June 30 2023
Demystifying GA4: 5 Key Concepts to Consider for Data Accuracy
GA4 Portal
The arrival of Google Analytics 4 (GA4) has brought new capabilities for users, but it has brought along some new challenges too. Businesses and marketers are seeking clarity on what the handover from Universal Analytics (UA) to GA4 will mean for them. This blog will dive into some of the key details around cardinality, data processing latency and more, so you feel empowered to wield GA4 to the best of your ability.
Cardinality: Counting the Uniques
GA4 uses a cardinality model to count the number of unique events recorded. This means that the reported number of events may differ from the actual number collected. When a dimension has high cardinality, like Path Paths, the data might end up in the (other) row. This catch-all row is used when data doesn’t fit elsewhere due to row limits. This can make it challenging to track the true performance of specific values for high cardinality dimensions. Keep an eye on this when analysing your data!
New Channel Grouping – “Unassigned”
In GA4, you may notice a channel grouping called “unassigned.” Just like the “other” channel in Universal Analytics, this grouping is used when there are no other channel rules that match the event data. You can use the source/medium dimension to identify whether data in this channel needs to be reallocated or if it truly belongs in the “unassigned” bucket. Check out this helpful guide from Google for more details on channel grouping:
https://support.google.com/analytics/answer/9756891?hl=en
“Not Set” in Multiple Reports – Rise in Volume
During the migration to GA4, it’s become more common to see “not set” landing pages appearing in the top 10 landing pages for many websites. But that’s not the only report where you might encounter the “(not set)” label. If you’re curious about why this occurs across different reports, this handy guide from Analytics Mania explains it all:
https://www.analyticsmania.com/post/not-set-in-google-analytics-4/
Data Processing and Latency
With GA4, there’s a processing latency window of 24–48 hours for data. That means the data in the platform can still change during this period as it’s being processed. If you’re pulling recent data, it’s important to be aware of this window and try to avoid pulling data within the first 48 hours. Additionally, there are slightly different limits when a website reaches over 200,000 sessions per day. Here’s a link to learn more about data processing in GA4: [Link:
Calculations and collections of data
GA4 tracks events at an individual level, which is a departure from UA’s session-based tracking. This means that some events may be tracked differently in GA4 compared to UA. However, there are various other settings within the platform that could lead to a distortion of the data that you’re collecting such as IP or traffic restriction filters, inaccuracies to cookie opt in collections or the number of domains being recorded.
We’ve observed strong levels of accuracy across many client accounts when benchmarking and comparing against core sources across all channels and devices. However, if you encounter discrepancies greater than 20%, it might be worth investigating individual tracking challenges within your account. The transition to GA4 presents industry-wide challenges, but we’re here to support you every step of the way.
We hope that exploring these concepts has provided some clarity on GA4 and addressed your burning questions. Embracing the transition to GA4 is an ongoing process, and we’re here to help you make the most of it. Feel free to reach out if you need further assistance.
