Unlike most other dimensions where Google Analytics classifies people and events for you, audiences let you customise how you slice and dice your website visitors. You get to put users into a range of categories. A user can belong to more than one category as well.

A couple of important things about audiences:

• It’s important when you define an audience! GA4 may be able to backfill the audience for up to 30 days but it’s not indefinite. If you want to look at audiences based on long-term historical data, you’re best to use the GA4 BigQuery data.
• Audiences can be used for reporting/analysis (in GA4) but can also be exported to other Google products (eg. Google Ads) for advertising/remarketing.
• In order to export audiences you’ll need to enable Google Signals.

Creating audiences

These are in the admin area. There are some pre-defined templates but the main place to build them is the custom audience builder. Here, you build a series of conditions which a user has to match either in sequence or not in sequence. A condition can be based on event names or parameters or almost any dimension in GA4.

If your property has enough traffic, you can also build predictive audiences, for example people GA4’s machine learning thinks are likely to purchase in the next 7 days.

You will also need to set how long someone stays in an audience once they match the conditions.

Example audiences

This can be almost anything but here are some ideas that we think would apply to a lot of readers. Note that standard properties have a maximum of 100 audiences so don’t go overboard.

• All visitors (mainly useful for general remarketing)
• People who visited the website but didn’t engage widely (eg. low engagement time, low pageviews, no conversions)
• People who engaged (eg. pageviews) but didn’t convert
• People who converted
• People who converted but not for a while
• Most important customers (eg. high average revenue in the last 90 days)
• People who started filling out a form (or the checkout) but didn’t complete it
• People who viewed a key page (eg. your contact page), video or downloaded a key PDF
• People from a geographic area you’re particularly interested in
• People who clicked on an ad from a marketing campaign you’re particularly interested in
• People who viewed a sales page and then viewed a contact form
• People who viewed a particular page or product category
• People who performed a site search
• People who logged into your website

Finally for analysis, it may be useful to define several complementary audiences to compare their size and conversion rate. For example, if there’s a key video and you want to see if it actually contributes to conversion, it might be useful to compare these audiences:

1. People who didn’t visit the page with the video
2. People who visited the page with the video and played it
3. People who visited the page with the video and did NOT play it

If for example the conversion rate of groups 2 and 3 are similar, this would suggest (at least as a preliminary finding) that the video is not actually convincing people to sign up.