GA4 Setup & Configuration

GA4 setup is the process of configuring Google Analytics 4 to collect event-based data from your website or app, replacing the older Universal Analytics (UA) system that Google retired in July 2024. If you’re still running on UA or haven’t configured GA4 properly, you’re flying blind.

“Most brands we audit have GA4 installed but barely configured. They’re collecting pageviews and nothing else. That’s like buying a Formula 1 car and driving it in first gear,” says Hardik Shah, Founder of ScaleGrowth.Digital.

This page covers everything we do when we set up GA4 for our clients: from migration planning to event architecture, conversion tracking, audience building, and BigQuery integration. We’ve configured GA4 for brands across BFSI, ecommerce, healthcare, and franchise operations since 2020, and the gap between a default install and a properly engineered setup is massive.

Why Did Google Replace Universal Analytics with GA4?

Universal Analytics was built for a web-only world where sessions and pageviews were the primary measurement unit. GA4 uses an event-based data model because user behaviour has changed. People move between websites, apps, YouTube, and email before converting. Session-based tracking couldn’t handle that.

Google officially sunset UA on July 1, 2024. If you created a new property after that date, GA4 is your only option. But here’s the problem: GA4 isn’t just “new UA.” The data model, the interface, the reporting structure, and the attribution logic are all different. Brands that treated migration as a checkbox exercise lost months of usable data.

We’ve seen companies where the GA4 property existed for over a year but had zero custom events configured. Default enhanced measurement was collecting scroll depth and outbound clicks, but the metrics that actually matter for their business were invisible.

What Does a Proper GA4 Setup Actually Include?

A complete GA4 implementation covers seven layers. Most agencies stop at layer two.

Layer 1: Property and data stream configuration. This includes setting up the GA4 property, connecting web and app data streams, configuring data retention (set it to 14 months, not the default 2 months), enabling Google Signals for cross-device reporting, and linking to Google Ads and Search Console.

Layer 2: Enhanced measurement. GA4 automatically tracks page views, scrolls, outbound clicks, site search, video engagement, and file downloads. But the defaults need tuning. For example, the scroll event only fires at 90% depth. For most content sites, you want 25%, 50%, 75%, and 90% thresholds to understand where readers drop off.

Layer 3: Custom event architecture. This is where real measurement begins. We map every meaningful user action to a custom event. For an ecommerce site, that’s add_to_cart, begin_checkout, purchase, and view_item_list. For a lead generation business, it’s form_submit, phone_click, chat_initiated, and pdf_download. Each event gets structured parameters that tell you the what, where, and context.

Layer 4: Conversion configuration. In GA4, any event can be marked as a conversion (Google now calls these “key events”). We identify the 5-8 actions that represent genuine business value and configure them as key events. Then we assign monetary values where possible, so you can calculate actual ROI in your reports.

Layer 5: Audience building. GA4’s audience builder is significantly more powerful than UA’s segments. We create audiences based on event sequences (users who viewed a product page AND added to cart but didn’t purchase), predictive audiences (users likely to purchase in the next 7 days), and custom dimensions. These audiences sync directly to Google Ads for remarketing.

Layer 6: Data layer implementation. For accurate ecommerce and form tracking, we implement a structured data layer using Google Tag Manager. The data layer pushes event data in a consistent format that GA4 can consume without relying on DOM scraping or CSS selectors that break when your dev team updates the site.

Layer 7: BigQuery integration. GA4’s free BigQuery export is one of its biggest advantages over UA. We configure the daily and streaming exports so you have raw, unsampled data in a SQL-queryable warehouse. This is critical for brands processing more than 500K sessions per month, where GA4’s interface starts sampling your reports.

How Is GA4’s Event-Based Tracking Different from UA’s Session Model?

In Universal Analytics, the fundamental unit was a session. A user arrived, did things during that session, and left. Every interaction was tied to that session container. Pageviews, events, transactions, and goals all lived inside sessions.

GA4 treats every interaction as an independent event. A pageview is an event (page_view). A purchase is an event (purchase). A scroll is an event (scroll). There’s no session container holding them together. GA4 reconstructs sessions after the fact using a 30-minute inactivity timeout, but the raw data is event-level.

Why does this matter practically? Three reasons.

First, cross-platform tracking works. When a user browses your app on their phone and then buys on your website from a laptop, GA4 can stitch those events together using User-ID or Google Signals. UA couldn’t do this natively.

Second, you can track things that aren’t pageviews without workarounds. In UA, tracking a button click required setting up a custom event with Category, Action, and Label. In GA4, you just fire an event with whatever parameters you need. The flexibility is dramatically better.

Third, the reporting changes. UA had pre-built reports for sessions, bounce rate, and pages-per-session. GA4 replaced bounce rate with engagement rate (sessions lasting longer than 10 seconds, or that had a conversion event, or that had 2+ pageviews). This is actually a better metric, but it means your numbers won’t match your old UA benchmarks. Don’t try to compare them.

What Is a Data Layer and Why Does GA4 Need One?

A data layer is a JavaScript object that sits on your web page and holds structured data about the page, the user, and the actions they take. Google Tag Manager reads from this data layer to fire tags and send data to GA4.

Without a data layer, your tracking relies on scraping information from the page itself. Your GTM triggers fire based on CSS classes, button IDs, or URL patterns. This is fragile. The moment a developer changes a class name or restructures a page, your tracking breaks and you don’t know until someone notices the data gap weeks later.

With a data layer, your development team pushes structured data objects when events happen. A purchase event pushes the transaction ID, revenue, tax, shipping, and item-level details into the data layer. GTM reads that object and sends it to GA4. The tracking is decoupled from the front-end code.

We implement data layers following Google’s recommended ecommerce schema. For non-ecommerce sites, we design custom schemas that capture the business-specific data points your team needs.

How Do You Set Up Conversions (Key Events) in GA4?

GA4 renamed conversions to “key events” in early 2024 to reduce confusion with Google Ads conversions. The setup process is straightforward but requires strategic thinking about what actually matters.

Step 1: List every action on your site that represents business value. Not just the final conversion (purchase, form submission), but the micro-conversions that indicate intent (added to cart, started checkout, viewed pricing page, downloaded a case study).

Step 2: Ensure each action fires a GA4 event. Some will be covered by enhanced measurement. Others need custom events through GTM.

Step 3: Mark the events as key events in the GA4 admin panel. You can do this under Admin > Events > toggle the “Mark as key event” switch.

Step 4: Assign monetary values. For ecommerce, the purchase event carries the transaction value automatically. For lead gen, assign estimated values based on your conversion rates. If 1 in 20 form submissions becomes a client worth Rs 5 lakhs, each form submission is worth Rs 25,000 to your reporting.

Step 5: Verify with the Real-Time report and DebugView. Submit a test conversion and confirm it appears in both places with the correct parameters.

We typically configure 5-8 key events per property. More than 10 dilutes the signal. If everything is a conversion, nothing is.

What About GA4 Audiences and How Do They Help with Remarketing?

GA4 audiences are dynamic user segments that update in real time as users meet or stop meeting the criteria you define. They’re far more capable than UA segments.

You can build audiences based on events and event parameters (users who triggered add_to_cart with item_category = “enterprise”), user properties (users whose lifetime value exceeds Rs 50,000), time-based conditions (users who visited in the last 7 days but not in the last 2 days), and sequences (users who viewed a product, then added to cart, then abandoned).

The real power comes from the Google Ads integration. Every audience you create in GA4 can be pushed to Google Ads as a remarketing list. Want to run ads specifically to users who started your checkout flow but didn’t complete it within 48 hours? Build the audience in GA4, and it’s available in Google Ads within 24 hours.

GA4 also offers predictive audiences. Based on machine learning models trained on your data, GA4 can identify users likely to purchase in the next 7 days or users likely to churn. These predictive audiences require a minimum of 1,000 positive and 1,000 negative examples over 28 days, so they’re only available for sites with sufficient traffic volume.

Why Should You Connect GA4 to BigQuery?

The BigQuery export is, in our opinion, the single most underused feature in GA4. Every GA4 property, including free ones, can export raw event data to Google BigQuery at no cost (BigQuery’s free tier covers up to 1TB of queries per month and 10GB of storage).

Three reasons this matters.

First, GA4’s interface samples data on properties with more than 500K events. If your reports show a yellow shield icon, you’re looking at estimated numbers, not actuals. BigQuery gives you the complete, unsampled dataset.

Second, you can join GA4 data with other business data. Want to see which marketing channels drive the highest lifetime value? Join your BigQuery GA4 export with your CRM data. Want to correlate website behaviour with offline sales? BigQuery makes that possible.

Third, you own the data. GA4 retains user-level data for a maximum of 14 months. BigQuery stores it indefinitely (within your Google Cloud project). For any longitudinal analysis beyond 14 months, BigQuery is your only option.

We configure both the daily export (batch, processes overnight) and the streaming export (near real-time, available within minutes) for clients who need up-to-the-hour reporting.

How Long Does a GA4 Setup Take?

For a standard business website with 10-30 pages and a contact form, we complete the full seven-layer setup in 5-7 working days. That includes the data layer implementation, which requires coordination with your development team.

For ecommerce sites with complex product catalogues, the timeline extends to 10-15 working days, primarily because of the ecommerce data layer work and the testing required across product pages, cart, checkout, and post-purchase flows.

For multi-property setups (separate properties for different countries or business units with a roll-up property), add another 3-5 days for the cross-property configuration and data governance documentation.

We don’t do “quick installs” where someone pastes the GA4 tag and calls it done. That takes 10 minutes and gives you 10% of what GA4 can actually do.

How Does ScaleGrowth.Digital Approach GA4 Setup?

Our GA4 implementation connects directly to the analytics practice within our Organic Growth Engine. Every GA4 property we configure feeds data back into the engine, which means your SEO, content, and paid media decisions are informed by the same measurement framework.

Here’s what you get:

Measurement plan document. Before we touch GA4, we map every business objective to KPIs, every KPI to GA4 metrics, and every metric to the events and parameters that feed it. This document becomes your analytics reference for the next 12+ months.

Full seven-layer implementation. Property setup, enhanced measurement tuning, custom events, key event configuration, audiences, data layer, and BigQuery export. Nothing skipped.

QA and validation report. We test every event across desktop and mobile, document expected values versus actual values, and deliver a validation spreadsheet your team can re-run quarterly.

Team training session. A 60-minute walkthrough of your GA4 property, covering the reports that matter for your business, how to build explorations, and how to interpret the data without getting lost in GA4’s interface.

Frequently Asked Questions

Is GA4 free to use?

Yes. GA4 is free for all websites and apps, including the BigQuery export. Google Analytics 360 (the paid version) exists for enterprises needing higher data limits, SLAs, and additional features like sub-properties and roll-up properties. Most businesses under 10 million events per month don’t need 360.

Can I still access my old Universal Analytics data?

Google shut down UA data access in July 2024. If you didn’t export your historical data before that deadline, it’s gone. We recommend exporting UA data to BigQuery or Looker Studio for any brand that needs historical comparisons.

How is GA4 different from Google Tag Manager?

GA4 is the analytics platform that collects, processes, and reports on your data. Google Tag Manager (GTM) is a tag management system that controls how and when data is sent to GA4 (and other tools). You can use GA4 without GTM by adding the tag directly to your site, but GTM gives you far more control over event tracking and is strongly recommended for any implementation beyond basic pageview tracking.

What if I already have GA4 but it’s not configured properly?

This is actually the most common scenario we encounter. We’ll audit your existing GA4 property, identify what’s missing or misconfigured, and bring it up to our seven-layer standard. The audit itself takes 2-3 days and results in a gap analysis document showing exactly what needs to change.

Do you handle GA4 for mobile apps as well?

Yes. GA4 natively supports both web and app data streams within a single property. For apps, we work with Firebase Analytics (which is the app-side SDK for GA4) and configure the same event architecture across web and app to give you a unified view of user behaviour.

Ready to Get GA4 Working for Your Business?

If your GA4 property is either missing or underperforming, we’ll fix it. Book a 30-minute call and we’ll walk through your current setup, identify what’s missing, and scope the implementation.

Book Your GA4 Consultation | View All Analytics Services