GA4 Setup Guide 2025: Custom Events, Audiences, and Attribution Explained

Most businesses that migrated to GA4 are only using 20% of its capabilities. They have set up basic pageview tracking but missed the custom events, audiences, attribution models, and configuration settings that make GA4 genuinely powerful for marketing decisions. After auditing GA4 setups for 150+ businesses, here are the gaps we find — and fix — every single time.

Critical Configuration Settings Most Businesses Miss

Before worrying about custom events or audiences, these configuration settings must be correct or all your data will be incomplete or misleading:

1. Data Retention: Admin > Data Settings > Data Retention. Change from 2 months to 14 months immediately. Do this today — you cannot recover data that has already been deleted.
2. Google Signals: Admin > Data Collection > Google Signals. Enable this to track users across devices, unlock demographic and interest reports, and enable remarketing audiences based on GA4 data.
3. Link Google Search Console: Admin > Product Links > Search Console. This populates the organic search queries report in GA4 — essential for understanding which search terms drive visits and conversions.
4. Link Google Ads: Admin > Product Links > Google Ads. Without this, GA4 cannot attribute conversions to your paid campaigns or build remarketing audiences for Google Ads.
5. Internal Traffic Filter: Admin > Data Streams > Configure Tag Settings > Define Internal Traffic. Filter your own team's traffic to prevent skewing engagement metrics.
6. Unwanted Referral Exclusions: Admin > Data Streams > Configure Tag Settings > List Unwanted Referrals. Add your payment processors (Razorpay, PayU, PayPal) to prevent them from appearing as traffic sources after checkout.

Setting Up Custom Events Correctly

GA4 auto-collects basic events (page_view, scroll, click, session_start, first_visit) but the events that actually matter to your business require custom setup via Google Tag Manager. The most important custom events for most businesses:

• contact_form_submit — someone submitted your contact form (mark as conversion)
• phone_call_click — someone clicked your phone number (especially important on mobile)
• whatsapp_click — someone clicked your WhatsApp button
• video_play — someone started watching a video on your site
• file_download — someone downloaded a brochure, guide, or case study
• add_to_cart and purchase — for e-commerce (GA4's enhanced e-commerce schema)
• booking_complete — for service businesses using calendar booking tools

Naming Convention Best Practices

Always use snake_case for event names (contact_form_submit, not ContactFormSubmit or contact-form-submit). GA4 event names are case-sensitive — inconsistent naming creates duplicate events in your reports. Use Google Tag Manager for all custom event implementation — never hardcode tracking directly into your website code, as this requires developer involvement for every future modification.

Marking Conversions Correctly

In GA4, every goal must be explicitly marked as a conversion — nothing is tracked as a conversion automatically unless you designate it. In Admin > Events, find each event you want to track as a conversion and toggle the 'Mark as conversion' switch. In the Reports section under Conversions, you will now see conversion counts, rates, and attribution.

Be selective about what you mark as a conversion. If you mark every click and scroll as a conversion, your conversion rate metrics become meaningless. Typically you should have 3–8 primary conversion events that represent genuine business value: form submissions, phone calls, purchases, bookings, and high-value downloads.

Building High-Value Audiences in GA4

GA4's audience builder is significantly more powerful than Universal Analytics' goal-based audiences. You can now build audiences based on sequences of events, time spent between events, predicted likelihood of conversion, and dozens of behavioural criteria. The most valuable audiences to build:

• High-Intent Visitors: Users who visited your pricing or services page AND did NOT complete a conversion — your hottest remarketing audience
• Engaged Non-Converters: Users who spent >60 seconds on site AND scrolled >70% of a key page AND did not convert
• Converters (for Lookalikes): Users who completed a conversion event — use for Google Ads Similar Segments
• Return Visitors: Users who visited the site 3+ times in 30 days without converting — high-intent, needs a different message
• Video Engagers: Users who watched >75% of a product video — demonstrated high interest
• Geographic Segments: Visitors from specific high-value cities — essential for local service businesses

Attribution Models in GA4

Which Attribution Model Should You Use?

GA4's default is Data-Driven Attribution (DDA), which uses machine learning to assign credit across touchpoints based on actual conversion data patterns. For most businesses with sufficient data (50+ conversions per month per channel), DDA is the most accurate model available. For smaller businesses or newer accounts, DDA lacks sufficient data to be reliable — use Last Click (Google paid) or Cross-channel Last Click instead.

Setting the Correct Conversion Window

In Admin > Attribution Settings, set your conversion window to match your sales cycle. Most B2B services businesses should use 30–90 days — a decision-maker who first visits your website often takes weeks to convert. Most e-commerce businesses: 7–30 days. Using the wrong window dramatically misattributes credit to recent touchpoints while ignoring the earlier touchpoints that initiated the buyer journey.

The 5 GA4 Reports You Should Review Weekly

1. Acquisition > Traffic Acquisition: Which channels drive the most engaged sessions (sessions with events, low bounce)? Use this to guide budget allocation.
2. Engagement > Pages and Screens: Which pages generate the most conversions? Which have high engagement time but low conversion? These pages need CTA optimisation.
3. Monetisation > Conversions (or Reports > Engagement > Events): Which conversion events are completing? At what rate per session?
4. Retention > User Retention: What percentage of new users return within 7 days? 30 days? Low retention signals an onboarding or product problem.
5. Advertising > Attribution > Model Comparison: Compare Last Click vs. DDA side by side to understand which channels are under or over-credited in your current reporting.

GA4 vs. Universal Analytics — Key Conceptual Differences

Never compare GA4 metrics directly to Universal Analytics metrics — they are measured differently. Sessions in GA4 are calculated differently from UA sessions (timeout thresholds differ; a single user can generate multiple sessions in GA4 for scenarios that generated one in UA). Bounce Rate does not exist in GA4 — it is replaced by Engagement Rate (percentage of sessions that lasted >10 seconds or had 2+ page views or a conversion event).

• GA4 Sessions typically count 10–30% higher than equivalent UA sessions — not a tracking error
• GA4 Users count differently — GA4 identifies users by Client ID and User ID together, reducing duplicate counting
• There is no Goal flow in GA4 — use Funnel Exploration in the Explore section instead
• Standard reports in GA4 are sampling-free — but some Explore reports do apply sampling at scale

Frequently Asked Questions