Most Google Analytics users will be familiar with these entries in the Traffic Sources report: google / organic, google / cpc, etc. Since GA groups all Google visits together. But wouldnt you like to see the actual coutry domains: google.es / organic, google.com / organic, etc. Here’s how. Go to Filter Manager and Add new filter and then fill in the form as follows:

NOTE: The expression in the Extract A field is: ^http://([^/]*)

Here is a nougat of information that surprisingly I havent heard spoken of much by SEO people. The length of time that a certain keyword has been associated with a certain page is a very important factor in ranking. For example: Suppose your competitors have pages optimised for the keyword “monkey feed” and then you come along and launch a page for “Indian monkey feed”. Soon you find you rank better then them for this keyword. So your competitors follow suit and change their pages to include “Indian monkey feed” just like yours. Who ranks better? Answer: You do. You will maintain your dominance for Indian monkey feed for some time dispite your competitors having better pages. Its not about the age of the page or the domain. Its simply about who was the first to include that keyword in their page. So the moral of the story is that when you launch a page, think about all the possible secondary keywords, like “Indian” that you may also want to rank well for - espacially the ones the competition haven’t yet thought of.

Is it better to have many small pages or one large page to cover a range of keywords? This is a discussion I have had many times. In fact, inexperienced webmasters tend to dedicate a whole page to a keyword. This has advantages in that you can put the kw in the page title. But here is evidence that such flagrant over-optimising of your site in not needed and can even be damaging. You can achieve better results by having larger pages that cover a range of related keywords. The key is proximity of the secondary keywords. What do I mean by this? Look at the screenshot. The primary keyword of the page is “Kölner Karneval”, but the secondary keywords are things like “2010″. The result: the page works well for decent keywords like “Karneval 2010″. Why? Because the secondary keyword appears three times in close proximity, and is not distributed around the page. This is interpreted by Google as natural language and can be contrasted with a scattergun approach to kw distribution, which does not work. Believe me, this natural language approach does works. I have many examples and here I give you one. Have a nice day!

Is Google becoming intelligent? Well here’s an example of seemingly intelligent behaviour. The interpretation of abbreviations. Look at the screenshot. I typed in “vuelos baratos a ny” (cheap flights to ny). Google, being Google, correctly interpreteted the ny to mean “Nueva york”. Note the embolded keywords in the results. It reminded me of IBM’s chess wizard Deep Blue beating Kasparov back in 1996. The debate was whether a good database of stored human behaviour processed by a super-fast computer could be more “intelligent” than a human. This is what I think is heppening at Google. Many thousands of people every day use abbreviations and since Google records what people click on it can ”learn” from us what the appropriate connections are. Conclusion: Google is not intelligent, its just good at emulation.

Unfortunately, I’m one of those people who have mates who send jokes and other curiosities by email and a while ago I recieved a photo of Toronto’s CN Tower. In fact it was a sexist joke because in the foreground was a young well endowed girl. What has this got to do with universal search? Well it seems many people saw this image and linked to the page that contains it with the words “CN Tower” in the anchor text. The result you can see in the screenshot. Or try typing CN Tower in Google.es. I’m sure this is not what Google intended would show up for this keyword but it goes to show just how “algorythmic” universal search really is.

At last search engines have started to allow webmasters to segment their pages into indexable and nonindexable parts. Up to now it was only possible to set a whole page as indexable or not. Yahoo is the first to do this, lets hope Google follows. 

First Google gave preferential treatment and more screen coverage to algorythimically determined “authority sites”. Fair enough. But now they have gone a step further and given some sites a Google map pointer aswell. Look at all the screenspace devoted to just one site. So much for variety in search results.