Le guide complet Google Analytics 4 pour les chefs d'entreprise

Google Analytics 4 is not an upgrade to Universal Analytics — it is an entirely different analytics platform built on a fundamentally different data model. Universal Analytics tracked pageviews as its primary unit of measurement. GA4 tracks events. Every interaction a user has with your website or app, a page view, a button click, a file download, a video play, a scroll to the bottom of a page, is recorded as an event. This shift changes how you think about measurement, reporting, and optimization.

The event-based model exists because user behavior has changed. People no longer follow linear paths from homepage to product page to checkout. They bounce between devices, interact with chatbots, watch embedded videos, and engage with dynamic content that never triggers a traditional pageview. Universal Analytics could not capture this complexity. GA4 can, because events are flexible enough to represent any interaction you care about, from a simple page load to a complex multi-step form submission.

For business owners accustomed to Universal Analytics, the transition can feel disorienting. Reports look different, metrics have new names, and some familiar features seem to have disappeared. But the underlying capabilities are more powerful. GA4 offers better cross-device tracking, built-in machine learning predictions, tighter integration with Google Ads, and a privacy-centric architecture that does not rely on third-party cookies. Understanding these differences is the first step to unlocking GA4's full potential for your business.

A correct GA4 setup is the difference between useful data and noise. We have audited dozens of GA4 implementations at GRADAX and found that most businesses rush through the setup, leaving default settings in place that produce inaccurate or incomplete data. Taking an extra hour during initial configuration saves hundreds of hours of confusion later when you are trying to make sense of your reports.

Start by configuring your data streams. A data stream is the connection between your website or app and GA4. For most businesses, you need one web data stream. Enable enhanced measurement, which automatically tracks page views, scrolls, outbound clicks, site search, video engagement, and file downloads without any additional code. Review each enhanced measurement toggle and disable any that are not relevant, for example, if your site does not have a search function, disable site search tracking to avoid polluting your data with false positives.

Next, set up your data retention to the maximum of fourteen months instead of the default two months. Configure cross-domain tracking if your business operates across multiple domains, such as a marketing site and a separate application. Exclude internal traffic by creating a filter that removes visits from your team's IP addresses. Finally, link your GA4 property to Google Ads, Google Search Console, and BigQuery if you plan to run advanced analyses. These integrations are free and dramatically expand what you can learn from your data.

GA4 offers hundreds of metrics, but trying to monitor all of them is a recipe for analysis paralysis. As part of our technical SEO and analytics services, we help clients identify the five to ten metrics that actually drive business decisions and build their reporting around those. For most businesses, the essential metrics fall into three categories: acquisition, engagement, and conversion.

Acquisition metrics tell you where your visitors come from. Track users by source and medium to understand which channels, organic search, paid ads, social media, email, referrals, are driving the most traffic. Pay special attention to new versus returning users. A healthy acquisition strategy brings in a steady stream of new users while retaining a growing base of returning visitors. If your new user count is flat or declining, your top-of-funnel marketing needs attention.

Engagement metrics tell you what visitors do once they arrive. GA4 introduces "engaged sessions", sessions that lasted longer than ten seconds, had at least one conversion event, or included two or more page views. The engaged session rate replaces bounce rate as the primary engagement indicator and is far more useful because it accounts for single-page visits where the user found exactly what they needed. Also track average engagement time per session, pages per session, and your most-visited pages to understand which content resonates and which falls flat.

One of GA4's most powerful features is its ability to map user journeys across sessions and devices. The path exploration report lets you visualize the sequences of pages and events that users follow on their way to a conversion, or on their way to dropping off. This is invaluable for identifying friction points, unexpected navigation patterns, and content that either accelerates or stalls the buying process.

To build a useful path exploration, start with a specific conversion event as your endpoint and work backward. If your goal is lead form submissions, set the form submission event as the ending point and let GA4 show you the most common paths leading to it. You will often discover that certain blog posts or resource pages appear disproportionately in successful conversion paths, even though they are not directly linked to the conversion page. These are your hidden conversion drivers, and they deserve more visibility and internal linking.

The funnel exploration report complements path analysis by letting you define a specific sequence of steps and measure drop-off at each stage. For an e-commerce site, this might be: product page view, add to cart, begin checkout, purchase. For a service business, it might be: service page view, pricing page view, contact form view, form submission. Identifying the step with the highest drop-off rate tells you exactly where to focus your optimization efforts. A 60% drop-off between "pricing page view" and "contact form view" suggests the pricing page is creating objections that need to be addressed.

In GA4, conversions are events that you mark as important to your business. Unlike Universal Analytics, which had a separate goal configuration interface, GA4 treats any event as a potential conversion. You simply toggle the "Mark as conversion" switch next to the event in your admin panel. This flexibility is powerful but requires intentionality — without a clear conversion framework, your reports will lack the business context needed to make decisions.

We recommend defining conversions at two levels: macro and micro. Macro conversions are the actions that directly generate revenue or leads, purchases, form submissions, phone calls, demo requests. Micro conversions are engagement signals that indicate progress toward a macro conversion, email newsletter signups, PDF downloads, video completions, pricing page views. Tracking both gives you a complete picture of your funnel health. If macro conversions drop, micro conversion data helps you diagnose where in the journey users are falling off.

For events that GA4 does not track automatically, you will need to create custom events. The most common custom events we set up for our digital advertising clients include form submissions (especially when using third-party form tools), CTA button clicks, scroll depth milestones, and specific page views that indicate high purchase intent. Use Google Tag Manager to create these events without modifying your website code. Define clear naming conventions from the start, using underscores, lowercase letters, and descriptive names like "lead_form_submit" rather than cryptic codes like "event_42."

GA4's standard reports cover the basics, but the real analytical power lives in the Explore section, where you can build custom reports tailored to your specific business questions. At GRADAX, we create a set of core custom reports for every client during onboarding, ensuring they have immediate access to the insights that matter most.

The first report every business needs is a content performance report that shows which pages drive the most engaged traffic and conversions. Configure a free-form exploration with dimensions for page path and page title, and metrics for users, engaged sessions, average engagement time, and conversion events. Sort by conversions to identify your highest-performing content, and sort by engagement time to find pages that capture attention but fail to convert, these are prime candidates for CTA optimization.

The second essential report is a channel attribution report that goes beyond last-click attribution. GA4's data-driven attribution model distributes conversion credit across all touchpoints in the user journey, giving you a more accurate picture of which channels actually drive results. This is critical for businesses running multi-channel strategies because it reveals the true contribution of top-of-funnel channels like organic search and social media that rarely receive last-click credit but frequently assist conversions. Without this report, you risk cutting spend on channels that are quietly driving results behind the scenes.

Linking GA4 to Google Ads is one of the highest-return actions you can take in your analytics setup. The integration allows GA4 conversion data to flow directly into Google Ads, enabling smarter bidding strategies, more accurate ROAS reporting, and the ability to build remarketing audiences based on granular behavioral data rather than simple page visits.

Once linked, import your GA4 conversions into Google Ads as primary or secondary conversion actions. Primary conversions are used by Smart Bidding algorithms to optimize your campaigns, these should be your macro conversions like purchases or lead form submissions. Secondary conversions are tracked for reporting purposes but do not influence bidding. Micro conversions like newsletter signups or pricing page views work well as secondary conversions, giving you additional funnel visibility without confusing the bidding algorithm.

The remarketing capabilities unlocked by the GA4-Google Ads integration are particularly powerful. You can create audiences based on any combination of events, user properties, and behavioral patterns. For example, you could build an audience of users who viewed your pricing page, spent more than sixty seconds on the site, but did not submit a contact form, and then serve them a targeted ad addressing the objection that likely caused them to hesitate. These behavior-based audiences consistently outperform generic remarketing lists because they allow you to match your ad creative to the user's specific stage in the buying journey. Contact us to set up your GA4-Google Ads integration correctly from day one.

The most common GA4 mistake we see is treating it as a drop-in replacement for Universal Analytics and expecting the same reports to exist in the same places. GA4 is a different tool that requires a different mental model. Businesses that approach it with fresh eyes and invest time in learning its event-based architecture consistently extract more value than those who try to recreate their old Universal Analytics dashboards pixel for pixel.

The second most frequent mistake is failing to set up conversions at all. GA4 tracks events by default but does not know which events matter to your business until you tell it. We routinely encounter businesses that have been running GA4 for months with zero conversions configured, which means they have no way to measure what is working and no data feeding their Google Ads bidding algorithms. Even a basic setup with three to five key conversions immediately transforms GA4 from a traffic counter into a business intelligence tool.

Other mistakes include leaving data retention at the default two months instead of fourteen, not filtering out internal traffic, ignoring cross-domain tracking for multi-domain setups, and failing to link GA4 to Google Search Console. Each of these oversights degrades data quality in ways that compound over time. The longer you wait to fix them, the more historical data you lose. If you suspect your GA4 setup has gaps, we recommend running a configuration audit sooner rather than later, the cost of bad data is always higher than the cost of fixing the setup. Get in touch to have our analytics team review your implementation.