Say hello
Let us know a little bit about you and your business, and we’ll get back to you.
Or contact us now on
As a marketer, staying ahead of your customers’ needs and delivering what they want before they even realise it is the ultimate goal. Predictive metrics, a new feature in Google Analytics 4 (GA4), play a vital role in achieving this objective by harnessing the power of machine learning. These metrics enable you to anticipate user behavior and preferences, allowing you to make proactive decisions.
In this article, we’ll provide a concise overview of predictive metrics and explore their potential benefits for your business. We’ll also discuss the best practices and optimal scenarios for implementing predictive metrics to maximise their impact on your marketing efforts.
What are Predictive Metrics?
Predictive metrics in GA4 utilise machine learning to forecast future customer behavior. Drawing insights from historical data, these metrics can accurately identify the customers most likely to convert, the channels with the highest conversion potential, and the preferred devices of your customer base.
Predictive metrics are valuable tools for marketers and business owners as they allow you to be proactive rather than reactive in your strategies. By understanding and leveraging these predictions, businesses can optimise their marketing campaigns, personalise user experiences, and ultimately improve their overall performance and ROI.
Key Predictive Metrics in GA4
Churn Probability
Definition: The probability that a user who was active on your app or site within the last 7 days will not be active within the next 7 days.
Churn probability is a crucial metric that helps predict the likelihood of a user discontinuing engagement with your platform. By identifying users who are at risk of churning, businesses can implement targeted retention strategies to retain valuable customers and reduce overall churn rate.
Purchase Probability
Definition: The probability that a user who was active in the last 28 days will log a specific conversion event within the next 7 days.
Understanding the probability of a user making a purchase is vital for e-commerce businesses. GA4’s purchase probability metric leverages historical data to estimate the likelihood of a specific user converting, aiding in refining marketing strategies to boost conversion rates.
Predicted Revenue:
Definition: The revenue expected from all purchase conversions within the next 28 days from a user who was active in the last 28 days.
Revenue prediction is a powerful metric that forecasts the potential revenue a business can generate over a specific period. This insight allows businesses to set realistic revenue goals, allocate resources effectively, and make informed decisions to drive growth.
Prerequisites
Due to the nature of GA4 being tied to machine learning, in order to train the predictive metrics, users must fulfill a set of requirements beforehand.
Enough Examples
There should be a minimum number of positive and negative examples of purchasers and users who stopped using the product or service (churned users). Within the last 28 days and over a seven-day period, at least 1,000 returning users must have triggered the relevant predictive condition (like making a purchase or churning), and at least 1,000 returning users must not have done so.
Consistent Model Quality
The model’s accuracy must be consistently high over time to be considered eligible. Taking specific actions can improve your chances of meeting the criteria for predictive metrics eligibility.
Event Collection
To be eligible for purchase probability and predicted revenue metrics, your property must send purchase (recommended for collection) and/or in_app_purchase (collected automatically) events. When you collect the purchase event, you also need to collect the value and currency parameters for that event. Learn more here.
Reminder: Once you’ve met the prerequisites, you’ll receive daily updates for predictive metrics. However, if the numbers fall below the standard requirements, you’ll no longer receive updates.
Best Practices
Enable Modeling Contributions & Business Insights
In your data-sharing settings, ensure that the Modeling contributions & business insights setting is turned ON. This allows Analytics to utilise shared aggregated data, enhancing model quality, and improving your predictive insights.
Utilise Event Recommendations
Maximise the use of event recommendations in your property. These recommendations help enhance the accuracy of predictive metrics.
Collect Relevant Events
Ensure you are collecting the purchase and/or in_app_purchase events. While in_app_purchase events are automatically collected, you need to link your Android app to Google Play via your Firebase account to see the in_app_purchase event. It is now recommended to use the purchase event instead of ecommerce_purchase.
Diverse and Meaningful Events
Collect a broader range or higher volume of recommended events that correspond to user behavior. This will contribute to improving predictive models. Conversely, minimising noisy events that do not hold significant meaning in terms of user behaviour will also lead to improved predictions.
Are you intrigued by the potential of predictive metrics in Google Analytics 4, but find it challenging to implement and grasp? We understand that the world of data analytics is constantly evolving, and as a business owner, you need to continuously learn and adapt. However, instead of spending valuable time deciphering complex data problems, wouldn’t you rather focus on nurturing more meaningful relationships with your customers?
Contact Metric Labs to discuss how we can take care of your data analytics so you can focus on delivering great customer experiences driven by accurate data and insights.
Like this blog post? Sign up to our email newsletter – Lab Report – and never miss a new one. Or, get it sent straight to your Messenger!