Artificial intelligence (AI) excels at handling large datasets, making it a natural fit for analytics. It’s no surprise that Google Analytics 4 (GA4) has incorporated sophisticated AI capabilities. This article delves into two core AI features in Google Analytics: predictive metrics and predictive audiences. These tools transcend mere data analysis; they empower businesses with foresight and strategic decision-making.
Table of contents
- How does Google Analytics 4 use AI?
- What are predictive audiences in Google Analytics 4?
- What are predictive metrics in Google Analytics 4?
How does Google Analytics 4 use AI?
GA4 utilizes AI in two primary ways. Firstly, through predictive audiences, where AI algorithms segment users based on their probability of performing certain actions, such as purchasing or disengaging. This allows marketers to fine-tune strategies for specific audience segments.
Secondly, GA4 integrates predictive metrics and segments directly into reports. This goes beyond simply displaying predictions; it enables creating segments based on anticipated data. This offers a nuanced understanding of potential customer behaviors, giving insights into how different segments might act in the future.
Directly incorporating these predictions into reports provides businesses a holistic view of their audience. This enables anticipating market trends and user actions more effectively. This advanced GA4 functionality marks a significant advancement in data-driven decision-making for businesses.
Speaking about data-informed choices… discover the true effectiveness of your Google Ads campaigns with our complimentary Google Ads Grader.
What are predictive audiences in Google Analytics 4?
Let’s start by examining predictive audiences within GA4. This groundbreaking feature, when connected to platforms like Google Ads or Display & Video 360, can significantly improve media performance across the Google Marketing Platform. They are particularly valuable for similar audience targeting, retargeting, and audience suppression strategies. Predictive audiences leverage predictive metrics, generated from machine learning models analyzing user behavior to forecast future activities.
Similar audience targeting
GA4’s predictive audiences enhance Google Ads’ similar audiences by identifying the top percentage of users most likely to convert. Marketers can target new users resembling this valuable group, optimizing reach more efficiently than traditional approaches like budget hikes or broader targeting.
Retargeting audiences
Predictive audiences refine retargeting by identifying users close to converting based on interactions like product views or cart activity. Targeting users likely to convert soon, or predicted top spenders, empowers marketers to personalize their approach for increased conversions and revenue. Conversely, these audiences can pinpoint users at risk of churn. Engaging them with tailored messages or offers can improve customer retention, minimizing churn, and fostering loyalty.
Audience suppression
Predictive audiences can inform strategies to avoid showing ads to users highly likely to convert independently, such as the top 10% expected to convert soon. This optimizes media spending by focusing on users needing additional encouragement to complete a purchase.
Optimize your Google Ads account for maximum results! Download Now » The Last Guide to Google Ads Account Structure You’ll Ever Need
What are predictive metrics in Google Analytics 4?
GA4 provides three predictive metrics:
- Purchase probability: Measures the likelihood of a user making a purchase within the next week based on their activity over the past month.
- Churn probability: Calculates the chance of a user not returning to the app or website in the following week.
- Predicted revenue: Forecasts potential revenue from a user’s purchases within the next 28 days.
Are you eligible to use predictive metrics in GA4?
Utilizing GA4’s predictive metrics requires fulfilling certain conditions for the machine-learning algorithms to learn effectively. Here’s what you need to set up:
- Event configuration: Ensure the ‘purchase’ event is set up correctly in your GA4 property settings. This is crucial for the algorithms to identify and learn from user purchases.
- Data volume: A significant dataset is needed for AI training. You need at least 1,000 positive (purchasers) and 1,000 negative (non-purchasers or churned users) instances. This ensures enough data for the AI to learn patterns.
- Model quality: Consistent, high-quality traffic generating purchase events over a minimum period, typically 28 days, is crucial. This allows the AI to recognize patterns for accurate predictions. GA4 will calculate daily predictive metrics for users meeting these conditions. If not met, or if user counts fall below the threshold, GA4 may stop updating these metrics, making them unavailable. To track your predictive metrics’ status, go to the ‘Audience builder’ section in GA4 and navigate to ‘Suggested audiences’. This offers insights into your predictions’ performance and reliability.
Predicted channel value report
Beyond audiences, integrating AI into reports can significantly improve analysis accuracy and insights, particularly when creating a predicted channel value report. Let’s explore how to do this: I’m using the GA4 demo account to create these reports, and I recommend you explore it too. Here’s a link to the demo account, providing real business data to experiment with Google Analytics features. I also demonstrate this in this YouTube video.
In your GA4 property, click ‘Explore’ on the right-hand menu.
Select the ‘Free form’ exploration template.
Clear the report template to start fresh. It should resemble this:
Under ‘Segments’, click the plus sign and choose ‘User segment’.
In ‘Add new condition’, scroll down to ‘Predictive’. Clicking this lets you segment data by predictive metrics. Choose ‘Predicted Revenue’, ensure ‘Most likely top spenders’ is selected under ‘Filter’, and click ‘Apply’.
Name your segment ‘Top Spenders (28 Day)’ and click ‘Save’ and ‘Apply’ (top right corner).
Add ‘First user source’ and ‘First user medium’ under ‘Dimensions’, and ‘Active users’ under ‘Metrics’.
Place ‘First user source’ and ‘First user medium’ in rows, and ‘Active users’ in values.
Change the cell type to a heat map. Your report should look like this:
This report provides insights into the Google Merchandise Store’s website analytics, predicting potential high-value customers for the coming month. The prediction is based on user acquisition channels, like direct traffic, referrals, or email campaigns. Such reports allow businesses to identify and focus on the most effective strategies for attracting valuable customers.
You can also include another segment for churn probability, resulting in something like this:
Adding this segment provides insights into which groups and channels have a higher risk of not returning to the Google Merchandise Store within a week.
These categories might reflect lower engagement or patterns previously associated with a lower likelihood of repeat visits or purchases. It also indicates that a subset of active users within the top spenders might not revisit the site.
Understanding these trends allows the store to create tailored retention strategies for these specific segments, minimizing churn and promoting continued engagement.
Use Google Analytics 4’s AI features
In conclusion, GA4’s powerful AI-driven features, like predictive metrics and audiences, are particularly valuable in a future without cookies. They allow marketers to effectively target audiences and anticipate user behaviors. While predictive audiences enhance advertising, using predictive metrics like purchase and churn probability effectively necessitates specific prerequisites. Moreover, integrating AI into reports, as demonstrated by creating predicted channel value reports, enhances data analysis precision and highlights impactful user acquisition strategies and potential churn risks. Utilizing AI in GA4 is crucial for remaining competitive and making data-driven decisions in the evolving digital world.