Thanks to Google Analytics 4, today’s PPC marketers have a powerful tool at their disposal: Predictive Audiences.
Using machine learning algorithms and predictive analytics, we can now create custom audiences built upon valuable user behavior data.
Long story short, it’s a boon for most digital marketers — as long as they’re using the feature correctly.
Today, we’ll demystify these predictive models, sharing some of our PPC team’s findings and recommendations since the launch of the automation feature a few years ago.
Keep reading for our take on using Predictive Audiences properly, or, if you’d rather get an expert’s help in making it happen for your brand, reach out to our team anytime to schedule your free consultation.
What are Predictive Audiences?
Predictive Audiences are audience subsets created in Google Analytics 4 that can also be used in Google Ads campaigns. These audiences are based on current website visitors that meet at least one condition based on a predictive metric — the possibility that a customer will complete a certain action, such as a purchase, add-to-cart, etc.
Google offers five suggested Predictive Audience templates that you can use:
- Likely 7-day churning purchasers: Purchasing users who are not likely to visit your website in the next 7 days.
- Likely 7-day churning users: Users who are not likely to visit your website in the next 7 days.
- Likely 7-day purchasers: Users who are likely to purchase in the next 7 days.
- Likely first-time, 7-day purchasers: Users who are likely to make their first purchase in the next 7 days.
- Predicted 28-day top spenders: Users who are predicted to generate the most revenue in the next 28 days.
Building Google Ads Audiences in GA4: Predictive Audiences & Other Options
You can also create custom audiences based on the predictive metrics, dimensions, and events of your choosing.
To use the Predictive Audience Builder, in your GA4 property, navigate to Admin > Audiences > New Audience > Predictive.
Prerequisites to Using GA4’s Predictive Audiences
Unfortunately, the power of Predictive Audiences isn’t available to every marketing team quite yet.
Like the rest of Google’s automation offerings, Predictive Audiences require a substantial amount of data to be effective, specifically:
- At least 1,000 positive and 1,000 negative samples: Typically, this means at least 1,000 purchases and 1,000 non-purchases (like add-to-carts). For those focused on lead generation, a positive sample could be a completed form, while a negative sample could be a “form begin” event.
- At least 1,000 signals sustained over a 28-day period: All of these data points must be collected within one month.
Of course, to actually capture this data, you’ll also need to have your purchase event configured in your analytics account. This is a crucial event for any eCommerce GA4 account and does not automatically carry over from Universal Analytics; if you’re missing it, you can use our GA4 eCommerce tracking toolkit to get yourself caught up to speed.
If your business isn’t eligible for Predictive Audiences, you do have other options, which we’ll discuss later on in this guide.
3 Best Practices for Using Predictive Audiences
Despite Google’s promises, its automation features are not foolproof.
As you test Predictive Audiences on your Google Ads campaigns, employ plenty of manual oversight from an experienced PPC marketer. Otherwise, you risk Google quickly running away with all of your ad spend — with little to show for it.
As you embark on your audience testing, keep these three best practices in mind:
1. Establish a baseline for comparison.
Before you start testing Predictive Audiences, we highly recommend creating a baseline audience segment against which you’ll compare all others. For most businesses, this will be an “all users” audience group.
By creating this baseline, you’ll better understand how your Predictive Audiences perform compared to your overall audience base. (Never forget the power of the control group in your scientific experiments!)
In the example below, you’ll see a comparison of one client’s Predictive Audiences (Likely 7-Day Purchasers, Likely First-Time 7-Day Purchasers) and manually created audiences (All Visitors, All Users).
Using the All Users audience as a baseline, we can see that the Likely 7-Day Purchasers audience is performing much more efficiently, delivering a 700% higher conversion value/cost and a 254% higher conversion rate.
On the other hand, the Likely First-Time 7-Day Purchasers is underperforming in comparison — with a 55% lower conversion value/cost and a 79% lower conversion rate.
We can use this data to better allocate ad spend across our predictive and regular audiences in Google Ads, in turn improving our overall ROAS. We could even consider pausing the Likely First-Time 7-Day Purchasers audience because it’s performing so poorly compared to our baseline audiences.
Bottom line: If you’re going to experiment with your Predictive Audiences (and we recommend you do!), make sure you’ve got a baseline audience running at the same time.
2. Strategically test based on your business goals.
The client example above illustrates the importance of testing as many Predictive Audiences as possible — but don’t do so willy-nilly.
For example, if you’ve recently run a promotional campaign aimed at bringing in new purchasers, consider pausing that First-Time 7-Day Purchaser audience until you have a better idea of the long-term value of those audiences. If your first-time purchasers rarely return after their first conversion, the cost required to convert them in the first place may not be worth the return they deliver.
In general, we recommend letting audiences run for an additional three to four weeks if their initial performance is complicated by external factors like a sale or holiday. Otherwise, you may preemptively make an audience adjustment based on inaccurate information — a choice that will haunt you when the audience fails to deliver later on.
Keep in mind that audience lists are only one factor in the success of your Google Ads campaigns. Make sure you’re also strategically testing your ad copy and ad creative to see what engages and converts those target audiences.
After all, you can have the best audiences in the world — but if your copy and creative fail to grab their attention, you’ll never motivate them to convert.
Need help creating engaging ad creative for your display or Performance Max campaigns? Reach out to see how Inflow’s High-Performance Ad Creative can help.
3. Keep the data flowing.
When it comes to Google’s automation features, there’s one rule that applies to all: The more information you can feed the system, the better it will work.
Whether we’re talking Performance Max or Predictive Audiences, you need to keep your stream of audience data as large as possible. More data allows the system to learn faster, which means Google’s automation will optimize more efficiently and deliver better results for your business.
We recommend frequently checking the health of your Predictive Audiences in GA4 to confirm their eligibility over time. If your website events and traffic slow down, any existing audiences may become obsolete and no longer be effective in your Google Ads marketing campaigns.
Don’t Have Access to Predictive Audiences? Use These Instead
The eligibility threshold for using Predictive Audiences is fairly high and can be difficult for even established brands to meet.
And, while Predictive Audiences do unlock valuable insights for advertisers, you can still successfully use Google Ads without them.
Here are the traditional audience segments we recommend for those of our clients without access to Predictive Audiences:
eCommerce
As mentioned above, we highly recommend creating a “baseline” audience of all users for your Google Ads campaign targeting. This gives you an idea of how your audience performs at a high level — a comparison point for your conversion rates, session durations, purchases, average order value (AOV), etc.
eCommerce advertisers should also consider creating audiences out of:
- All-time purchasers
- Frequent purchasers (i.e. multiple times a month or year)
- High-AOV purchasers
While these audiences may be small to start, they will gradually grow over time. Once they’re large enough, you can start creating lookalike and similar audiences based on these demographics — putting your campaigns on the fast track to higher purchase rates and better ROAS.
We also recommend investing in audience segments at the top of the funnel. You can build audiences based on website event data (add-to-carts, newsletter signups, etc.), allowing for later remarketing that pushes them through the funnel to ultimately become a customer.
Lead Generation
If your business operates in the lead-generation space, you probably have one goal: to get an audience’s contact information through a demo or quote signup. This limits the number of goals you can optimize for in your audience segments and can make a full-funnel targeting strategy more difficult.
However, if your GA4 configuration is robust enough to identify when a contact form has been started but not completed, we recommend creating a separate audience for those users. Similarly, you can create an audience of users who spend two or three times your sitewide average session duration; this is likely to be an audience heavily researching your brand and product.
The same strategy applies to creating an audience with multiple page views per session.
Here’s a basic list of audiences to add to your lead-gen-focused campaigns:
- All users
- Contact form start
- Contact form completion
- Long session duration
- Multiple page views
Build Your Google Ads Audience Strategy With Inflow’s Help
As Google’s latest development in its campaign toward full automation, Predictive Audiences can be a great addition to any PPC marketer’s toolbelt.
However, with great power comes great responsibility. If you want to start testing these audience segments on your Google Ads campaigns, you’ll need to do so smartly.
The best practices in this guide are a great starting point, but if you want to expedite your success with these audiences, consider working with a paid search ads expert like the ones at Inflow. That way, you can protect your valuable ad spend and trust that your PPC marketing efforts are in good hands, regardless of whatever updates Google throws your way.
Find out what our paid search marketers recommend for your campaigns by requesting a free Google Ads proposal below. After reviewing your account configuration and audience targeting, we’ll develop a data-driven strategy to improve your ROAS and scale your revenue growth in accordance with your business goals.
Get started by contacting us now:
0 Comments