Nevertheless, all of them have a goal – at least since GDPR popped in – to secure the data of the user. Of course, giving the user the best experience is also important, but isn’t privacy a part of that?
Firefox, for example, blocks third-party cross-domain cookie tracking by default, while Chrome wants you to label your third-party cookies to be transparent.
Safari, on the other hand, takes a different approach to the user’s sensitive data. The modus operandi of Safari is to simply prevent cross-site tracking entirely by blocking third-party cookies.
How can this work together when we use enterprise solutions like the Google Marketing Platform? More importantly: How does this impact your data measurement and attribution model? Keep in mind: about 8-15% of all internet users worldwide use Safari as their browser either on their mobile device or desktop.
What is happening?
We have seen conversion behaviour change within Analytics over the past few years and especially months when it comes to browsers – and especially concerning Safari users.
This is due to the fact that Safari’s browser engine, Webkit, released the first Intelligent Tracking Prevention (ITP) in June 2017. Following several adaptations and developments, its latest release in December 2019 (ITP 2.3 enhanced) drastically changed the way we interpret data.
The biggest difference between ITP 1.0 and ITP 2.3 enhanced is that Safari does not track third-party cookies and limits the lifetime of first-party cookies to 7 days, and in some instances, it’s only 24 hours. This happens if the domain sending the user to the website is classified with cross-site tracking capabilities and the final URL of the navigation has a query string and/or a fragment identifier.
Excursus: third-party cookies & first-party cookies Third-party cookies are created by domains other than the domain the user is visiting, i.e. from a banner on your website. A first-party cookie is generated from your website when the user lands on your website.
In a Nutshell, the outcome in the GMP will be that:
- More conversions will be attributed to new users instead of returning, and also, to the wrong channels
- View-through conversions measurement is impacted
What does it look like?
You might be thinking if the release of ITP 2.3 just occurred last December and there are almost 2.5 years between it and the launch of the original version, it might not affect your data. This assumption will result in inaccurate measurement as we see this behaviour happening:
The line of ITP 2.2 is also present before its release date to simplify the reproduction of the total combinations of device, browser version and operating system.
You can see with the introduction of each new version, the previous ones decrease and keep a portion of traffic over time. This is due to users who do not update their devices or the safari app.
ITP 2.3 is the most recent version, which negatively impacts the ability of tracking returning users in GA and of course, Floodlights.
As you can see, every previous ITP Version will be dominated by the updated version.
Especially with ITP 2.3, you see a lot of your previous conversions are attributed to a lower time frame than before (i.e. instead of a conversion delay of 10 days, it is now just a few days). In the worst case, you only see one channel or a single campaign, when previously there have been multiple touchpoints. As a result, your conversion may not be allocated to the right channel. You might also, see huge discrepancies between different measurement tools, like Campaign Manager and Google Analytics.
How to act now?
An important note is that we do not recommend any workarounds. This is because each new ITP is built to combat any workarounds to previous versions. We recommend sticking with your partner’s suggested solutions at all times.
We often see fancy server-side script solutions, forcing the lifetime cap to extend. Safari is aware of that, so they will always try to prevent your actions to solve tracking prevention.
A Solution for the GMP:
Within a Google Marketing Platform context including Google Ads, once you implement the Global Site Tag for your conversion tags/Floodlights, your cookie lifetime will extend to 24 hours of those originally third-party cookies. The click-ID created when the user interacts with your ad will be read on the website in a first-party context. Besides updating your tags to the Global Site Tag, this requires the activation of Auto Tagging (Google Ads) and/or Enhanced Attribution Data (Campaign Manager). Alternatively to the Global Site Tag, you can also use the Google Tag Manager to get your data right.
Why is it still limited to 24 hours? If the Safari browser detects URL decoration by containing DCLIG parameters, it will limit the cookie lifetime to 24h if the referring URL has been classed as a tracking domain like for example doubleclick.net.
Your Google Analytics data will be valid for 7 days (as no navigation from a domain classified by ITP is happening).
Excursus: Google is estimating your conversion data Google is estimating your conversion data based on your history data. Therefore you might see conversions attributed in Safari under ITP 2.3 for a longer period of 7 days.
What is the impact on your attribution data?
Building a comprehensive attribution strategy for conversion measurement is now limited. You have to rely on intra-channel measurement and have the cross-channel measurement and attribution within GA360 as a final solution (by using data from Campaign Manager, Google Ads, SA360 and Google Analytics itself).
As mentioned in the beginning, there are changes in third- and first-party cookie context for Chrome (Labelling of third-party cookies) & Firefox (ETP), as well as the use of Ad Blockers & Incognito Mode. But the impact of all this stays the same: attribution and measurement of your data changes.
Many thanks to Giuseppe Muto providing extensive knowledge on this topic.
- Marktanteile der führenden Browserfamilien an der Internetnutzung in Europa von Januar 2009 bis Januar 2020, Tenzer, F., Statista, 2020
- Marktanteile der führenden Browserfamilien an der Internetnutzung in den USA von Januar 2009 bis Januar 2020, Tenzer, F., Statista, 2020
- Intelligent Tracking Prevention, Webkit.org, 2017
- Preventing Tracking Prevention Tracking, Webkit.org, 2020
- AdBlocking in the US, eMarketer Report, 2017
- An Analysis of Private Browsing Modes in Modern Browsers, Aggarwal et. al, Stanford University, 2017
- Illustration 1: own representation
- Illustration 2: own representation based on ‘The New Attribution Eco System’ by Jesús Martín, February 2020