Google Analytics 4: The Complete Setup Guide and FAQ Every Advertiser Needs

GA4 is not a minor update to the analytics platform most marketers have used for the last decade. It is a fundamentally different system — a different data model, different attribution logic, different audience architecture, and different reporting philosophy. The transition from Universal Analytics was not just a feature migration. It changed how measurement works at the foundation level.

That's why most advertisers who "set up GA4" by adding a tracking tag and calling it done are running campaigns on incomplete data. Their bid strategies are optimizing toward proxies. Their audiences are thin. Their conversion numbers don't match reality. And they don't know why.

This guide covers what GA4 actually changed, how to set it up properly for paid media use, and how to fix the problems that appear in almost every account we audit.

GA4 vs Universal Analytics — What Actually Changed

The most important change in GA4 is the shift from a session-based data model to an event-based data model. In Universal Analytics, every visit created a session, and interactions within that session (pageviews, events, goals) were measured relative to it. GA4 has no primary session concept — every interaction is an event, and sessions are just one type of event that GA4 derives from others.

This changes more than it seems. It means that what counts as a "conversion" in GA4 is an event, not a goal. What counts as a "user" is measured differently. What counts as a "session" uses a different timeout logic. And what counts as attributed revenue depends on a model that GA4 controls independently of the old last-click default UA used.

The Key Differences That Matter for Advertisers

• Data model: Event-based (GA4) vs. session/pageview-based (UA). Every GA4 interaction has an event name and up to 25 custom parameters.
• Attribution default: GA4 uses data-driven attribution by default. UA used last non-direct click. This alone causes significant differences in how credit is assigned across channels.
• Session counts: GA4 sessions run 20–35% lower than UA because GA4 aggressively filters bot traffic and spam that UA counted as real sessions.
• Cross-device tracking: GA4 is designed for cross-device and cross-platform measurement from the ground up. UA required separate web and app properties with manual stitching.
• Data retention: GA4 retains raw event data for a maximum of 14 months (2 months by default — extend this immediately in Admin). UA had no limit on historical data.
• Bounce rate replaced: GA4 uses Engagement Rate instead of Bounce Rate. An engaged session is one that lasts more than 10 seconds, has a conversion, or includes 2+ pageviews.

The practical consequence: you cannot compare GA4 numbers to UA numbers and draw meaningful conclusions. They are measuring similar things using different definitions. Treat GA4 as a new baseline, not a continuation of UA history.

Setting Up GA4 Correctly From Day One

Most GA4 setups we audit have three things in common: they added the base tag, they left Enhanced Measurement on its defaults, and they never configured custom events or conversion marking. The result is a property collecting a fraction of the data it could — and feeding incomplete signals to Google Ads bidding.

The Complete GA4 Setup Checklist

• Create the GA4 property and add your data stream (web). Use Google Tag Manager for tag deployment — direct tag implementation is harder to manage and debug.
• Extend data retention immediately: Admin → Data Settings → Data Retention → set to 14 months. The default is 2 months. Failing to do this is permanent data loss.
• Configure Enhanced Measurement deliberately. Auto-tracked events (scrolls, outbound clicks, file downloads, video engagement) sound useful but can inflate event counts and trigger false conversions. Review each one and disable what doesn't align with your business model.
• Define your conversion events explicitly. Mark only events that represent meaningful business actions as conversions: purchases, qualified form submissions, phone call initiations, booked appointments. Marking too many events as conversions confuses Smart Bidding.
• Set up custom dimensions and metrics for data your business tracks but GA4 doesn't capture by default — user type (logged in vs. guest), plan level, product category, lead source from your CRM.
• Enable Google Signals in Admin → Data Settings for cross-device reporting (requires accepting additional data usage terms).
• Configure Looker Studio or GA4 custom explorations for the reports your team actually uses — GA4's standard reports are not a replacement for UA's reporting structure.

The GA4 setup process we use with clients takes significantly longer than dropping a tag — typically 2–3 weeks to get from a blank property to a properly instrumented, conversion-ready account. The difference in data quality is measurable.

Key GA4 Events You Must Track for Paid Media

GA4 auto-collects a set of standard events for every property. But auto-collected events alone are not sufficient for paid media optimization. The events that matter most for Google Ads performance — qualified leads, revenue values, funnel stages — require deliberate configuration.

Auto-Collected Events (Available Without Configuration)

• session_start — fires when a new session begins
• first_visit — fires when a user visits for the first time
• page_view — fires on each page load (and virtual pageview)
• user_engagement — fires when the page has been in focus for 1+ second

Enhanced Measurement Events (Enabled in GA4 Settings)

• scroll — fires at 90% page scroll depth
• click (outbound) — fires when a user clicks a link leaving your domain
• view_search_results — fires on internal site search
• video_start / video_progress / video_complete — for embedded YouTube videos
• file_download — fires on PDF, doc, and other file download links

Recommended Custom Events for Paid Media

• generate_lead — fires on form submission with parameters: form_id, lead_source, estimated_value
• purchase — with transaction_id, value, currency, and item parameters for proper revenue tracking
• begin_checkout — critical for e-commerce funnel analysis and remarketing
• schedule_call / book_appointment — for service businesses where phone and calendar bookings are primary conversions
• view_item / add_to_cart / remove_from_cart — full e-commerce event set for funnel analysis
• login / sign_up — to distinguish new user acquisition from returning user behavior

The events that have the most direct impact on paid media performance are the ones with revenue values attached. When GA4 passes purchase events with accurate transaction values to Google Ads, Smart Bidding can optimize toward actual revenue — not just conversion volume. Without value data, Target ROAS bidding is operating blind.

GA4 + Google Ads Integration: The Essential Setup

Linking GA4 to Google Ads is not optional for advertisers who want their campaigns to perform at full capability. The link enables two things that have an outsized impact on performance: conversion import for bidding optimization, and audience sharing for remarketing and Smart Bidding signals.

Step-by-Step Linking Process

• In GA4: Admin → Property → Google Ads Links → Link → select your Google Ads account → confirm
• Enable "Enable Personalized Advertising" during linking — this allows GA4 audiences to be used in Google Ads
• In Google Ads: Tools & Settings → Measurement → Conversions → New conversion action → Import → Google Analytics 4 properties → select your conversion events

Conversion Import vs. Auto-Import

GA4 offers "auto-import" of conversions to Google Ads, but this is not the same as a deliberate conversion setup. Auto-import will bring over every event you've marked as a conversion in GA4 — including ones that are too high in the funnel for bidding optimization. The best practice is to manually import only the conversion events that represent genuine business value: purchases, qualified leads, booked appointments.

Enhanced Conversions

For any advertiser using Google Ads, Enhanced Conversions for Web is one of the highest-leverage setups available right now. It supplements cookie-based conversion tracking with hashed first-party data (email address, phone number) that Google matches to signed-in users — closing the gap caused by cookie blocking and iOS privacy restrictions. Enable it in Google Ads Conversion Settings and deploy the required tag via GTM.

The combination of properly imported GA4 conversions plus Enhanced Conversions gives Smart Bidding the signal quality needed to actually optimize toward revenue. Accounts with this setup consistently outperform those relying on standard cookie-based tracking alone.

Understanding GA4's New Attribution Models

Attribution is where GA4 creates the most confusion — and the most consequential misunderstandings — for advertisers reviewing campaign performance.

Universal Analytics defaulted to last non-direct click attribution. This meant the last marketing touchpoint before a conversion received 100% of the credit. GA4 defaults to data-driven attribution (DDA), which uses a machine learning model to distribute credit across all touchpoints in proportion to their actual contribution to conversions.

What This Means in Practice

If a user clicked a Google Search ad two weeks ago, then saw a Display ad last week, then converted via an email link today, last-click would give all credit to the email. Data-driven attribution would distribute credit across all three touchpoints based on how much each one actually influenced the conversion path.

The result: Google Search and Display campaigns typically receive more credit under DDA than under last-click. Email and direct channels receive less. This does not mean the old attribution was wrong and the new one is right — it means the picture is more complete, and budget decisions based on GA4's DDA will tend to favor upper-funnel channels more than last-click analysis did.

Attribution Models Available in GA4

• Data-driven (default) — ML-based credit distribution, requires minimum conversion volume to qualify
• Last click — 100% credit to the last touchpoint before conversion
• First click — 100% credit to the first touchpoint in the conversion path
• Linear — equal credit distributed across all touchpoints
• Time decay — more credit to touchpoints closer in time to conversion
• Position-based — 40% to first, 40% to last, 20% distributed across middle touchpoints

You can compare models in GA4 under Advertising → Attribution → Model Comparison. Before making major budget decisions based on GA4 data, understand which attribution model is active and what it implies for your specific channel mix.

GA4 Audiences for Remarketing and Smart Bidding

GA4 audiences are one of the most underused capabilities in most ad accounts we review. When configured properly, GA4 audiences passed to Google Ads provide Smart Bidding with behavioral signals that platform-only data cannot replicate — and they enable remarketing precision that goes well beyond "visited the website."

Proper GA4 event setup increases remarketing audience quality by up to 45%, because the audiences are built from actual behavioral signals (specific page visits, form interactions, product engagement) rather than just cookie-based site visit data.

High-Value GA4 Audiences for Paid Media

• Cart abandoners — users who triggered add_to_cart but not purchase within a 7-day window. The single highest-converting remarketing segment for e-commerce.
• High-engagement visitors — users with 3+ sessions or 5+ minutes on site. Strong signal for B2B and considered-purchase categories.
• Checkout starters who did not complete — begin_checkout without purchase. Different intent level from cart abandoners and warrants different creative.
• Past purchasers by product category — segmented by item_category parameter for cross-sell and upsell campaigns.
• Lead form visitors who did not convert — users who reached the contact or pricing page but did not submit. Strong intent signal for B2B.
• Predictive audiences — GA4 offers "likely purchasers" and "likely churners" as ML-generated segments, available once your property has sufficient data.

How to Create and Share GA4 Audiences

In GA4: Admin → Audiences → New Audience → build conditions using events, parameters, and session behavior → save. Once the GA4-to-Google-Ads link is active with personalized advertising enabled, new audiences appear automatically in Google Ads under Audience Manager within 24–48 hours.

Set your membership duration deliberately. For high-intent audiences (cart abandoners, form starters), 7–14 days is appropriate. For general site visitors used as observation segments, 30–90 days gives broader reach for Smart Bidding to learn from.

Also see our guide on PPC trends and first-party data strategy for how GA4 audiences fit into a broader paid media architecture.

Common GA4 Problems and How to Fix Them

Every account we audit has at least one of these problems. Most have three or four.

Problem: GA4 conversion numbers don't match Google Ads

Why it happens: GA4 and Google Ads use different attribution windows, different attribution models, and different session definitions. GA4 counts conversions on the day the conversion event fired. Google Ads counts them on the day the ad click happened. These differences alone cause reporting gaps of 15–30%.

Fix: Stop trying to reconcile the two numbers directly. Use GA4 for understanding user behavior and conversion paths. Use Google Ads for evaluating campaign-level efficiency. Use your CRM as the source of truth for actual revenue attribution.

Problem: GA4 showing far fewer sessions than Universal Analytics did

Why it happens: GA4 filters known bot and spam traffic that UA counted. GA4 also uses a different session timeout and session definition. A 20–35% reduction in session counts is normal and expected — it means GA4 is showing you more accurate human traffic data, not that tracking broke.

Fix: Establish a GA4 baseline and stop comparing to UA history. If you need historical data, export UA data to BigQuery or CSV before it expires. See our conversion tracking checklist for additional context on cross-platform measurement.

Problem: Remarketing audiences in Google Ads are too small

Why it happens: GA4 audiences require a minimum of 1,000 users (for Display) or 1,000 active visitors in the past 30 days (for Search) before they become eligible for targeting. Small sites may never reach this threshold for granular segments.

Fix: Build broader audience definitions for sites with lower traffic. Combine behavioral signals — for example, "visited any service page OR spent more than 2 minutes on site" — to reach the minimum threshold. For very small sites, customer match (uploading an email list) is a better starting point than behavioral audiences.

Problem: GA4 conversion events firing multiple times per session

Why it happens: Most commonly caused by GTM tags firing on every page instead of only on thank-you or confirmation pages, or by Enhanced Measurement auto-tracking an event that is also being tracked via a custom tag — causing duplicates.

Fix: Use GTM's Preview mode and GA4 DebugView together to trace exactly what triggers each conversion event and when. Add page path conditions to conversion tags so they only fire on confirmation URLs. Disable the Enhanced Measurement equivalent if a custom tag is covering the same event.

Problem: GA4 audiences not appearing in Google Ads

Why it happens: Either the GA4-to-Google-Ads link was created without enabling personalized advertising, or the audience was created after the link but before the audience collected enough users to be shared (minimum 100 users for Search, 1,000 for Display).

Fix: Verify the link in GA4 Admin → Google Ads Links → confirm "Personalized Advertising" is enabled. If it isn't, delete and re-create the link with it enabled. Then allow 24–48 hours for audiences to propagate. Check Google Ads → Tools → Audience Manager → Your Data Segments.

If you want comprehensive support with GA4 setup, event tracking, and Google Ads integration, our analytics team handles everything from initial property configuration through ongoing data quality monitoring. A correctly instrumented GA4 account is foundational — it is the infrastructure that makes every campaign decision more reliable.