Effective May 13, 2022 (or earlier), the Google Analytics cookie will be recategorized as a performance cookie and will no longer be saved for users who opt out of performance cookies (or all cookies), or who don't dismiss the cookie bar. In my opinion, this is a great move for privacy, and I highly support it.

However, Google Analytics data is used to power the "traffic sources" tab of the site analytics page, to which access is granted as a privilege to users with 25,000 or more reputation.

Given the prior design, the data would include all users, regardless of their cookie settings (so long as they didn't block the cookie using an extension or disable all cookies). However, effective 30 days from the day the Google Analytics cookie is recategorized as an optional cookie, data from users who've not enabled performance cookies (or who are browsing the site without dismissing the cookie bar) will no longer be reflected in the data shown to users with the site analytics privilege (and in the meantime after it's recategorized, it will only include data from prior to the recategorization).

The help center page lists out a few "oddities" in the data shown in the tab. A bullet explaining that the tab doesn't include data from those who've not enabled performance cookies should be added to the list:

This tab gives you access to select Google Analytics data from the last 30 days. There are a few oddities in the data:

  • DuckDuckGo shows up as a referring site and not a search engine.

  • The numbers are sampled so it's not uncommon to see implausible results such as several search engines at exactly 211 visits.

  • We have a limited API quota, so arbitrary date ranges are not supported.

  • Search terms are so unreliable that we just aren't showing them.

  • Data from users who've not enabled performance cookies is not included, as such users aren't served a Google Analytics cookie.

(The first bullet above, regarding DuckDuckGo showing up as a referring site, should be removed as it now shows up as a search engine instead.)

Can this edit, along with the one requested in the above link, please be made?

  • 1
    I feel like this doesn't really make a difference since the data is already sampled and not specific anyways, per the second bullet point. It definitively does not include all users, only a small percentage of them. Which makes this addition pointless in the grand scheme of the data being shown.
    – animuson StaffMod
    May 4 at 0:38
  • 2
    @animuson I was referring to the population out of whom it's sampled. Still, it's nice to list it, given that it does list oddities and this is indeed an oddity. May 4 at 0:42
  • 1
    I know, but the point of sampling is to make small inconveniences like this irrelevant. So if it's irrelevant, why mention it at all?
    – animuson StaffMod
    May 4 at 0:44
  • (Note: my comment above did not contain the sentence beginning with "Still, it's nice to list it" at the time animuson responded to it; I edited it in right before the response and the system doesn't auto-load edited comments.) 4 hours ago


You must log in to answer this question.

Browse other questions tagged .