Yaron Rosen, Director of Data Product
Apple ITP. Marketers need to be aware of the evolution in consumer privacy issues, such as changes in Apple’s Safari browser, and how they can impact marketing tracking and measurement.
For the last few years, Internet giants such as Facebook, Google, and Apple have been making changes to help protect user privacy on the web.
In September 2017, Apple started adding Intelligent Tracking Prevention (ITP) to Safari browser versions 11 and above. Apple ITP is a feature embedded in the browser which uses machine learning to identify trackers based on usage patterns. It then applies various restrictions on cookies and other capabilities to prevent trackers from sharing user identifying information cross domains.
ITP looks to limit the ability of tracking service providers to read their cookies from a website beyond the first 24 hours. It targets cookies that Apple believes track users across multiple sites—specifically third-party cookies and cookies that are read in a third-party context. It also blocks first-party cookies when those are identified or suspected to be used for tracking. This is based on the context which sets them (using JS vs. server side, or landing pages referred to by a tracking domain)
Intelligent Tracking Prevention is significant to all marketers—but especially ones in the U.S. because over half of all mobile browser activity is on Safari.
What is new about ITP 2.2?
When ITP 1.0 was first activated, Kenshoo released the Kenshoo global tag to ensure accurate measurement. However, in April, Apple released ITP2.2 in Safari 12.3, not only blocking third-party cookies but also limiting first-party cookies under some conditions.
In many respects, Apple ITP 2.2 makes analytics, measurement, retargeting and many other user data-based practices more challenging. Publishers such as Google and Facebook—together with independent solutions like Kenshoo—will need to rethink methodologies for accurate conversion measurement from users on Safari as well as multi-touch attribution and the machine learning optimization systems that rely on these measurements.
Attribution: Safari conversion attribution is impacted when the click preceded is more than 1 day before the final conversion event. The actual impact will be in measuring conversions, therefore you may see fewer conversions in your grid.
We have analyzed client conversion attribution behavior in the last year to learn the potential impact of ITP 2.2 on measurement and estimate the potential maximum impact on the attribution 1 day after a click.
The maximum ITP 2.2 potential impact is subject to the percentage of users using Safari 12.3 and the percentage of conversions that occur more than 1 day after the click.
Audience: User identification via a cookie is purged after 1 day. This means that a tracked user is “forgotten” after a day and will be considered as a new user after this period. Therefore, you may see shorter paths than before, as each path won’t be longer than a day.
Optimization: Optimization is based on the actual performance data. Due to the effect on the attribution and the limitation of accuracy in Safari, the optimizations are impacted and therefore may behave differently.
Custom Domain Tracking: Custom Domain Tracking is a new way to place the Kenshoo tracking cookie via HTTP response (using the Kenshoo server) which is not affected by ITP 2.2. This solution requires some work on the advertiser side.
How it works: When a user clicks an ad and is redirected to a client’s domain, the Kenshoo Global Tag calls the Kenshoo tracking servers (via a CName record) in order to set the cookie under the client’s domain. It is at that point that the user is tracked.
The solution will be available for Kenshoo clients soon.
Incrementality: Kenshoo can use an incrementality test to understand the true impact of your mobile ads.
How it works: Kenshoo’s incrementality testing will run a holdout test and measure the overall impact on ALL conversions (not just attributed). In addition, with our secondary insights analysis, we will look to measure the overall incremental effect by mobile browser. With incrementality testing, we will be able to not only capture lost conversions due to Safari ITP, but we will also see any conversions that are captured by other channels (SEO, direct, etc.) to understand the true value of your mobile investment.
After running the incrementality test, we will end up with an incrementality factor comparing your incremental KPI (incremental ROAS or CPA) to your attributed KPI (ROAS or CPA). For example, if we found mobile conversions to have a 130% incremental factor overall and you had a $50 CPA, you could now optimize to a $50 * 130% = $65 CPA.
Incrementality testing is available today as a service with a beta SAAS solution available in May 2019.
Modeling: By modeling the behavior of users we can track from other browsers, it is possible to estimate the number of conversions that occurred on Safari but were not attributed to paid media due to the Apple ITP.
This approach does not require any work on the advertiser side and no future ITP enhancements are expected to block it.
To learn more about Apple ITP 2.2 impact on your traffic, Kenshoo clients can reach out to their Kenshoo reps. For more information about Kenshoo solutions, contact Kenshoo today.
Request a Demo of Kenshoo