A year and a half into Google Analytics 4 (GA4), merchants are still wrapping their heads around it. While GA4 packs some impressive new capabilities, it also introduces subtle changes that can significantly impact your reporting accuracy.
In this article, we'll share seven practical tips to help you get the most out of GA4. Whether you're wrestling with attribution or just want your reports to make sense again, we've got you covered.
1. Don't Obsess Over New vs. Returning Metrics
New vs. returning user metrics have always been tricky in Google Analytics, but GA4 makes this even more complex.
While Universal Analytics would auto-generate user IDs and default to "new" users, GA4 requires pre-existing anonymous user IDs and session IDs for every hit. Without these, you'll see those frustrating "(not set)" values in your reports. Even worse, GA4 sometimes marks events from the same user in the same session as both "(not set)" and "new" or "returning."
Instead of fixating on these metrics, a layer cake analysis using CRM data can help you understand the split between new and existing customers.
2. Consider Server-Side Tracking
Traditional browser-based tracking is increasingly unreliable. Between iOS updates, ad blockers, and cookie restrictions, you're likely missing significant portions of your customer journey data. Server-side tracking solves these challenges by collecting data directly from your e-commerce platform.
A solid server-side solution (like Elevar), ensures 100% of your transactions make it to GA4, while maintaining more consistent user identification across sessions and collecting richer customer data.
3. Use Data-Driven Attribution (With Caution)
Data-driven attribution is a new attribution model that aims to distribute credit more accurately across your marketing channels. Unlike other attribution methods (like last click), it analyzes customer journeys to determine which touchpoints were most important for driving conversions, then weights them accordingly.
However, approach these insights carefully. Data-driven attribution sounds great, but some skeptics have noted that Google has a vested interest in showing that its own ads are performing well. And since the model is a black box, there's no way to verify its decisions.
The reports in the advertising section of GA4 let you compare last-click non-direct) versus data-driven performance.
In our analysis of five properties, Google's own channels (like Paid Search and Cross-Network) showed the biggest gains under data-driven attribution in four of them. While your results may vary, it's worth considering how your channel performance shows up in different attribution models.
In the Attribution Settings section, you can choose which attribution model to apply to the data in your account:
4. Export to BigQuery for Deep Analysis
BigQuery is Google's cloud data warehouse that lets you analyze large datasets. Exporting your GA4 data to BigQuery gives you access to raw, event-level data—essential for analyses that GA4's UI simply can't handle. Want to understand assisted conversions like you did in Universal Analytics? Or need to build custom attribution models? BigQuery makes it possible.
While it requires some technical expertise, the payoff is worth it. You'll have complete control over your data analysis, without being constrained by GA4's predefined reports and metrics.
Learn more: How to Export Data from GA4 to BigQuery
5. Accept That Sessions Are Estimates
Here's something that might surprise you: GA4's session counts are estimates, not exact numbers. That's why you might see seemingly contradictory numbers—like daily sessions not adding up to your monthly total. This isn't a bug, it's a feature (though we might debate whether it's a good one).
Don't let small discrepancies send you down lengthy troubleshooting rabbit holes. Instead, use session data for trend analysis and focus on more concrete metrics like transactions and revenue for exact reporting.
7. Understand How Sessions Impact Channel Attribution
Sessions in GA4 are fundamentally different from Universal Analytics, which affects your channel performance data.
In Universal Analytics, a new traffic source would spawn a new session. If someone came from Facebook and then returned via Google 15 minutes later, that counted as two sessions, with each channel getting credit.
In GA4, you only get a new session after 30 minutes of inactivity. This means in the same scenario, there would be just one session—credited entirely to Facebook.
This change particularly impacts channels that typically appear later in the purchase journey.
For example, if a customer starts on your site, searches for promo codes, and clicks through an affiliate link to complete their purchase, the affiliate won't get credit in GA4 like they would have in Universal Analytics.
To get a more complete picture, use the Advertising Performance reports in GA4 to compare different attribution models, and consider using raw data in BigQuery to build custom attribution models that better reflect your business reality.
