Site   Web

April 6, 2011

Google Analytics Misrepresenting its Own Search Traffic Numbers – A SPN Exclusive

Google logo

I came across a curious issue with Google analytics the other day. I had just posted a blog entitled “Online marketing explained with reference to cows” at www.siteprebuilder.com/content/online-marketing-explained-reference-cows.

I could see that the post was generating a lot of traffic and wanted to know where it was coming from, so I went to Google.com and typed in “online marketing explained” to see if the post had already been indexed and if it was ranking highly for that keyphrase.

Sure enough, the blog post had been indexed and it was showing up in the first or second spot in Google, ranked amongst the posts indexed in the last week. Awesome! Now I wanted to know if all the traffic was coming from that keyphrase or were there others, so I headed to analytics.

As it turned out, only one visit had come from that keyword in Google. Instead, the influx of traffic was coming from social networks like Stumbleupon and reddit and not organic search.

Here’s the strange part:

The visit recorded by Google analytics for the keyword search “online marketing explained” was credited to my service provider and not Google search. But obviously, I never actually clicked through to my page because I already knew what the post was about.

I tried the experiment again, and sure enough, Google analytics was registering a visit from Google search despite there being no actual visit.

In other words, a user only has to search for a term that results in (I assume) a first page result for your site and Google will lead you to believe that it sent a visitor to your page. It’s treating a search result as an actual visit, which seems really underhanded.

The upshot of all of this is that webmasters who rely on Google analytics to determine how well their online marketing campaigns are performing, or how well their SEO efforts are paying off, are not being given an accurate picture of the traffic they are receiving.

I would have to spend more time analyzing the way the visit hits work in analytics and Google search in order to determine precisely by how much Google is inflating its search result visit numbers.

At a rough guess, if Google records a visit for every first page result for your website or webpages, and assuming a rough actual click-through rate of around 20% (which is very generous), then Google is over-representing its traffic by a factor of 5 for each given keyword.

This means that your website could be getting 5 times less organic search traffic than you are being led to believe.

The situation is made far worse when you consider that the analytics figures reported by Google are then used by webmasters to determine their traffic amount and how much they can charge advertisers.

Advertisers are then paying cash for traffic that doesn’t exist. So Google’s inflated numbers effectively lead to a situation in which, mislead web businesses end up overcharging their advertisers based on effectively fictitious numbers.

Worse, by inflating their perceived search traffic, Google gains an unfair advantage over its competing search engines, because everyone wants to go with the search engine that brings in the “most traffic”.

Now, maybe they have a good explanation (I will ask them), but it seems to me that this is a subtle, but deliberate, way to boost the perceived importance of Google search for webmasters using analytics (which is a lot of webmasters).


David Mercer, http://www.siteprebuilder.com, is one of the most experienced technical writers in the world today, having contributed to books that have have been translated into virtually every major language in the world.

2 Responses to “Google Analytics Misrepresenting its Own Search Traffic Numbers – A SPN Exclusive

    avatar John Stocks says:

    Personally I don’t use Google for tracking my traffic as I always had weird figures from them., Statcounter seemed a lot more accurate.

    avatar Ed Rude says:

    Server records are usually the most reliable records. Many hosting companies (unfortunately not all) even sort your webtraffic by search term and search engine.
    This helps us know where the traffic coems from, so that we can adjust the content to the expressed needs (what the search terms were).
    Other than that, the best use of analytics is to follow the
    pathway(s) that get used to navigate the overall site. This helps us in other ways.
    The problem shows up in definition fo what a website visitor is – some definitions are “refreshed browser,” which apparently is what Google was using, while others demand an actual click (click counters are particularly useful for CTRs).
    But when all the fine measurements of even the Sacerd Cow of Google Analytics goes awry, and everyone is stuck wondering why the results they got were the ones they geot, then to quote David Mercer’s blog-post, which started off this discussion, “Don’t have a cow, man.”

Submit a Comment

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






You may use these HTML tags and attributes: <a href="" title="" rel=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

Please leave these two fields as-is:

Protected by Invisible Defender. Showed 403 to 3,861,817 bad guys.

css.php