Even though GA4 is technically the successor to Universal Analytics, the two platforms operate in so many different ways that they could easily be put into completely separate product categories.
Here are some of the areas where GA4 and UA take separate paths.
Many of your reports and explorations can take 24-48 hours to process data from your website
This quote comes straight from GA4 documentation.
In our experience this can range from a few hours to fully 48 hours depending on the load on Google Cloud. After all GA4 is a free product, and can't be prioritized over Google's paid products.
So we have to get used to not having the freshness we’re used to from Universal reports - this doesn't affect realtime reports but rather the main reporting and explorations.
The session metric is not likely to match, because both platforms are calculating them differently, and the same action might trigger a new session in GA4 but not UA. Property settings and filters can also affect session count on one or both platforms. A few reasons might include:
Let’s think of a very common ecommerce scenario - a visitor adds a product to cart and abandons the browser. When they reactivate the browser to complete the purchase (given that more than 30 mins passed!) the original session has expired, and a new one starts.
We tested attribution for this particular case - while in Universal the session attribution is passed from add-to-cart session to the purchase session, it doesn’t happen in GA4: the rules of attribution apply strictly here - meaning that an expired session in the browser tab (the one that was abandoned for more than 30 mins) gets attributed to direct/none in GA4.
If your currency of events in GA4 is not in USD there will be small daily discrepancies on the purchase level.
GA4 does conversion at the moment of data processing, EVEN if the ecommerce currency is correctly being sent and the property is set to the correct currency.
While this isn’t a concept in Universal Analytics, GA4 uses machine learning to model conversion attribution according to the trends of overall traffic on your store. This could mean that your attribution in GA4 will be different than the one in Universal Analytics. You can read more about modelled conversions here.
In Universal Analytics, we have an overall view of the traffic sources in the Source/Medium report.
GA4 has changed that and offered us two ways to look at traffic. User Acquisition report will show you how new users arrive at your website for the first time; Traffic Acquisition focuses on where new sessions came from - regardless if the users are new or returning. As you will notice in your reports, these two will be different more often than not.
There are two common scenarios seen often in practice that expose how referrals are being handled differently in GA4 vs Universal Analytics.
A) Brand sends paid traffic to Landing pages residing on subdomains
Since “by default, Analytics will not identify traffic as referral when the referring website matched the same domain of the current page or any of its subdomains” i.e. subdomain traffic is treated as ignored referral - meaning it ends up in (direct/none) traffic source by default in GA4.
How to fix this?
Place the same GA4 tag that exists on the main domain onto every subdomain. It’s important to note here that Page title parameter becomes the main identifier because there’s no way to separate homepage on the main domain and subdomains using page path only (it’s “/” in both cases)
B) Brand controls a network of promotional websites, affiliate ones, comparison sites, reviews, blogs etc
In that case, if the brand sends paid traffic to an intermediary website i.e. site that softly promotes what’s been sold on the main domain - that paid campaign will remain as a traffic source ONLY for that intermediary website but NOT for the main domain, which will see the traffic as a referral from the ‘middlemen’.
What to do about it?
Add the intermediary website to the referral exclusion list.
Littledata tracks recurring orders back to marketing campaigns where possible.
Only recurring orders from subscription checkouts that happened AFTER the Littledata Shopify to GA4 connection will be attributed. This is because the checkout session happened before GA4 was setup and so is not tracked. Many accounts have better recurring orders attribution in UA compared to GA4.
There are many more instances where a direct comparison between UA in GA4 may cause more damage than good, and we’ll keep on updating this article as we come across the real life use cases.
While comparing Universal Analytics reports with Google Analytics 4 reports can be useful, you should be aware that these two analytics platforms have different tracking mechanisms and data models, which means that their reports may differ in terms of metrics and dimensions. .