What Are Predictive Metrics in Google Analytics 4?

Your Google Analytics 4 property is fantastic at telling you what happened yesterday, last week, or last month. But what if it could help you predict what's going to happen next? That's exactly what predictive metrics are designed to do. This guide breaks down what these AI-powered features are, the requirements for unlocking them, and how you can use them to make smarter, more proactive marketing decisions.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

What Are Predictive Metrics, Really?

In short, GA4 predictive metrics use Google's machine learning models to analyze your existing data and forecast a user's future behavior. Instead of just reporting on past actions like pageviews and purchases, it identifies patterns among your users to make educated guesses about what they're likely to do next.

Think of it like this: traditional analytics provides the rearview mirror, showing you everywhere you’ve been. Predictive metrics are like a GPS showing you potential routes and ETAs for where different customer segments are headed.

This allows you to shift from a reactive to a proactive strategy. Instead of analyzing last month's churn report and wondering what went wrong, you can identify users who are at risk of churning and intervene before they’re gone for good. It's about getting ahead of the curve.

The Three Core Predictive Metrics in GA4

GA4 focuses on a few key behaviors that are critical for e-commerce stores and lead-generation businesses. These predictions are centered around a user who has been active on your site or app within the last 28 days.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

1. Purchase Probability

What it is: This metric calculates the probability that a user who was active in the last 28 days will make a specific conversion event (like a purchase) within the next 7 days.

How to use it: Purchase Probability is perfect for identifying users who are on the verge of converting. Imagine a user who has visited a product page multiple times, added an item to their cart, but hasn't completed the purchase. This user would likely have a high purchase probability score. You can group these "window shoppers" into an audience and target them with a remarketing campaign on Google Ads offering a small discount or free shipping to nudge them across the finish line.

2. Churn Probability

What it is: Churn Probability is the inverse of engagement. It predicts the likelihood that a user who was recently active will not visit your site or app in the next 7 days.

How to use it: This metric helps you spot at-risk customers before you lose them. If a previously loyal customer's engagement drops and their churn probability rises, you can proactively try to win them back. Create an audience of "Likely to Churn" users and target them with a re-engagement campaign — perhaps a "we miss you" email with a special offer, or a survey asking for feedback on their experience. You can also exclude this audience from your regular conversion-focused ad campaigns to avoid wasting ad spend on users who have lost interest.

3. Predicted Revenue

What it is: This predicts the total revenue you can expect from all purchase conversions within the next 28 days from a user who has been active in the last 28 days.

How to use it: Predicted Revenue is your key to identifying potential high-value customers. It lets you look beyond a user’s past purchases and estimate their future worth. For example, a customer who made several small purchases might have a higher predicted revenue than a user who made one large purchase and never returned. You can create an audience of users with the highest predicted revenue and prioritize your ad spend to attract more customers like them, focusing your budget where it will have the greatest impact.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

The Catch: Meeting the Eligibility Requirements

Unfortunately, predictive metrics don't work "out of the box" for every GA4 property. Google’s machine learning models need a significant amount of high-quality data to make accurate predictions. Your property must meet specific data thresholds before these features become available.

To be eligible, over a recent 28-day period, you need to meet the following minimums:

• Triggering Users: You need at least 1,000 returning users who have triggered a specific event (like a purchase). This group provides the positive signals for the model.
• Non-Triggering Users: You also need at least 1,000 returning users who have not triggered that same event. This group provides the negative signals, which are just as important for training the AI.

Essentially, the model needs to see what both purchasers and non-purchasers look like to understand the key differences in their behavior. The model's quality depends on maintaining these thresholds consistently. If your user numbers drop below the minimum, the feature may become unavailable until the thresholds are met again. GA4 will notify you in the Audience builder when your property becomes eligible.

Finally, your property must be sending purchase and/or in_app_purchase events for purchase probability and predicted revenue models to work. You'll also need to have successfully implemented user consent requirements for analytics storage.

How to Use Predictive Metrics: Building Smarter Audiences

The most powerful way to leverage predictive metrics is by creating Predictive Audiences. These are dynamic lists of users that you can use for ad targeting in Google Ads. Here's a simple breakdown of how to create one:

1. Navigate to the Admin section of your GA4 property.
2. In the Property column, click on Audiences and then select New audience.
3. Under the Suggested audiences section, click on the Predictive tab. You'll see templates like "Likely 7-day purchasers" and "Likely 7-day churners."
4. Select the audience you want to create, for example, Likely 7-day purchasers.
5. You'll see a slider that lets you define the segment of users to include, based on their probability score. By default, it's set to include users from the 80th to 100th percentile — meaning the 20% of users most likely to purchase. You can adjust this slider to make your audience broader or more specific.
6. Set a membership duration (how long a user stays in the audience) and save it.

Once saved, you can link your GA4 property to your Google Ads account and start using this audience for targeted campaigns. Create a "high intent" retargeting campaign for your "Likely purchasers" audience, or a "win-back" campaign for your "Likely churners."

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Beyond Audiences: Analyzing Predictions in Explorations

You can also use predictive metrics directly in your analysis within GA4's reporting tools. Specifically, the "User lifetime" report in the Explorations section is a great place to put this data to work.

You can create a User Lifetime exploration to see metrics like "Predicted revenue" broken down by different user attributes, like their acquisition source, medium, or campaign. This allows you to identify which marketing channels are bringing in users with the highest future-value potential, helping you optimize your marketing mix for long-term growth, not just immediate conversions.

Final Thoughts

GA4's predictive metrics give you a powerful look into the future of your user behavior. By moving beyond historical reports, you can identify your most valuable customers, spot those at risk of leaving, and build incredibly focused audiences to make your marketing budget work harder and smarter.

While exploring these metrics within GA4 is a great step forward, we know that getting fast, clear answers from your data can still be a challenge. We built Graphed because we believe anyone should be able to get business insights without a degree in data science. You can connect your Google Analytics account in seconds and simply ask questions in plain English, like "Show me a chart of my predicted revenue by traffic source" or "Create a dashboard of my most engaged users this month." No navigating complex menus or waiting for data models to process — just simple conversations that lead to instant, actionable reports so you can get back to growing your business.