MeasureCamp Rome – Remarks on Intelligent Tracking Prevention 2.1

MeasureCamp idea has been conceived in 2012 when a bunch of analysts gathered in a pub and discussed a need for a different type of conference. The notion carried couple of elements. Firstly, the event should be free to attend. Secondly, everyone should be given opportunity to speak. Finally, there should be no announced speakers to secure informality and knowledge sharing without hidden agenda.

And so, the term unconference was born. To get the most professionally involved analysts who share passion for data the event is organised on Saturday – outside normal working hours. Peter’s O’Neill the founder of MeasureCamp organised the first MeasureCamp in Melbourne in 2012. Since then, there were more and more events held every year in more locations around the globe. In 2019 it is expected that there will be 22 MeasureCamps in 20 countries (state for October 2019).

MeasureCamp Rome took place as the second event in Italy after Milan in 2018. Responsible for the organisation, second year in the row was Anna Khan and Oleksandr Komarevych. Both members of the organising committee are locally involved analytics specialist. Among many internationals who travelled to Rome there were also many Romans and Milanese specialists.

Since the schedule was confirmed in advance, the participants collected their session cards upon arrival. You can select a session slot on the session whiteboard. There were 5 simultaneous tracks of presentations, so it was difficult to choose.

MeasureCamp Session Card

Source: MeasureCamp Italy, Facebook

Resolution participated in MeasureCamp in Italy, which was held on 18th of May 2019 in Rome. Here are the remarks regarding the event and one of the sessions attended.

MeasureCamp Alban Gerome ITP Session Participants

Source: MeasureCamp Italy, Facebook

ITP 2.1 with Alban Gerome

One of the presentations I chose was given by Alban Gérome, a Tag Management and Implementation Lead at Barclays bank in UK. Alban held two more sessions that day which I have attended but I was eagerly waiting to get to know more about this specific one as it was about Intelligent Tracking Prevention on newest Safari version. Alban outlined the problem that ITP version 2.1 and 2.2 pose for the cookie-based tracking and discusses solutions previously described in the famous Simo Ahava blog post. Since the solutions; localStorage and iFrame, reverse-proxy edge cache, CNAME record change and server-side response cookies are only temporary solutions to a pressing problem, Alban was curious whether anyone came up with other ideas. The session’s participants could only speculate on how severe the impact is and what the Digital Giants are going to come up with to substitute cookie-based data collection and advertising.

ITP Temporary Technical Solutions

Kilde: Alban Gerome, SlideShare

ITP 2.1 origin

Let us start from the beginning. In December 2015, John Wilander became WebKit Security Engineer at Apple. Banning the cross-domain data exchange has become his mission as he refers to in multiple Twitter posts quoting ‘Cross-site traffic is bad for the web’ and ‘Cross-site tracking is bad for the user’. Intelligent Tracking Prevention was primarily targeted into banning cookies which send behavioural data of web usage to advertising systems. A solution was developed to limit the cookie expiration period to 7 days, after which the cookie would be erased in Safari browser. This means that if the user does not return to a website within 7 days, he or she will be regarded as a new user. Now, it is speculated that with ITP 2.2 this expiry period will be limited to 24 hours.

Primarily, the invention impacted cookies which send anonymous user data between multiple domains, e.g. from inventory provider to advertiser allowing for behaviour-based marketing. However, beside the advertising IDs many user preferences are also stored in cookies. Hence, limiting them or disabling them entirely may in fact generate a negative user experience. The data trade-off ultimately lies in how much of privacy the users are willing to give in return for the usability and lack of friction they are accustomed to when browsing the web.

One of the key elements of providing frictionless, seamless user experience is sharing cookie data among different domains of the same company. The invention backfired on this key principle of good UX with eradicating cross-domain tracking (e.g. example1.dk and example2.dk) as well as passing a cookie from a top-level domain child domain (e.g. example.dk and www.my.example.dk). Apart from user experience, quality of data in rollup properties in analytical tools was also hit. The severity of the impact is affected by the percentage of traffic generated by Apple devices compared to other devices as well as the percentage of Apple users updating their operating system to 12.1 version or newer, since these roll out ITP automatically.

ITP 2.1 impact

Inspired by Alban’s session on MeasureCamp, We created the ITP – the Impact Assessment Tool, which answers three questions:

  1. What browsers do your users use?
  2. How big is the portion of Safari users who switched to version 12.1 or newer?
  3. How big is the impact on your data (7 days cookie expiry cap)?

ITP 2.1 Impact Assessment Tool

Source: Google Data Studio – Intelligent Tracking Prevention: Impact Assessment Tool

 

Answering these questions allows us to approximate the effect of ITP on each Google Analytics account and propose some temporary technical fixes.

Cookie alternative

For a couple of years now, the general trend within the data industry is to look beyond cookie-based tracking. The popularity of CRM data integration into online tracking provided another touch point of user identification and so did passing anonymised ID along user login authentication. The trend implied that multiple cookie-based sessions could be stitched together if the user identifies him or herself.

In Adobe, a unique visitor ID has been assigned while in Google a user-ID view feature has been released. Both features connects sessions or visits under single user. Also, both solutions implied user identification within their company networks. Additionally, Google and Facebook released a third-party site sign-up feature using personal data stored in accounts of the users. This technique got a name SSO (Single-Sign-On) and is now highly popular for users registering to new services with their existing Facebook or Google accounts. Similarly, Apple identifies users through Apple ID.

A popular speculation regarding how these developments are interconnected and connected with introduction of Intelligent Tracking Prevention is that we can expect a shift from cookie-based data to user-based data through networks of identified users.

 

ITP 2.1 implications

A shift to user-based data would push many websites to build a sign-up feature, even if the purpose of having the website may be strictly to market products and services. In industries like news or hobby-related publishing users are accustomed to getting information for free.

Even though these users’ data is being shared with advertisers in exchange for access to free information, users often argue that there is no real cost associated with accessing these services. The reality of the SSO web is possible but it does not mean that it will put a stop to advertising on the Internet.

Just as it is possible to write an advertising ID into a cookie, it is also possible to store it in a user ID. So, in fact, the ITP invention is likely to disrupt the technology (cookie) but it will not shift the status quo.

A users own data as a form of currency is becoming popular as being discussed in the Danish Parliament. In such scenario, the data becomes an asset that a user can decide to share or not to share in an exchange.

There is a new browser – Brave on the market that awards users for browsing. The awards can the be given to online creators the users care about. If a browser like Brave gains enough momentum it may even expand to offering instant gratification in a forms of crypto currencies.

This ultimately will put the user in the real control by commercialising their own data. The advertisers would then only be able to advertise to those who decides to share their data, hence they provide an informed consent and can benefit financially from their assets.