Zero visits from keyword in Google Analytics

[RyanKent]

Hi Atul.

What you have is common in the SEO world: two different SEOs who offer opposing explanations on the same topic. In this case they cannot both be correct. Some things to consider:

The first article was written by Anil Batra. I have never heard of him, but that's OK! He's probably never heard of me either. He lists his credentials at the top of the page which seem satisfactory.

Anil's article was well presented and he offered a screenshot of his results along with a meaningful description of the test he performed to draw the conclusion he ultimately made. The article was written in April 2008.

The second article was written by Will Critchlow of Distilled. Will has written other articles I have read and I find him to be a credible source of SEO information.

Will's article is also well presented. Will specifically shared he contacted Avinash Kaushik from Google who would be considered an expert on the topic. Will's article was written in Jan 2009.

Personally I would choose to accept Will's response being that it is more recent and I offer higher credibility to his Google contact. BUT, I am also hesitant to discount anyone's ideas, especially when they are well presented such as Anil's article.

The great news: we can easily try a new test and find the answer! It's been over 2 1/2 years since the most recent article. I'll go ahead and try some tests and share the results.

[Devin_Anderson]

The SEOmoz article you linked was published more recently (2009) than the other blog post (2008), so if you have to choose I would go with that one, since Google is constantly updating.

It might help to look at it as the number of 'unique visits'. You got a visit for "business engagement in outsourcing", but that same visitor subsequently did another search and came to your site from a different keyword. As a result, you got several visits for different keywords, but only one unique visit. GA attributed this visit to a different keyword, so "business engagement in outsourcing" is listed as drawing 0 visits.

..if that makes any sense?

[seoug_2005]

Thanks Ryan. I would also go with the second article. But as you mentioned, its difficult to trust one source over another. I will be waiting for the results of your test.  Here is another link.

http://www.google.com/support/forum/p/Google%20Analytics/thread?tid=0d04370f123bc16c&hl=en  on the same topic, which suggests that  first keyword will show 1 visit, and the second keyword 0 visits. ( Again contradicting seomoz post )

[seoug_2005]

Thanks Devin.

If i understood you correctly, GA is attributing the visit to a different keyword ( other than "business engagement in outsourcing" )

[seoug_2005]

Thanks Devin.

If i understood you correctly, GA is attributing the visit to a different keyword ( other than "business engagement in outsourcing" )

[RyanKent]

My test results support the first article you shared.

I duplicated Anil's test. I searched three times for the same site only using a long tail phrase with four words. I altered the 4th term each time.

The first phrase showed "1" visit, the other two phrases appeared in the report but showed "0" visits.

The other results were combined. Even though I did bounce on my first visit the Bounce Rate showed 0%. My bounce on the first visit was immediate but the Average Time on Site was 1:12 which is clearly the average of my three visits.

You can easily perform this test and have the results the next day. Choose a key phrase where you perform well in SERPs but is unlikely to be used in a search. Alter the last term a repeat the steps in Anil's test. The next day, check your GA.

[seoug_2005]

Thanks a lot Ryan. So, we can discount seomozz post ?

[RyanKent]

Yes.

In this instance the testing I performed this week and the testing Anil performed in 2008 showed consistent results, and are in alignment with Google's explanation of how their sessions work.

A session begins when a user first visits your site. That session is maintained for 30 minutes even if the user leaves your site and returns to Google. Everything makes sense and synchs perfectly with the results.

If you wish to pursue this any further I can only suggest either repeating the test yourself, or contacting Will or Avinash concerning the prior article. I can only presume there was a misunderstanding in Will's article.

[willcritchlow]

Thanks for checking this out and not just taking it at face value guys. I have to admit I didn't test it in detail. It seemed plausible because GA is generally last touch, but I'm glad to see someone testing it. I have updated the post.

[RyanKent]

Thanks for following up Will.