In the Google Marketing Platform, Audiences are how we pass collections of users between tools – like sharing a Google Analytics audience with Google Ads, Google Display & Video 360, or Google Optimize. While there are many ways to accomplish the same objective, using simple audience definitions in Google Analytics can improve your flexibility and accuracy when remarketing to users through Google Ads.
By keeping each audience definition modular and relying on tool-specific features, you can avoid situations that waste your money and annoy users.
What is a Google Ads Audience?
Remarking audiences in Google Ads allow you to target specific users based on a set of criteria you get to define. These audiences can be created in Google Ads or they can be created in and imported from Google Analytics.
You can then use multiple audience definitions when targeting a remarketing audience in Google Ads. For example, you could create 2 separate audiences in GA. Then in Google Ads, you could target a remarketing audience that includes everyone from the first audience and excludes everyone from the second.
This post by Michael explains in detail how to set up audiences in Google Analytics and how to import them into Google Ads.
Tell Me If This Sounds Familiar
Perhaps you’ve been in this situation: 1) You’re shopping for something online. 2) You buy that thing. 3) You proceed to get bombarded by ads for that thing. A quick search shows this is not a unique problem.
How Does This Happen
Anecdotally, I think most of can recall to a time that advertising has failed – which seems particularly infuriating in digital platforms where we expect/hope there’s a greater form of the good kind of personalization. While there are many reasons why search and display advertising can fail, one particularly manageable problem is how people define the audiences they target.
There are more advanced retargeting methods for ecommerce websites, like dynamic remarketing, but let’s go through an example where we might need to remarket to someone who visited a valuable page on our site – either a service or product that we’re trying to promote.
Say you want to remarket a specific product to users who have added that product to their cart but did not purchase it. You could include all of those criteria in a single audience definition and use it to create your remarketing audience in Google Ads:
The problem arises when a user comes back and ultimately purchases the product – whether they return later that day or two weeks from now, they’re still going to be in that remarketing audience because, at one time, they abandoned that product in their cart. Meeting the criteria of the audience adds them to the audience, but purchasing the product does not remove them from the list.
BUT IT CAN – as long as you have your remarketing audience set up correctly.
Streamlining Your Audience Definitions
The solution to the audience issue can be simple: rework your audience definition into something that isn’t so specific.
You can and will have audience definitions that involve multiple criteria, but you need to think through the definition to make sure you aren’t trapping users in a remarketing loop.
To remedy the situation I’ve created above, define one audience that includes all users who added that product to their cart:
Then, create a separate audience for users who purchased that product:
When creating your remarketing audience in Google Ads, you can include all users from the add to cart audience and exclude all users that are in the transaction audience.
Voilà: a remarketing audience that automatically removes users who have purchased that product.
Every Coin Has Two Sides
There’s a second principle that we should all adopt, and this applies to almost every problem we try to solve. When we attempt to target an ad to users by creating an audience to target, we need to answer both questions: Who should see this Ad? as well as Who should not see this Ad? These questions can help guide the audience creation and setup inside of Google Ads.
This applies to other challenges as well – when we’re adding tagging to certain pages on our site or creating experiment targeting in Google Optimize, we have to answer similar questions or we’ll end overcounting our conversions or showing our experiments to too many people.
Complicated Audience Definitions are a Bad Idea, Cont.
If you aren’t convinced by the lone scenario above, we have a few more reasons complex audience definitions are a bad idea.
1. Your audiences don’t collect enough cookies to be useful.
If your audience applies to just 3 people it is way too specific. You want/need to find that balance between targeting the most appropriate groups while also collecting enough cookies.
Bigger audiences are better. Audiences with 1,000 cookies can be used anywhere. Less than that, and they can only be used in display.
2. You think you’re targeting one group of users when, really, you’re targeting this other group.
The example I’ve given above applies here. You were inadvertently keeping users who had purchased the product in that remarketing audience.
This can cause the data your collecting (or at the very least your interpretation of that data) to be incorrect. It can also waste money.
3. You end up with a million audience definitions.
If your audience definition is specific, chances are you’re going to end up with a lot of those very specific audiences.
Keep in mind: once an audience has been created, it can’t be deleted – it can only be closed.
Evaluate the data before you create the audience. Set up your audience as a segment in GA first to ensure you aren’t running into any of the issues above. You can either look at the segment preview or apply that segment to your GA reports and click around to see if the data the segment is pulling in makes sense.
“I think people also try to make audiences very specific because they forget that they are able to combine it with targeting native to the tools or other dynamic elements. Dynamic attributes instead of product-level audiences help to scale. Targeting in conjunction with keyword and topic targets helps to contextualize ads.”
Shoutout to my colleague Stephen Kapusta for contributing to this post!