Predictive Metrics in GA4: Purchase Probability and Churn Probability

What GA4 Predictive Metrics Are

GA4 includes machine learning-powered predictive metrics that estimate the probability of future user actions based on their past behaviour. These are not just standard analytics metrics — they are forward-looking predictions.

The three predictive metrics available in GA4:

• Purchase probability: the likelihood that a user who was active in the last 28 days will make a purchase in the next 7 days
• Churn probability: the likelihood that a user who was active in the last 7 days will NOT be active in the next 7 days
• Revenue prediction: the predicted revenue a user will generate in the next 28 days

Requirements for Predictive Metrics

GA4's predictive models require enough data to train on. Minimum requirements:

• At least 1,000 users who triggered the predictive condition (e.g. returned purchase) in the past 28 days
• At least 1,000 users who did NOT trigger it

For small businesses with fewer than 1,000 monthly purchasers, predictive metrics may not activate. You will see a "Not enough data" indicator in the Predictive section if thresholds are not met.

Where to Find Predictive Metrics

• Reports then Retention (shows predicted user churn)
• Audience Builder (use predictive metrics as audience conditions)
• BigQuery export (predictive fields included in daily export tables)

Using Predictive Audiences for Google Ads

The most practical use of predictive metrics is building audiences in GA4 that can be pushed to Google Ads for bidding:

Likely-to-purchase audience:

1. Admin then Audiences then New Audience
2. Predictive tab
3. Select "Likely 7-day purchasers"
4. Save and link to Google Ads

In Google Ads: use this audience as a bid modifier (+30-50%) in Smart Bidding campaigns. These users are predicted to purchase soon — paying more to win their click is rational.

Likely-to-churn audience:

1. Build a "Predicted churn users" audience
2. Target with re-engagement campaigns before they lapse
3. Use email remarketing or display ads with special retention offers

Accuracy Limitations

Predictive metrics are statistical models, not certainties. A user in the "high purchase probability" bucket has a higher-than-average chance of purchasing — but most will not. Think of them as probabilistic signals, not guarantees. The value is in the aggregate: bidding more on a cohort with 15% purchase probability vs 3% average is rational even though 85% will not purchase.

Summary

GA4's predictive metrics are machine learning estimates of future user behaviour. The most practical application is building predictive audiences for Google Ads bidding (likely-to-purchase) and re-engagement targeting (likely-to-churn). They require 1,000+ monthly events to activate and work best as bid modifiers rather than strict targeting criteria.

See our GA4 Setup & Audit service for predictive audience setup.

Want to use GA4 predictive audiences in your campaigns? Contact Adslytics.