BLOG POST

If you work in SEO or you are curious about Google Analytics 4 (GA4), you might have noticed that the landing page report in GA4 is different from the one in Universal Analytics (GA3). You might wonder why the top landing pages in GA4 are not the same as the ones you used to see in GA3 for the same website and date range.

In this blog post, we will explain the two main reasons for this discrepancy and how to fix it:

1. Reporting Identity

Your customers may use different devices and platforms to interact with your website or app. For instance, someone may visit your website on their tablet in the morning, check it again on their mobile while travelling to work, and finally buy something after dinner. These are separate sessions on your website, but GA4 can combine them into a single customer for a complete cross-device user journey. GA4 does this by using four different methods that we will explain below.

• User-ID: A unique identifier that you assign to each user of your website or app.
• Google signals: Data from Google accounts that are signed in and have opted in to ads personalization.
• Device ID: A unique identifier that is generated by the device or browser that your customer is using.
• Data Modeling: A machine learning technique that estimates the missing data when your customers decline cookies on your website. This method requires the implementation of Google Consent Mode.

GA4 allows you to choose different reporting identity options based on which of these methods you want to use:

• Blended: Uses the user ID if it is collected, then Google signals, then device ID, and finally modeling.
• Observed: Uses the user ID if it is collected, then Google signals, then device ID.
• Device-based: Uses only the device ID and ignores all other IDs that are collected.

If you have enabled Google Signals in your GA4 property to get the benefits of demographics and interests data, as well as audience and conversions export to Google Ads, you might encounter data thresholding issues. Data thresholding is a privacy feature that prevents anyone from identifying your customers based on the data in the reports. You can see if data thresholding is applied to your report by looking at the data quality indicator at the top of the report, as shown in the screenshot below:

To remove this alert and access all the data in the report, you can change your reporting identity to “Device-based” by going to Admin > Data display > Reporting identity. This option does not rely on Google Signals, so you will not see data thresholding alerts in your report, and you will see that the report will be closer to what you see in GA3. Note: Only users with Editor or Administrator roles in GA4 can change the reporting identity. They can switch this setting as many times as they want without affecting the GA4 property, as it is backwards compatible. Also, the reporting identity changes are applied to everyone who has access to the GA4 property.

Google has recently added an option to disable Google Signals in reporting while still using it for audiences and conversions export to Google Ads accounts for remarketing and bid optimisation. This option can be found in Admin > Data collection. This will reduce the chances of encountering data thresholding issues, but also limit the enriched data provided by Google Signals in reporting for a holistic user-journey on your website or app. Use this option at your discretion. Note: Turning off the option “Include Google signals in reporting identity” is not backwards compatible. This means that if this option was enabled in the past, your reports will still show limited data due to data thresholding, and only from now on this will stop happening.

2. Estimated Data via Google Consent Mode

The other common reason is if you have implemented Google Consent Mode on your website or app. This allows you to communicate your customer’s cookie or app identifier consent status to Google. For example, if your customer denies cookies, then cookieless pings are still sent to Google, and Google can use machine learning to estimate the data that would otherwise be missing in GA4 because consent was not granted. Google models the behaviour of customers who decline cookies based on the behaviour of similar customers who accept cookies. Modelled data allows you to gain useful insights from your reports while respecting your customer’s privacy.

You can see when data is being estimated by clicking on the data quality icon, as shown in the screenshot below:

Because GA3 does not use modelled data, you may see differences when comparing it to GA4. You may find more traffic landing on certain pages because Google is recovering this data that would otherwise be lost.

Conclusion

In this blog post, you learned how to align the landing page in GA4 to what you were used seeing in GA3 by understanding the different methods that GA4 uses to identify your customers across devices and platforms. You also learned how to turn off Google signals in reporting to avoid data thresholding issues, and how using Google Consent Mode to fill in the gaps in your data whilst respecting customer privacy can contribute to differences in reporting.

Thank you for reading this blog post. If you have any questions or feedback, please leave a comment below or Contact us. We would love to hear from you!

Vitor Morais

I am the Martech Ninjas founder, covering all things digital analytics. With 10+ years’ worth of experience in marketing and analytics, and having the privilege of working across 4 different countries, I have seen it all! Something I have always seen in the world of digital analytics is how poor data quality and the lack of standardisation is hurting businesses in their data-driven decisions to achieve growth. When I’m not immersed in the world of digital analytics, I enjoy working out and try new cuisines – yes, in this order!