Excluding internal traffic with IPv6 addresses

Digitalization is advancing relentlessly, and with it, the continued development of network technologies. IPv6 addresses are becoming increasingly important as the IPv4 address space is slowly being exhausted. But how does this fit into the world of Google Analytics 4, which was primarily designed for IPv4? The use of IPv6 in GA4 presents both opportunities and challenges. In this blog post, we’ll explore how to use IPv6 addresses in GA4 to filter internal traffic and address the hurdles that arise from GA4’s focus on IPv4.

IPv6 has long been the new standard for IP addresses. Nevertheless, it is often still neglected in many companies. Many companies face the requirement to filter out IPv6 addresses as internal traffic in Google Analytics 4. However, the entire documentation primarily focuses on IPv4 addresses, and even the GA4 interface treats IPv6 addresses as a minor issue. If you still need to filter out IPv6 addresses as internal traffic, we’ve compiled some useful information for you below:

In principle, all operators for filtering internal traffic with IPv6 addresses can currently be used in GA4. However, Google itself has so far provided only one example of CIDR notation in its documentation regarding IPv6 address filtering.

Quote: “If you are using IPv6 addresses and want to specify a range, use the same ‘slash’ suffix to indicate how many bits of the range are fixed. For example, if the range is 0:0:0:0:0:ffff:c080:ff00 – 0:0:0:0:0:ffff:c080:ffff, specify it as 0:0:0:0:0:ffff:c080:ff00/120 (the first 120 bits are fixed).”

Source: https://support.google.com/analytics/answer/10104470?sjid=12656966587956890926-EU#zippy=%2Ccidr-notation-verwenden

I'll explain this in more detail using the following example. The following IPv6 addresses should be filtered out as internal traffic in the future:

• 2001:9e8:c048:1d00:286c:6432:85af:efc3

• 2001:9e8:c048:1d00:286c:6432:85af:efcf

• 2001:9e8:c048:1d00:286c:6432:85af:efe8

• 2001:9e8:c048:1d00:286c:6432:85af:effa

First, we need to determine the range that these IP addresses fully encompass. In this example, that would be the range from

2001:9e8:c048:1d00:286c:6432:85af:efc0 to 2001:9e8:c048:1d00:286c:6432:85af:efff

To convert this to CIDR notation, we use appropriate online tools, such as an IPv6 subnet calculator. With just a little prior knowledge of subnets, I can quickly determine that the CIDR notation is:

2001:09e8:c048:1d00:286c:6432:85af:efc0/122

122 bits of the IP address are defined as mask bits, which means that 64 addresses in this range would be filtered out for internal traffic in the future.

But is this properly taken into account in the new world of Google Analytics 4?

While there is still a corresponding entry in the Google Analytics Help for Universal Analytics, there is currently no documentation on this topic for GA4. It is well known that IP anonymization in GA4 is now performed automatically and no longer needs to be configured. However, when filtering out internal traffic, this anonymization no longer needs to be taken into account. Here, you can specify the full IP addresses or IP ranges, and only this traffic will then be filtered out.

This is an improvement over the logic in Universal Analytics, where IPs could only be filtered up to the anonymization level. With IPv6 addresses, this resulted in a disproportionate impact when masking the last 80 bits, leading to an excessively large address space being excluded. Google has now clearly found a much better approach to IP anonymization in GA4.

One final important note on testing

Google points out that filters can take up to 24 or even 48 hours to take effect. Occasionally, however, a filter takes effect immediately. This can tempt you to start testing too early. Since filters do not reliably take effect immediately, test cases should be planned carefully and executed with sufficient time. A good best practice tip here is to alternate between positive and negative test cases. This way, you can reliably determine from the testing data (in the “Test data filter name”) in GA4’s real-time traffic whether the newly configured filter is already active.