Audiences in GA4: building segments for remarketing
Parts 9 and 10 built out real event tracking for both businesses. A GA4 Audience is a saved, real definition of people who match specific event and parameter conditions — usable both for GA4's own reporting and, critically, exportable directly into Google Ads and Meta Ads for actual ad targeting.
Building a real audience: Bright Leaf Coffee cart abandoners
Audience name: "Cart Abandoners — 14 Days"
Include condition: triggered add_to_cart
Exclude condition: triggered purchase
Membership duration: 14 daysThis is the exact same real audience concept the Meta Ads series built directly as a Custom Audience — someone who added a product to cart but never completed checkout — now defined once, in GA4, using the real add_to_cart and purchase events already tracked in part 9's full e-commerce sequence.
The real relationship to platform-specific audiences
Meta Ads Custom Audience: built directly from the Meta Pixel's own
website-visitor tracking (Meta Ads series, part 6)
Google Ads remarketing audience: built from Google's own conversion
tracking (Google Ads series, part 11)
GA4 Audience: built from GA4's own event data, exportable to link
directly into either platform's advertising accountThese aren't redundant — a GA4-defined audience, once the property is linked (part 14), can be exported directly into Google Ads for remarketing without needing entirely separate audience logic rebuilt natively inside Google Ads itself. This is a genuine efficiency: define an audience's real conditions once, in GA4, and reuse it across every linked advertising platform rather than maintaining the same logic redundantly in three different places.
A real audience for GreenDesk: engaged, not-yet-converted leads
Audience name: "Engaged Content Readers — Not Yet a Lead"
Include condition: view_item (viewed 3+ pieces of content) AND
session engagement time > 2 minutes
Exclude condition: triggered generate_leadThis mirrors part 10's funnel directly — someone genuinely engaging with GreenDesk's content (the Content Marketing series' own Awareness-stage guidance) but not yet converting to a lead is a real, specific MOFU audience worth a distinct retargeting message, rather than lumping every non-converting visitor into one generic "everyone who visited" bucket.
Predictive audiences: GA4's own machine-learning segments
Predictive audience: "Likely 7-day purchasers" — GA4's own model,
identifying people it predicts are likely to purchase within a week,
based on real behavioral patterns across the property's historical dataBeyond audiences built from explicit, manually-defined conditions, GA4 offers predictive audiences built from its own machine learning model — genuinely useful once a property has accumulated enough real conversion history for the prediction to be meaningful (a similar data-volume dependency to the Meta Ads series' own learning phase guidance), and a real option worth testing once that volume exists, rather than relying purely on explicit rule-based audiences forever.
Exporting an audience into linked ad platforms
GA4 Audience: "Cart Abandoners — 14 Days" → linked Google Ads account
→ available as a remarketing list for Search and Display campaignsOnce a GA4 property is linked to a Google Ads account (part 14), any GA4 audience becomes directly available as a remarketing list inside Google Ads, with no separate audience rebuilding required — a real, practical efficiency for keeping audience logic consistent across both platforms.
Building the same conceptual audience — cart abandoners, engaged non-converters — separately and slightly differently in GA4, Google Ads, and Meta Ads, rather than defining it once and exporting where possible. Beyond the redundant effort, slightly different conditions across platforms produce audiences that don't actually represent the same real group of people, making cross-platform performance comparisons unreliable.
Next: Exploration reports — funnels, path exploration, and segment overlap, GA4's more advanced, flexible analysis tools beyond the standard built-in reports.