Device usage patterns
A very common pattern that can be seen in analysis of web device usage is that the users start the day using portable units such as mobiles and tablets. Later in the day desktop computers dominate, and in the evening the portable units increase in volume again. Device switching needs to be taken into account in web analytics, and as a result cookie deduplication.
The number of units utilized per user is measureable if a key that connects them together exists, one such key is a login but not limited to only that. In fact the more keys the better (up to a point...).
The essential key
With a user login as a key the deduplication of cookies is possible, that is connecting multiple device (and in some cases browsers on the same device) to a single key value in order to eliminate the painfully evident cookie inflation that exists.
While the user login is the obvious choice there are many other keys that can be used alone or in combination with the user login.
In the second image the obfuscated user id's make it clear that some users have be counted as 5 unique browsers using cookies, using the login to deduplicate user counting provides numbers that better reflect reality. In just a day a single user can be counted as many more... As seen in the example an overall 20% inflation becomes visible.
Cross device/browser usage
Taking as an example one of the top inflated users found in a study of deduplication, the count of unique browser might be 5 times or even more, but there are also users that use a single device when they log in.
Deduplication however has the "downside" (which is why some avoid resolving it) that the number of unique visitors over a given period will drop, however it does seem like a better way to rely on correct measurements when taking decisions rather than inflated metrics.
The more keys used to identify the same user over multiple devices the closer to the real number the metrics will be, but there is no exact prediction that can be made on the impact.
Since the reality is that users switch devices over the hours of the day having the proper solutions in place to deduplicate the unique browsers is increasingly becoming important.
With a select set of keys suddenly the web analyst can elevate term "unique user" in order to replace "unique browser" which is becoming more unreliable for every day that passes.
Key data can be harvested from such diverse things as targeted emails to users, even without logging in the user is known upon clicking on a link in the email as link parameters can be connected to address data from a CRM system.
For those using dodgy math, i.e. adding the number of unique browsers on their mobile site with the unique browsers from their desktop site to get a total unique browser count, the time has come to get a more accuracy into web analytics. Even responsive sites face the same issue.
The upside is that, despite cookie deletion inflating metrics, getting a user id in place and users to use it provides an easy route to deduplication of unique browsers. Additionally for websites based on subscription this is no brainer, the user login is in place so it is time to start reaping the benefits.
It is time to start increasing the accuracy in your web analytics.