Why I’m Not Worried About Intelligent Tracking Prevention – A Digital Analyst’s Frustration
July 22, 2019 -
TL;DR: Most of what a web analyst can (and should!) do is still possible in the world of ITP 2.1 and 2.2.
Apple’s Intelligent Tracking Prevention (ITP) 2.1 released in February this year. Just as people were starting to get their bearings the most current version, ITP 2.2, rolled out. The web was a flurry of activity as digital marketers and agencies panicked and searched frantically for answers. Major outlets looking to capitalize on the chaos and confusion quickly ran stories loudly proclaiming statements resembling “The Internet As We Know It Is Dead” and other such click-bait nonsense authored by so-called “industry experts”.
Thankfully, actual industry experts (Like Simo Ahava, Tealium’s Ty Gavin, Jeffrey Mann at Rise Interactive, and our own Pat Cooney) who have informed opinions on what ITP actually means to the future of digital analysis weighed in to the commentary. As an Analyst myself, I’d like to add my two cents to this conversation. To put it succinctly:
Some data-driven decision-making will be affected by these releases that you should be aware of but, overall,
ITP has surprisingly little impact on the vast majority of the data analysis that can be done in web analytics tools
My point is really simple – there are more applications of first-party web data than remarketing.
In order to make this quick and easy, let’s consider what analysis can be done with web data by defining a group of teams that action decisions based on web data in an organization. In 10 of the 11 below use cases, ITP restrictions have no impact on analytic ability. This is not the result of cherry picking. ITP has a surprisingly little impact on how I use data to help companies every day.
Note: I’m presuming an advanced implementation that leverages data uploads and a plethora of custom tracking related to business goals. These are all use cases I have seen enabled.
The single use case impacted by ITP is marked with an asterisk ‘*’.
|Team||Example of Relevant Data||Example of Action|
|Digital Marketing – SEM*||Conversion rate for a strong performing campaign*||Adjust budget to increase spend on successful campaign*|
|Ad Ops||Ad revenue per session, by page||Submit a business case to the UX team to make a higher performing page have more navigation options from the homepage|
|Front-End Dev||Latest Safari version is experiencing significant increases in bounces in-line with increases in load times||Adjust team priorities to have regression testing performed on Safari immediately to determine whether there is a technical root cause to address|
|SEO||Landing page bounce rate for SEO sessions||Adjust SEO strategy if a set of terms is good at generating sessions, but those sessions are poor quality leads|
|UX or UI||List of preceding pages for the 404 error page by exits||Determine navigation buttons that require adjustment and submit to front-end dev team|
|Merchandise||Customer review average/median for a product SKU by year produced||As production facilities for different products change, leverage data to correlate changes in customer satisfaction and avoid using production facilities that produce products our customers do not like|
|Order Fulfillment & Inventory Management||Median wait time to ship by products for specifically perishable goods||Develop forecasted inventory and revenue loss from specific perishable products. Inform timing of purchases from suppliers to prevent revenue loss and opportunity cost of warehousing|
|Blog or Site Content Team||Value of support pages based on customer lifetime value. Additionally permutate this data based on author||Establish estimated return on support resources producing additional support content vs focusing those resources on retaining customers by being available on the phone|
|Helpdesk / Customer Support (Call Center)||Success rates of support pages based on author||Determine which employees should be prioritizing producing additional support content vs focusing those resources on retaining customers by being available on the phone|
|Helpdesk / Customer Support (Site Content)||The last five pages and last two internal search terms used by people who called help center||If I know which support content pages are not sufficient for customer needs, I will adjust them based on customer needs|
|Does using the same photographer for the email images and the product pages on the PDP linked result in higher sales of that product?||If I know that consistency of photographer improves conversion, I will adjust my email content or linked PDPs to optimize revenue post-clickthrough|
|Affiliate Marketing||Do any of my affiliates generate fraudulent clicks and try to blame low conversion on the UX/content teams?||If I know there are non-human indicators in high proportion among specific affiliate sources, I will threaten them with legal action and/or demand specific language banning them from my affiliate programs through my vendors the next time we renegotiate terms|
In the case of our first example (conversion rates for a particular campaign), sessions being attributed to that campaign will be reduced under the restrictions of ITP and therefore you should expect to see an increase in sessions and transactions attributed to direct traffic instead of paid marketing efforts. In no other listed example does ITP impact an analyst’s ability to provide metrics and inform data-driven decisions.
But, Ed, you’re washing over the fact that inter-session analysis is no longer possible!
A few important responses to that claim:
- Inter-session analysis is still possible on Chrome, Internet Explorer, Edge, and Opera, which make up well over 60% of web traffic. There is no public information indicating any intent to adjust 3rd party cookies on these browsers.
- Cookie-based inter-session analysis has always been biased by the fact that cookies are not users, no matter how much Google Analytics, IBM, & Adobe like to pretend they are.
- Inter-session analysis is still possible through User ID tracking, and while that’s only going to be possible for logged in users, it does allow you to isolate for the behaviour of a particular person, instead of a cookie.
There are hundreds of default dimensions available in web analytics tools that are unaffected by ITP. Get out your thinking cap and focus on the ones that still work.