GA4 audience builder: deep dive

The four building blocks

Block	What it does
Conditions	Define inclusion or exclusion rules based on events, parameters, demographics, devices, or user properties
Scope	Apply the condition across all events, within one session, or within one event
Sequences	Require conditions to happen in order, optionally within a time window
Exclusions	Remove users who match a separate condition

Scope: across users, sessions, or events

Scope is the most misunderstood concept in GA4 audience building. The same condition behaves very differently at different scopes:

• Across all sessions (default). The condition is true if the user ever did it. Good for "users who have ever viewed a product page."
• Within the same session. The condition must happen within one session. Good for "users who viewed a product and added to cart in the same session."
• Within the same event. The condition must be true within a single event. Good for parameter-based filtering like "page_view where page_title contains 'pricing'."

Sequences: the time-aware audience

Sequences let you require conditions to happen in order. The two sequence rules:

1. Order matters. Step 1 happens before Step 2 before Step 3.
2. Time constraint (optional). Step 2 must happen within X minutes/hours/days of Step 1.

Common sequence audiences: "Added to cart but didn't purchase within 24 hours" (sequence: add_to_cart THEN [no purchase] within 24h, scoped to all sessions). "Visited pricing page then visited demo page within 7 days" (sequence: pricing page_view → demo page_view within 7d).

Predictive audiences

GA4's predictive metrics — purchase probability, churn probability, predicted revenue — power predictive audiences. Requirements:

• Ecommerce data with purchase events.
• Minimum 1,000 returning users with the relevant event in the past 28 days.
• Minimum 1,000 users without the event for comparison.

The predictive audiences GA4 ships with:

• Likely 7-day purchasers. Users with high predicted probability of purchasing in the next 7 days.
• Likely 7-day churning users. Existing users with high predicted probability of churning.
• Predicted 28-day top spenders. Users predicted to be in the top 20% of revenue contributors.

Each can be activated as a Customer Match audience in Google Ads — bid higher on likely purchasers, lower on likely churners.

The seven audience recipes every account needs

1. All converters. Users with at least one conversion event. Suppress from prospecting campaigns; target for retention.
2. High-intent non-converters. Cart-adders without purchase, demo-pricing-page viewers without lead form, plan-comparing users without signup. The prime retargeting audience.
3. Engaged sessions (deep readers). Users with 3+ pageviews or 2+ minutes per session. Quality awareness audience.
4. Past customers by tier. Single-purchase, multi-purchase, high-LTV. Different lifecycle campaigns per tier.
5. Lapsed customers. Past purchasers who haven't returned in 90+ days. Win-back campaigns.
6. Brand searchers. Users who entered via branded organic or paid. High-intent audience for upsell or cross-sell.
7. Predictive likely purchasers. GA4's predictive audience. Bid amplification for paid campaigns.

Activation: where audiences flow

Destination	How
Google Ads	Native link between GA4 property and Google Ads account. Audiences flow automatically once linked.
Display & Video 360 (DV360)	Native link from GA4 property; audience flows to DV360 audience library.
Campaign Manager 360 (CM360)	Native link via Floodlight; audiences activate as Floodlight remarketing lists.
Meta Ads	Indirect — export the user list via BigQuery, hash, upload to Meta as Customer Match.
TikTok, LinkedIn, Reddit, etc.	Same indirect path — BigQuery export → hash → platform Customer Match upload.
CDP	BigQuery export to your CDP (Segment, mParticle, RudderStack) or via Reverse ETL (Hightouch, Census).

BigQuery export: the activation engine

GA4's BigQuery export (free for most accounts) is the bridge for activating audiences in platforms beyond Google's. The flow:

1. Define audiences in GA4. They populate based on the GA4 event stream.
2. Audiences sync to BigQuery as a daily user_pseudo_id table.
3. Join against your CRM in BigQuery to attach email/phone to each pseudo_id.
4. Hash the emails and phones using SHA-256.
5. Export to a Reverse ETL tool (Hightouch, Census) or directly to each platform's API.
6. Activation: Meta Custom Audience, TikTok Custom Audience, LinkedIn Matched Audience, Klaviyo segment, etc.

Common audience-building mistakes

Building audiences that won't populate. The audience needs at least one matching user; without that, downstream activation fails silently. Audit audience size in the Admin → Audiences view; anything under 100 users typically won't activate effectively in Google Ads.

Forgetting the 14-day lookback. GA4 audiences populate based on the last 14 days of data by default (can be extended to 30 with a setting). A condition based on "users who purchased in the past 90 days" won't work as a single audience — you need a custom dimension or a BigQuery-derived audience.

Not linking accounts. An audience defined in GA4 doesn't appear in Google Ads until the accounts are linked. Same for DV360. Verify the link before assuming an audience is missing.

Imprecise naming. 50 audiences named "Cart abandoners (v2 final)" and "Audience 1 - DON'T USE" are the legacy of poor naming discipline. Establish a convention (e.g., RGM-CartAbandon-7day-Excl-Customer) and stick to it.

How do I activate GA4 audiences in Meta?

Indirect path: export GA4 audiences to BigQuery, join with your CRM to attach email/phone, hash the identifiers, upload to Meta as a Custom Audience. Reverse ETL tools (Hightouch, Census) automate this flow.

What's the minimum audience size for Google Ads?

GA4 audiences need at least 100 active users (in the last 30 days) before they can be used in Google Ads. Display campaigns need at least 1,000 users for serving. Search Remarketing Lists for Search Ads (RLSA) need at least 1,000.

Can I use GA4 audiences in PMax?

Yes — GA4 audiences can be added as audience signals in Performance Max. They direct the algorithm's initial population but don't restrict targeting. Strongest signal types: customer lists, high-intent non-purchasers, and predictive likely purchasers.

How long does a GA4 audience take to populate?

24-48 hours for most audiences. Predictive audiences require model training, which can take 7-14 days when first enabled.

What's the difference between GA4 audiences and Google Ads audiences?

GA4 audiences are defined in GA4 and sync to Google Ads (when linked). Google Ads audiences include GA4 audiences plus Customer Match lists (CRM uploads), Similar Audiences (lookalikes), Affinity, In-Market, Custom Audiences (search-based), and built-in audience categories.

Do GA4 audiences work without third-party cookies?

Yes for first-party signal-driven audiences (events, parameters, conversions). The audience is built on first-party event data. Activation in Google Ads remains effective; activation in Meta and other platforms depends on the platform's own identity resolution.

Operating checklist

1. Define the business outcome the audience supports before building.
2. Confirm GA4 has the events and parameters the audience condition needs.
3. Build the audience in GA4 with clear naming conventions.
4. Link GA4 to Google Ads, Meta Ads, and any DSPs that will activate it.
5. Verify audience populates within 24-48 hours.
6. Set a refresh cadence and ownership for the audience.
7. Document the audience purpose, definition, and downstream usage.