evaluation methods.

One of the key features of GA4 is Predictive Audience. This powerful tool uses machine learning to identify users who are most likely to take a desired action, such as making a purchase or signing up for a newsletter. By targeting Predictive Audiences with ads, businesses can improve their campaign performance and reach the right people with the right message at the right time.

What exactly is GA4 Predictive Audience?

Machine learning is now a key part of Google Analytics 4 (GA4), and it’s poised to revolutionize the way we target and measure advertising campaigns. In Universal Analytics (UA), machine learning was used in “smart goals”, but with GA4, it’s used in a much wider range of features, including “Predictive Audiences”.

Predictive audiences are a powerful new tool that allows you to target users who are most likely to take a desired action, such as making a purchase or signing up for a newsletter. GA4 uses machine learning to analyze data from various events on your website, such as page views, product views, and add-to-cart events, to identify these users.

Once you’ve created a predictive audience, you can use it in Google Ads to target your ads. This can help you to improve your campaign performance and reach the right people with the right message at the right time.

For example, imagine that a user adds an item to their cart on an e-commerce website. Many users who shop at this store tend to visit the relevant help pages to check the delivery time and shipping costs. 

In the past, marketers would need to identify this behavior through manual analysis. However, GA4 Predictive Audience can use machine learning to analyze the behaviors that users often take to make a purchase without the need for manual verification. This is a truly revolutionary feature that can help marketers to create audience lists of users who are highly motivated to purchase.

While machine learning is already being used extensively in Google Ads, there is a major difference between GA4 and Google Ads: machine learning in Google Ads is mainly based on data on responses to ads, while GA4 learns behavior within the user’s own site. By basing its predictions on both on-site and off-site behavior, GA4 is expected to be able to reach users with even higher value.

5 types of Predictive Audiences

So far, we have touched on an overview of the functionality, but now it is time to put it into practice. There are five types of predictive audiences available. I would like to briefly list the usage scenarios for each of them.

1.  Existing customers who are likely to leave within 7 days

• Distribute special offers to a limited number of users to encourage them to return to the store.
• Based on the predicted audience, further narrow down the list of users to those who have accumulated a certain amount of points, and distribute the special offer.

2.  Users most likely to leave within 7 days

• Place ads that guide users to special offers, which recommend products with a high likelihood to convert.

3.  Existing customers likely to purchase within 7 days

• Enable user exclusions from ad targeting because certain users are more likely to buy, and you’d want to exclude them from ad targeting and maximize the return on ad spend.
• Increase bids and acquire impressions to avoid losing out to competitors in search advertising.

4.  Users most likely to make their first purchase within 7 days

• Run ads that appeal to acquire first-time buyers.
• Run ads that highlight the unique benefits of your shop, offer, or product to lower the psychological barriers of first-time purchasers.

5.  Users predicted to be the top spenders within 28 days 

• Exclude fans from ad serving targets, as they are likely to make purchases without spending money on advertising.
• Distribute ads promoting high-value products.
• Distribute ads promoting new products, as fans are likely to be interested in them.

It will help attract the interest of users by inferring the psychology and behavior of each user and delivering ads with appeals tailored to them.

Useful for data visualization of campaign performance

“Predictive Audiences” can be used for more than just targeting. By setting audience triggers, you can generate an event when a user matches the conditions of the audience list and is added to the list. This event data can be used to visualize the performance of awareness-raising measures.

Traditionally, awareness-enhancing advertising performance is evaluated using pseudo KPIs such as new traffic, reach indicators, and assist conversions. However, by combining predictive audiences and audience triggers, you can supplement and visualize the change in user attitude toward purchase as a conversion.

For example, you could create a predictive audience of users who are likely to make their first purchase within 7 days. Then, you could deliver ads to this audience that are aimed at increasing their willingness to purchase and moving them into the top of the funnel.

You could also create an audience list of users who have visited your site and whose purchase probability is in the middle (List A). You could then distribute ads to this list. Additionally, you could create an audience list of users with high purchase probabilities (List B). When a user is added to List B, you could use an audience trigger to fire a specific event.

If you regard reaching the top audience as the conversion for awareness-enhancing ads, you can visualize the degree to which your advertising is contributing to increasing the purchase intent of List A users. You can also import the event into Google Ads and set it as a campaign conversion so that you can optimize your campaign for increased willingness to purchase.

The event is triggered when a user matches the criteria in the audience list and is added to the list (source: Principle).

The key to unlocking your marketing potential

In the past, marketers relied on their experience and intuition to identify users with high purchase intent and analyze their characteristic behaviors. However, machine learning has now revolutionized the process, allowing us to leverage analytics based on volumes of data that humans cannot process.

If you’re not already using predictive audiences, I encourage you to implement GA4 on your site and take your ad serving to the next level. 

Predictive audiences are a powerful tool that can help you improve your advertising performance and reach your target customers more effectively. By taking advantage of predictive audiences, you can stay one step ahead of the competition and achieve your marketing goals.