The advanced segmentation feature in Universal Analytics has been a popular tool which has allowed organizations to group their users together based on shared attributes, such as demographic characteristics and past behavior. In Google Analytics 4 (GA4), this feature has been taken to the next level with the introduction of Predictive Audiences, a tool which now allows organizations to group and analyze users based on how they are likely to behave in the future, rather than solely relying on their past behavior. Using various machine learning techniques, Google can make predictions about users’ likelihood to make a purchase, their expected revenue, or their likelihood to churn. These predictions are then used in predefined audience segments in GA4, such as “likely 7-day purchasers” or “predicted 28-day top spenders”, or can be combined with other conditions to build out custom audiences.
Since Predictive Audiences are shared with any advertising accounts linked to your GA4 property (e.g., Google Ads, Display & Video 360, Search Ads 360), they can be a powerful tool for driving media performance across the entire Google Marketing Platform. Three examples of Predictive Audiences being leveraged to enhance media effectiveness are similar audience targeting, retargeting, and audience suppression.
The Similar Audiences feature in Google Ads is a powerful yet simple tool which can be used to efficiently grow your organization’s reach by targeting new and qualified customers based on their similarities to your existing high-value audiences. The addition of Predictive Audiences in GA4 has augmented the benefits of Similar Audience targeting even further by introducing new high-value audiences that can be used as a basis for lookalike modeling.
For example, using Predictive Audiences, you can define an audience of the 20% of your users who are most likely to purchase. By pushing that audience to Google Ads, you can then grow your reach by running campaigns that will specifically target net-new users who are similar to the people who are already most likely to buy from you. This is a much more efficient way to scale your audience than simply increasing your budgets, buying more keywords, expanding your targeting criteria, and so on.
Predictive Audiences in GA4 have made retargeting strategies more effective than ever by streamlining the process of identifying users who are likely to convert. Google analyzes website behavior to look for users with strong indicators of converting, such as studying product details or adding items to a cart, and uses machine learning techniques to uncover patterns based on your unique GA property to make predictions. You can then follow up with users from Predictive Audiences, such as likely 7-day purchasers who are on the threshold of converting, and target them with specialized marketing to encourage them to complete the conversion process. Predictive Audiences based on revenue, such as predicted top spenders, can also be targeted with specialized marketing tactics, which can help increase overall revenue.
On the other hand, Predictive Audiences can also be used to easily identify users who have previously engaged with your website or made a purchase who are at risk of churning. Using suggested audiences in GA4, such as likely 7-day churning purchasers, marketers can improve customer retention by retargeting users with specialized messaging or offers. By reengaging these audiences and reminding them of the value your business can offer them, you can prevent churn and drive an ongoing and loyal base of customers.
Alongside remarketing to users who may need extra encouragement to convert, the Predictive Audience feature can also be used to suppress advertising to audiences who may already have a high likelihood of converting. For example, the top 10% of customers who have the highest likelihood of converting within the next 7 days may already have made the decision to convert and likely do not require additional advertising. Since these users are already likely to convert, you can consider suppressing ads for this Predictive Audience to save media spend and invest that money in marketing to other audiences who might need additional encouragement to convert, thereby making the most efficient use of your marketing investments.
Read our post about Audience Creation and Management in GA4 to learn more about audiences across Google Analytics.
Predictive Audiences are built using conditions that are based on predictive metrics. Unlike standard metrics in GA, which measure observed behavior, predictive metrics rely on machine learning to train predictive models to identify patterns in user behavior. These models are then used to make predictions about how users will behave in the future.There are currently three predictive metrics available in GA4 which can be used to build audiences:
After the prerequisites for the predictive metrics have been met (see below), you can select from a variety of suggested Predictive Audiences that include users who exceed certain thresholds based on the metrics. For example, the “Likely 7-day purchasers” uses the purchase probability metric to include users whose likelihood of churning is above the 80th percentile, which will return the top 20% of users who are most likely to purchase in the next 7 days. These percentiles can also be adjusted, as shown in the image below. In addition, you can layer in additional criteria like Region or Device Type to narrow down the audience.
Before taking advantage of the benefits that Predictive Audiences have to offer your organization, a few prerequisites have to be met. First, your organization must have a property set up in GA4, as the Predictive Audiences feature is not available in Universal Analytics. Second, Predictive Audiences require a sufficient amount of positive and negative examples of users who churned or made a purchase within 7 days, over the last 28 days. For example, to satisfy conditions for the purchase probability metric, there need to be at least 1,000 returning users who have made a purchase during the required time period, and at least 1,000 returning users who have not made a purchase. Finally, the model quality must be sustained over a period of time to be eligible for use.
For the first time, Google Analytics gives marketers the ability to tap into advanced machine learning techniques like predictive modeling directly within the GA platform itself. Now, instead of taking action on insights that are purely retrospective, you can proactively take action based on unique insights into the future.
Predictive Audiences in GA4 dramatically simplifies the process of predicting users’ future behavior, which can be a powerful tool for your organization. Predictive Audiences can turn GA4 into an audience hub, which can be used to maximize the efficiency of your media investments across the entire Google Marketing Platform. More specifically, this will help you optimize performance and grow your business by empowering you to retarget more effectively, reallocate media budgets to focus on users who need that extra nudge in order to convert, and profitably grow your reach.
As consumers become increasingly digitally savvy, and more and more brand touchpoints take place online,…
Marketers are on a constant journey to optimize the efficiency of paid search advertising. In…
Unassigned traffic in Google Analytics 4 (GA4) can be frustrating for data analysts to deal…
This website uses cookies.