Google Analytics 4 (GA4) is replacing Universal Analytics (GA3). To prepare everyone for the change, I’ve created a series of blog posts to help users understand the difference between the two versions. Today’s post is a guide to helping you compare your GA3 reporting with your GA4 reporting. You’ll learn how increases and decreases of certain metrics may be more representative of your shift to GA4, rather than optimizations you made as a marketer.
Sessions
Sessions are calculated differently in GA4 than in GA3. Your sessions may be lower in GA4 because:
- Session Timeout: While the automatic session duration is defaulted to 30 minutes in both versions of Google Analytics, GA4 allows you to extend that session timeout to 7 hours and 55 minutes. GA3 had a maximum timeout of only 4 hours. That means if you extended your session timeout to GA4’s maximum, someone that stepped away from your website for five hours would be counted as two sessions in GA3, while in GA4 they would be counted as one.
- Midnight Restart: In GA3, all sessions break at midnight. In GA4, this isn’t the case. This midnight cutoff is based on the time zone you set within your Google Analytics settings. For example, let’s imagine you’re an ecommerce vendor who gets a lot of traffic from across the U.S. on Thanksgiving. While your fellow East Coast residents have likely gone to bed at midnight, it’s only 9:00 p.m. on the West Coast, and people may still be up shopping. Any session that spans both 8:59 p.m. PST and 9:00 p.m. PST will appear as though you’ve gotten 2 sessions in GA3. In GA4, the session will persist and more accurately reflect the one session.
- New Campaign Parameters: If someone encountered two of your campaigns within a short time frame while browsing the web, each of those campaigns would cause a new session to start in GA3. However, in GA4, the session won’t break. For example, if someone first clicks on a paid ad, goes to your website, immediately goes back to Google and clicks on your Google Business Places profile, GA3 would count that as two sessions. In contrast, GA4 would only count that as one session.
- Automatic Bot Filter: GA4 has a bot filter that removes known bots and spider traffic from your website. Google populated this list based on their own research and the International Spiders and Bots List. GA3 did not have such a filter.
There’s one reason, however that your GA4 numbers may be higher than your GA3 numbers:
- Manual Filters: At the time of writing this post, GA4 does not allow for filters beyond IP filters. This lack of filters means that if you were filtering out hostname traffic in your GA3 property settings instead of configuring your Google Analytics tag within Google Tag Manager, your unwanted hostnames will appear in GA4.
Conversions
GA3 groups all website activity within a session. In doing so, Google also deduplicated any conversions within that session. If someone filled out a form 8 times, Google Analytics counted that as one goal completion. Conversely, GA4 isn’t a data model based on sessions. As a result, in GA4, those 8 form fills would be counted as 8 conversions. This difference means you will likely see an uptick in GA4 conversions compared to GA3, whether or not you make any website optimizations.
It’s also important to point out that because of this lack of deduplication, the transition to GA4 is a great time to test the validation on your forms. If your conversion event fires when you click a “submit” button without ever filling out a form, you likely need to tighten up your event logic to more accurately reflect your lead count.
Average Session Duration
To calculate the average time on page metric in GA3, Google would subtract the time you visited page 1 from the time you visited page 2. The inherent flaw was that Google Analytics had no way of measuring the time someone spent on your exit page because there was no subsequent pageview from which to subtract.
Because Google could never figure out the length of time someone spent on an exit page, GA3 always undercalculated average session duration. Sometimes this was undercounted by mere seconds. Other times, it was under calculated by 10 minutes or more.
GA4 measures time differently by sending a timestamp with every event. I dive deeper into the comparison in another blog post, but the bottom line remains the same. When you compare your average session duration metrics in GA3 to GA4, GA4 will likely have the higher number.
Bounce Rate
GA3’s bounce rate metric was largely influenced by the amount of event tracking you added to your Google Analytics account. The more events you had, the less bounces you likely had. As a result, there is a good chance anyone with lots of event tracking would naturally have a lower bounce rate in GA3. In contrast, account owners that did not implement any custom event tracking would naturally have a higher bounce rate.
In another blog post, I explain why GA4 bounce rate is a little harder to inflate. With GA4, event tracking will only decrease your bounce rate if you count that event as a conversion. The result is that if you had a super low bounce rate in Universal Analytics due to lots of event tracking, your GA4 bounce rate will likely be higher. Inversely, if your bounce rate was incredibly high in Universal Analytics, don’t be surprised if it lowers in GA4.
Summary
Below is an easy takeaway of what you should expect when you compare your GA3 to your GA4 reporting.
| Metric | Results in GA4 Reporting |
|---|---|
| Sessions | Likely lower than GA3, but it depends |
| Conversions | Likely higher than GA3 |
| Average Session Duration | Likely higher than GA3 |
| Bounce Rate | It depends |
Make sure to take this new measurement into consideration when doing analysis so you don’t inadvertently make an optimization decision based on incorrect assumptions. GA4 is continuously rolling out new features. Check back for updates or dive into other comparisons of Universal Analytics and GA4.