Predictive audiences are an exciting new feature of Google Analytics 4. The automatically composed audience segments give you a shortcut to cheaper paid campaigns and a higher ROI (Return on Investment). That’s a double win. Let’s look if and how you can profit from this.
Get started with GA4 predictive audiences in no time with these 7 questions (and answers).
1. What are predictive audiences in GA4?
Predictive audiences are a novelty in GA4. Google uses your e-commerce or app data (visitors, conversions) and Artificial Intelligence to predict the purchase behavior of your users for the next 1 to 4 weeks. You can use and adapt 5 suggested predictive audiences and target these in your paid campaigns.
Don’t panic if you don't understand everything at this point. I will explain step-by-step what you need to know in the following questions.
But first I want to highlight one important aspect.
Predictive audiences don’t exist in GA Universals
As a digital marketer, you know FOMO (Fear Of Missing Out) can speed up the decision process. But there are many more reasons than this one cool feature to overcome objections to switch to GA4.
It may, however, take some time to get access to predictive audiences.
Let’s quickly find out if your Google Analytics 4 property is ready for the future of marketing.
2. Is your GA4 property eligible for predictive audiences?
To profit from the power of predictive audiences, your website or app needs to meet several requirements. You can quickly check whether your GA4 property is eligible by following these 5 steps.
1. Click on Configure in the menu on the left.
2. Click on Audiences.
3. Then click on New Audience.
4. Scroll down to Suggested Audiences.
5. Click on Predictive.
“Not eligible to use”
This means you will have to wait until you have enough users who made a purchase on your app or site.
Marketers and waiting. That’s a rare combo.
Instead of wasting your time away, you can spice up your skills with one of our free or paid resources for digital marketers.
Or you can generate traffic with Google Ads without predictive audiences to feel the difference later on.
“Ready to use” predictive audiences
This status means you can jump into the future and to question 7.
But I recommend you first understand the differences between the available predictive audiences.
3. What are suggested predictive audiences in GA4?
Google Analytics 4 has five different predictive audiences. Those are in fact built-in audience segment templates. You can use them as they are, or fine-tune the metrics to your specific needs.
The nuances between the predictive audience templates are subtle: likely, predicted, purchase, spend, churn, users, purchasers.
If you are familiar with the GA4 Lifetime Value Report, this will already make sense to you.
If you are confused, the cheat sheet below may be helpful. For now, you can ignore the campaign column. I will deal with it in the last question.
Google can identify and target two types of users and divides these further into audiences.
- Users who are likely to open their wallet soon
- Users who are likely to slip away soon
Nice, but what to do when your business needs more predictive audiences?
Fortunately, Google has kept data analysts in mind when building GA4.
4. Can you create custom predictive default audiences in GA4?
Next to the 5 suggested predictive audiences, you can create predictive audiences yourself in two ways:
- Edit one of the 5 suggested predictive audience templates
- Create an audience from scratch and add predictive metrics
GA4 offers a lot of metrics to define your predictive audience.
My recommendation: do not tamper with the suggested audiences, unless you are an experienced GA4 data analyst. Keep in mind that the more narrowed down your predictive audience, the more time it may take to become an eligible audience.
5. How does Google compose predictive audience segments?
Simply put, Google divides users into audience segments in 4 steps. Track. Collect. Learn. Predict.
- GA tracks behavior on the internet and your website with cookies.
- Your website data is stored in huge databases.
- AI (Artificial Intelligence) processes the data and learns more about how your users behave.
- Finally, your users are grouped into segments based on predictions about their future behavior.
The more data (visitors, conversions) your site feeds to Google’s hungry machine learning algorithm, the more accurate its predictions.
When you have too few conversions, the predictive model becomes nothing but a wild guess. Google (and you) wouldn’t like that. Hence the not eligible status.
Luckily for you, you can drive more traffic to your website or app with… paid advertising on Google.
What a delightful coincidence.
Or is it?
6. Are predictive audiences an upselling technique for GA4?
You can’t use predictive audiences in Google Ads if you don’t use GA4. If you don’t have enough conversions, you cannot use predictive audiences.
To get more conversions, you are likely (see what I did here?) to turn into a Google Ad user.
Smart move Google.
But let’s keep it real.
Predictive audiences are a win-win. It makes running paid search marketing a lot easier.
7. What can you do with GA4’s predictive audiences?
The purpose of predictive audiences is to get more from your paid marketing efforts. In the end, you save time and money by focusing only on users who are on the verge of slipping away, or are likely to pay.
Your GA4 predictive audiences are automatically imported in your linked Google Ads account. (That’s coincidentally a lot easier than taking 11 steps to connect GA4 with Search Console).
Google knows your users.
You know Google will show your ads to them.
And both of you are more certain than ever before that laser focus targeting can result in more revenue.
For the both of you.
How can you put this into practice?
Remarketing campaigns
Instead of showing ads to everybody who visited your website, you can show it to specific audiences who are more LIKELY to buy from you.
In your remarketing campaigns, you can target 3 suggested predictive audiences:
- Likely seven-day purchasers
- Likely first-time seven-day purchasers
- Predicted 28-day top spenders
Re-engagement campaigns
Unless you are a huge brand, users who know your app or website rarely think of you all the time. They explore other options (from your competitors), or are too occupied with spending time and money on other important aspects of life.
7 days is long enough to stick around in the memory of your users. When you show your ads to churners, you have a higher chance that they will remember you. And click on your ad.
Tip for paid campaigns with predictive audiences
Of course, it is not all moonlight and roses. Unless you do your homework properly, your paid campaigns will likely fail, despite the power of AI.
A good thing about predictive audiences is that it
- frees up time for creativity
- it makes you fully aware of where your target audience is residing in your sales funnel
Driving people who visited your homepage again to your homepage is not effective. Instead, you can lead them with adapted ads to adapted landing pages, with more (or fewer) testimonials, an adapted pricing offer, a special promotion, etc.
Key Takeaways
Predictive audiences in Google Analytics 4 are a huge step forward. From reporting what happened to predicting what will probably happen makes GA4 more powerful than Universal Analytics.
To use predictive analytics, your website needs a substantial amount of visitors and converters. Google’s machine learning can only work when it has enough data from your property.
Setting up predictive audiences is easy. You can start with the 5 predefined suggested audiences.
If you want to narrow them down, you can customize a predictive audience template. Or you can create a new audience and configure the predictive metrics.
If you use Google Ads, your predictive audiences are automatically imported from your linked GA4 property.
You can then reach these audiences effortlessly with adapted remarketing and re-engagement campaigns.
Final note:
This post is not promoted with Google ads, so Google could not predict you would be here. And yet, here you are. I sincerely hope the above information will help you set your data at work for your future campaigns.