The importance of reporting Identities in GA4
1st of July 2024 marked an important day, the sunset of Universal Analytics (UA). This means all users were forced to transition to Google Analytics 4 (GA4) and let go of their beloved UA. By now, most of us have had time to familiarize ourselves with GA4’s interface and tools. However, understanding the data we’re analyzing is just as, if not more important as navigating the platform.
GA4’s introduction wasn’t just about new interfaces, it brought a fundamental shift in how data is collected, processed, and reported. With the increasing focus on user privacy, the adoption of Consent Mode—whether in its basic or advanced form—has become widespread. However, this raises an essential question: what does Consent Mode mean for the data we see in our reports?
The answer lies in GA4’s reporting identities, a feature that determines how user data is attributed and modeled. Unlike UA, where data attribution followed relatively rigid rules, GA4 employs a flexible identity system that can switch between observed (user-provided) and modeled (system-generated) data based on the level of consent provided by users.
This distinction is no longer a minor technical detail, it’s key in data interpretation. When consent is withheld, GA4 employs advanced modeling techniques to fill in gaps, ensuring continuity in reporting. Yet, these modeled insights differ from directly observed data in critical ways. Without a clear understanding of this difference, analysts risk misinterpreting their findings, leading to skewed strategies and decisions.
The different reporting identities in GA4
The differences within the reporting identities lies in how user interactions are unified across devices and sessions. This directly impacts the accuracy of your data analysis and how you should interpret the data.
GA4 offers three reporting identity options:
- Blended: This method prioritizes identifiers in the following order: User-ID, Device ID, and then modeling. If a User-ID is available, GA4 uses it to track user interactions across devices. If not, it resorts to the Device ID. When neither identifier is available—such as when users decline analytics cookies—GA4 employs modeling to estimate user behavior based on similar users who have provided data. This approach aims to provide a comprehensive view of user activity, balancing accuracy and data availability.
- Observed: This option relies solely on directly collected data, using User-ID first and then Device ID if the User-ID is unavailable. It does not incorporate modeled data, making it suitable for analyses that require strictly observed user interactions without estimations.
- Device-based: This method exclusively uses Device ID to track user interactions, ignoring User-IDs and modeled data. It treats each device as a separate user, which can lead to overcounting individuals who interact with your business on multiple devices. This approach may be less accurate in representing unique users but is straightforward in its application.
Which reporting identity should I use?
Selecting the appropriate reporting identity depends on a lot of different factors but you should mainly consider the specific insights you aim to derive.
- Blended: Provides the most holistic view by combining observed and modeled data, which is particularly useful when user consent varies. However, it's essential to recognize that modeled data introduces estimations, which may affect precision.
- Observed: Offers a clear picture based solely on actual user-provided data, ensuring that all reported interactions are directly observed. This can be advantageous for compliance and accuracy but may result in data gaps when user identifiers are missing.
- Device-based: Simplifies tracking by focusing on devices, but it may inflate user counts due to the inability to link multiple devices to a single user. This approach can be useful when user identification is not feasible, but it provides a less accurate representation of unique user behavior.
Before beginning any analysis, it’s essential to clarify what you want to examine and choose the reporting identity that best aligns with your goals.
Finally it is important to note that you can adapt your reporting identity at any time. It will not change your data in itself, it will only change the representation of it. Be careful though, when you adapt it , it will be adapted for everyone looking at the GA4 property.
Lotte Vranckx
| LinkedinThis email address is being protected from spambots. You need JavaScript enabled to view it.
