The browser you are using is not supported.

Try one of them Google Chrome, Mozilla Firefox, Opera

SAFARI – INTELLIGENT TRACKING PREVENTION (ITP): PROBLEM, CASES AND SOLUTION RESULTS

SAFARI – INTELLIGENT TRACKING PREVENTION (ITP): PROBLEM, CASES AND SOLUTION RESULTS

Let’s start with what the problem is:

The implementation of ITP has reduced how long information can be stored in cookies for Safari users, currently at a limit of 7 days or as little as 24 hours if the user is routed through a flagged tracking domain before arriving at the website. These updates were put into place with ITP 2.1 (Feb 2019) and 2.3 (Sept 2019) mainly. Please do get in touch with us if you have any questions or if you need more information.

The effect of this is that tools using cookies to keep track of user data are negatively impacted, affecting Analytics, A/B testing and marketing/Ads tools.

This reduction in cookies lifetime leads to both over reporting of “unique users” and “exposure to unique users” metrics, as well as decreasing effectiveness of marketing/ads tools trying to retarget users in timeframes over 7 days.

With Safari being one of the most used browsers in Scandinavia (especially on mobile) this is important for all website owners to know about and consider in terms of evaluating their performance and for understanding the effectiveness of any tool relying on cookies.

Why do we see this as so important?
For us this relates to one of our cornerstone values for our work, that all data used for performance evaluation and daily optimisation should be as qualitative as possible, for both short- and long-term effectiveness and results. Because of this we not only supply SEM/SEO services but also Analytics and Tracking services to make sure we can deliver the highest quality in our partnerships for our clients.

What results can we see in cases with and without our solution in place?

Bellow we have 5 different cases (3 without solution and 2 with solution) of websites from different verticals with short/medium/long purchasing cycles. Showing how these sites have been affected, and the results of the sites that had the solution implemented.

Case 1: Business with medium length purchase cycles
We see a clear pattern emerging after both ITP updates, where Safari is clearly having an increase in % New Sessions compared to the average and peaked out at close to 70% in Feb 2020 and in May 2020 having the largest gap compared to chrome to date with Safari at 63% and Chrome at 41%.

Business with medium length purchase cycles ITP Case Study

Case 2: Business with long purchase cycles
We see an even more clear situation than Case 1, where the divergence is very high where Safari is reporting 55% New Sessions in May 2020 while Chrome is reporting closer to 30%.

Here we can see that the calculated % New Sessions is likely overreporting around 80% compared to what would be reality without these shorter cookies (55% vs 30%).

Business with long purchase cycles ITP Case Study

Case 3: Business with short purchase cycles
With short purchasing cycles we expect to see a smaller discrepancy, which we do compared to case 1 and 2.

Here we see Safari having a slower increase over a period from ITP2.1 while Chrome initially have been stable or decreased until Jul 2019 where it clearly is decreasing compared to average while Safari continues.

Later in Dec 2020 we see what looks like some issues where 90% of chrome users have been flagged as new (this is something specific for this site and not something we see in general, likely due to some change related to the website), after that we are seeing a decrease over time in Chrome while Safari is increasing.

Business with short purchase cycles ITP Case Study

Case 4: Business with long purchasing cycles – With Solution implemented
Here we see a quite big increase in % New Sessions in Safari immediately after the release of ITP 2.1 that stabilises around ITP 2.3, after the implementation of our solution we see a clear shift where Safari converges towards the Average and Chrome.

Chrome as can see has had a long term decrease which is expected for a long purchasing cycle, however in the beginning of the year we can see an increase happening which is related to an increased influx of new users, the only reason the Average is also decreasing in this period is that Safari visitors are normalising after the implementation.

Business with long purchasing cycles - With Solution implemented ITP Case Study

Case 5: Business with short purchasing cycles
In this case we see a site that has majority of users on Safari and mobile, connected to the short purchasing cycle, this site has also during this period seen a huge amount of growth and we expect to see high numbers of % new visitors.

After ITP 2.1 we directly see a sharp increase in Safari that is much larger than on chrome that continues through ITP 2.3 until the solution was implemented and we now see % New Sessions returning to numbers as they were before ITP 2.1.

To note specifically here is that both are below the average which is very high due to in-app Safari that also after the solution is causing the average to decline sharply.

Business with short purchasing cycles ITP Case Study

How to review this issue with your GA data:
By breaking down data from Google Analytics (GA) for “% New Sessions” by the main browsers like Safari and Chrome we can look at how this has affected different websites and you can review how it has affected yours.

Every website is affected differently often connected to user behaviours in different verticals, doing this will show how your website is affected.

Follow these steps in GA to see how this affects your website:

  1. In GA go to Audiences -> Technology -> Browser & OS
  2. Select a long timeframe from Jan 2018 till end of most recent month
  3. Change the graphs granularity to month instead of day
  4. In the graph change the metric to % New Sessions
  5. In the bottom table on left side check the rows for Safari and Chrome
  6. Then click the Plot Rows button just above
  7. And done!

You would now in GA see a graph that looks something like this – where in this case we can see a clear divergence happening with Safari compared to Chrome users:

ITP divergence between Chrome and Safari users

Conclusion
In conclusion we can see that all sites are impacted by this no matter what vertical, and we can see that quickly after implementation numbers are normalising. On the % New Sessions data giving more accurate numbers on total # Users and % New User acquisition.

Implementation would help guarantee that the data quality increases for decision making and follow-up.

These cases show you how big the effect is on user data for GA, and the same is happening for all other tools and platforms relying on cookies for IDs/user data.

Looking at these number we can see situations like Case 2 where you have gone from 30% to 55% which is almost an 100% increase, meaning that for every person visiting the website through safari they are commonly logged as 2 unique users instead of 1, both fragmenting the data by splitting user history and increase the absolute # of users registered.

Sites with shorter cycles don’t see as dramatic an effect, but still see large enough effects to create issues with data fragmentation and inflation of numbers.

Beyond just the issue with data-quality there is also the complications this has on all other tools used, making any long-term analysis or retargeting of users less qualitative or inflated in numbers.

For more details regarding what impacts this has on your website or getting information regarding our solution please get in contact with [email protected].