Follow

Why are the numbers in Cxense Advertising statistics and my third-party tracker different?

Statistical discrepancies between Advertising and third party systems are occasionally seen when it comes to counting clicks, impressions, etc.

There are a number of reasons why two systems measure events differently.  We will describe a few examples below.

Navigation is aborted

When navigation is aborted, either by the user or through an error in the javascript, the tracker that measures earlier will always measure more events than the tracker that measures later.

Non-unique URLs

If the URL to the later tracker is a URL that's been encountered previously, the browser may retrieve the page contents from the browser's local cache or from a proxy cache configured on the user's internal network, instead of actually calling to the remote site.  As a result, the later tracker will never know that the page was requested again, and will not register the event.


Malformed click tracking URLs

There may be a different route for handling the click that does not pass through the inner tracker.  For instance, there may be some ads for an advertiser that do not have the correct click tracker URL.

URL interception

Sometimes click tracking URLs are intercepted by the browser, so that the outer tracker (Advertising) will record the click, but the click tracker URL is rejected.  Sources can be ad blockers, firewalls, character-encoding issues, URL encoding issues, etc.

Different tracker scopes

Sometimes the two tracking systems are not in fact monitoring the same set of pages, so that extra clicks and impressions can be recorded by one tracker that can be missed by another.


Different time periods

Similarly, two trackers may be looking at a similar but not identical period of time.  This can be due to a configuration problem, time zone difference (i.e. local time zone versus universal time), or differing definitions of what defines the start of a day or week.

Different definitions

Often, two systems are configured differently as far as what should actually be counted.  Examples include whether or not to include robot traffic, or whether impressions are counted or not when javascript is disabled.

Conclusions

As you can see, it is often that discrepancies with third-party tracking systems are a result of the configuration of the web server or of the third party system, which is rarely possible for Cxense to diagnose.

We will stand by our numbers.  Cxense can provide a detailed log of events that were received by Advertising -- everywhere a page impression occurred, where it was clicked, who made the click, etc.  While we cannot perform an in-depth analysis of a customer's website to determine why the discrepancies occur, this report can be used by our customers to verify against what they have logged in their own tracker.

 
Have more questions? Submit a request

Comments

Powered by Zendesk