User ID – Cross-Device Tracking

According to the ARD/ZDF Online Study 2013, “every online household now has 5.3 internet-capable devices available.” The topic of multi-screen usage, where two devices are used simultaneously, is consequently becoming increasingly important. Overall, 28% of Germans use devices parallel, meaning they use a second device like a tablet or smartphone in addition to watching TV, and 11% are second-screen users. With second-screen usage, TV viewers interact directly with the program via their tablet or smartphone, e.g., through Twitter or an app. This user behavior is already being extensively researched by market research to enable significantly more precise targeting of advertisements.

This examination is now also supposed to make its way into online business. Many online shops want to get to know their customers better and learn more about their customers' usage behavior. This also includes learning portals and other websites that offer services for a very specific target group and want to address them more accurately, regardless of whether it's on a computer, smartphone, game console, tablet, etc.

This is where Google Analytics Universal comes into play. The most widely used web analytics software has offered the “User-ID” feature since the beginning of April 2014, which allows users to be tracked across different devices. This subsequently gives companies the opportunity to recognize which devices customers used to navigate on their website. A possible presentation of the data is shown in the following image.

[caption id="attachment_13663" align="aligncenter" width="349"]

Screenshot: Device overlaps on a website. As of: 04/11/2014[/caption]

How does User-ID work and what requirements must be met?

As you can gather from the above description, Google Analytics Universal follows a significantly more user-centric approach than its predecessor version. While with the old Google Analytics “Classic,” the tracking code automatically generated and assigned a so-called Client-ID, now in Google Analytics Universal, it can be independently assigned to track users across various devices.

The old Google Analytics would generate a “Hit” with a randomly and uniquely generated “ID” for each page view by a user. This ID was unique for each user and the device used and was only generated on the first page view. It was then stored in a cookie on the user's device. If a new ID was generated, Google Analytics counted a new user. If Google Analytics recognized an ID, it treated the user as a returning visitor.

These IDs could be deleted by the user clearing their browser cookies. Simultaneously, the same user received another ID on a different device, e.g., their smartphone, as they were using a new device. In this case, the user could not be recognized.

To solve this problem, companies can now overwrite the IDs in the sent hits and equip them with their own IDs. Companies often utilize data from CRM systems, databases, or independently generated IDs for this. It's important that these IDs are correctly read from the database or newly generated and transferred to the website before appearing in the Google Analytics Universal source code of the website.

But how can it be ensured that a user's browsing behavior can be tracked across various devices? For this, the website must be able to identify users. This is usually achieved via a login.

If a user logs into a website and then browses it, they are assigned their very own personal ID. This ID remains constant, so they receive the same ID on every new login. It doesn't matter whether they use their PC, smartphone browser, or any other device. Logging into a smartphone app can also be captured this way. Since the ID is randomly and uniquely assigned, and the user's IP address is anonymized, the user cannot be identified or located. Consequently, there are no data protection-related issues.

To also consider the browsing behavior of users who were not logged in at the start of a visit and thus received a new ID, the optional “Session Unification” feature can be used. With this, Google Analytics Universal can combine the browsing behavior of a user who has two different IDs (logged in and not logged in). However, this only applies to the browsing behavior that occurred before logging in. If a user continues to browse the site after logging out, this browsing can no longer be assigned to the user.

How can you evaluate User-ID data and adjust your website accordingly?

To evaluate User-ID tracking, an additional User-ID data view must first be created. This view lists all visits with User-ID tracking, allowing companies to differentiate between visits from users with and without User-ID.

Within the data view, user interactions can be analyzed according to users with User-ID and without User-ID. Additionally, the order of devices used during access, the generated revenue (with e-commerce tracking enabled or goals attributed a monetary value), etc., can be analyzed. Particularly interesting is the