What Can Ga4 Predictive Audiences Do And How Do I Use Them In My Dv360 Campaigns
Management Summary
Get the most out of GA4 Predictive Audiences and learn how to create and use them in your campaigns!
What are Predictive Audiences?
A predictive audience is a target group with at least one condition based on a predictive metric. For example, you can target users who are likely to make a purchase in the next seven days.
Predictive metrics are ML-enriched metrics that can predict the future behavior of users. These are: Purchase Probability, Churn Probability and Predicted Revenue.
The availability of Predictive Audiences depends on whether the underlying Predictive Metrics meet all requirements for use:
- You need a minimum number of positive and negative examples of buyers and churned users. At least 1,000 returning users must have triggered the Predictable condition (purchase or churn) on 7 days in the previous 28 days. This must not have been the case with at least 1,000 users.
- The model quality must be maintained over a certain period of time.Learn more about what you can do to make your property eligible for forecast metrics
- In order for purchase probability and forecast sales data to be returned for a property, it must pass the “purchase” events (recommended for manual entry) and/or “in_app_purchase” (automatically recorded) send. When you capture the “purchase” event, you must also capture the “value” and “currency” parameters for the event.More information
Create predictive audiences
Predictive audiences can be created very easily in the GA4 interface using existing templates.
In GA4, under “Admin” (gear icon), navigate to the “Data Display” section and click “Audiences”. Then click on “New audience” (you need at least marketer role at property level for this)
Then click on “Predictive” under “Suggested Audiences”.
Suggested Predictive Audiences that meet the requirements for predictive modeling are marked “Ready to use”. Click on one of the templates and adapt it to your requirements using the target group lists.
If you want to edit a prediction condition when creating an audience, select one of the configuration options:
- Most likely… (including the top N% of users)
- Least Likely… (including bottom N% of users)
- Custom (enter percentage range or use slider on right to select a range)
If you use a custom range, you can use the sliders to see how many users are included and how likely they are to meet the predicted condition. A larger range will contain more users, but a larger proportion of them will be less likely to meet the condition.
Advantages and areas of application of Predictive Audiences
One of the advantages of GA4 Predictive Audiences is that they can be used across all linked products, such as Google Ads, DV360 and SA360.
The use of these audiences is particularly useful in remarketing.
Users who are close to a conversion are easier to convince to carry out that conversion or key event. For example, if users have viewed product details or added items to their shopping cart, they are signaling a strong interest in these products. With a convincing follow-up campaign via a well-thought-out remarketing campaign, you can give users the final push to complete the conversion.
The area of application in campaigns where users who have lost their interest for some reason should be re-engaged is also extremely valuable. Such users have a higher churn rate, but they have also been interested in the past.
With this strategy, you can appeal to these users again and remind them, for example, of the variety of products, their quality and price, or the practical shipping and return options. You can also show users how important they are to you with special offers.
Conclusion:
In summary, Predictive Audiences in GA4 offer a way to use data-driven predictions to design personalized and efficient marketing campaigns that can both strengthen customer loyalty and increase conversions. With GA4’s machine learning, deep behavioral patterns specific to your property can be determined, which ensure a particularly valuable database for remarketing and re-engagement campaigns.
