Does the “private mode” kill the cookies? (3/7)

Screenshot from Google Chrome incognito-mode

Every one of us online marketers knows it and every one of us has used it before – the “private” or “incognito” mode. In the private mode of the browser you surf without leaving permanent traces. The browser history and form data are not stored and cookies, as well as website data, are deleted after the application is closed. So far everything is known – but what influence does the private mode have on the attribution of conversions and, above all, how widespread is its use?

More widespread than expected

According to a study conducted by duckduckgo in 2017 (“PDF: A Study on Private Browsing: Consumer Usage, Knowledge, and Thoughts“), 46% of 5,700 US citizens surveyed know about the “Private Browsing Mode”. Of these, 33% use it daily and 25% at least once a week (desktop + mobile) – in other words, ¼ of those surveyed use it at least regularly. Interesting: the usage on desktop and mobile is about the same.

Private Browsing Frequency by Device Type
Screenshot: duckduckgo study

The reasons for using the private mode are also very revealing – after all, 25% said they explicitly used it when shopping:

Reasons for Using Private Browsing by Device Type
Screenshot: duckduckgo study

Those who use it daily use it much more often for “general searches” and shopping:

Reasons for Using Private Browsing by Frequency
Screenshot: duckduckgo study

Taboola comes to a similar conclusion in a blog post from 2019 – what is interesting is that Germany uses the private mode the most:

% of traffic with session cookies
Screenshot: Taboola

The Cookie Killer

First of all, you need to know that the private mode is not as private as you think – duckduckgo’s study will go on to answer the questions of what users think and what it actually offers. A good summary of what it is and what it is not can be found here: “Private Browsing – What It Is & What It Isn’t

In brief: Private mode does NOT provide tracking protection – at least not in the way you expect. In the private mode, cookies can be set and read in a regular way – the only difference to the normal browser mode is that they are limited to the lifetime of the application. That means, if a user closes the browser OR the tab, the cookies are deleted. If the user visits the page again, he/she will receive a new session and thus new cookies. In some cases, such as in Safari under iOS, the cookies are also regularly deleted if a page is open for a longer period of time and the smartphone is not used for hours.

Why does this influence the attribution?

Since in private mode the lifetime of the cookies is influenced by the browser or the user, the risk of “losing” them increases. The result is that the user receives a new session and thus the information where the user came from is also lost. In Google Analytics (and all other web analysis systems), the user is then assigned to a different channel than the original one (Direct, Other or Referal), which in the case of a conversion distorts the attribution as already known.

Example: A user uses the private mode and clicks on a Facebook ad. He is redirected to the shop including UTM parameters. Google Analytics records the impression and correctly assigns the call to the Facebook channel. The user is browsing around in the shop, does not buy directly, but closes the tab or simply puts his smartphone aside (in the case of Safari/iOS). In the evening he visits the shop (again in private mode) and now makes his purchase. In normal browser mode, the customer would now usually have a cookie, with which Google Analytics would assign the purchase to the Facebook channel. In private mode, however, all cookies were deleted by closing the tab. This means that the customer is a completely new user for Google Analytics who has visited the shop directly. Ergo, the visit and especially the ecommerce turnover is assigned to another channel (probably “Direct”).

The Facebook channel is thus weakened, since the visit is recorded but the conversion cannot be assigned to the channel. This means that “Direct” becomes more and more profitable at the expense of all other channels as more users use the private mode.

Intentional or unintentional – the result is the same

In my next article, “The Reason for increasingly poor conversion of paid channels on mobile” (Part 4) I will like to get to the bottom of a myth (or at least refute it a little). But I would like to start with it already at this point in the context of the private mode.

In addition to the variant of opening the private mode specifically for a single action (e.g. for a purchase or for “sensitive searches”), there is another application scenario that can cause problems: The permanent use of the private mode.

While on the desktop this is only possible with explicit settings and even quite awkward with some browsers, on mobile it is completely normal for many people to surf permanently in private mode. Many users switch to private mode once and then forget that they are still using it – or they deliberately stay in it. On Mobile, NO new browser application/window is opened, but the current browser instance simply remains open. On iOS, Private mode reloads even after Safari is completely closed, if it was last used.

I.e. the permanent surfing in private mode for days on end is in my opinion more realistic on mobile than on desktop – and with it all the problems for attribution.

You want numbers?

How much the private mode actually influences the attribution is unfortunately very difficult to measure. Until June 2019 it was possible to find out with a few “hacks”, but current solutions work only very unreliably. Nevertheless, I recently included such a private mode check on everysize for a few days for testing purposes. To compensate for the unreliability (e.g. it doesn’t work in Safari under iOS 13) I statistically extrapolated the numbers and came up with a private-mode rate of about 8%!

Hat dir der Beitrag gefallen?

Author: Eugen

CTO & Co-Founder of everysize.com

Leave a Reply

Your email address will not be published. Required fields are marked *