seo

I wrote before about the rel nofollow monstrosity. It seems Google guys did finally realize something is terribly wrong with their invention. What Google anti-SPAM engineer Matt Cutts told us the other day is this:

Instead of preventing a “PageRank leak” (read: pathologically saving every drop of the inbound linking credit that a site gets) as can be seen in this “rel NoFollow” article, Google decided to simply exclude the PageRank which is nofollowed from the equation, effectively, reducing the size of PageRank that circulates through the site (and the web). This will, by the way, introduce some global level disturbances in the page-rank matrix, one of the major algorithms in the google rankings (even thought the number of nofollowed links is less than 3 percent, as they mostly come from big sites).

Is this change good or bad?

I really like to think that Google realized how bad their invention was, and what consequences it had in terms of search results which were more and more originating from a small number of big sites (which, by accident? almost all use nofollow on all external links). Google’s responsibility is immense, as they become in a way a public service company. (I’ve heard that few Google data centers are even on the list of US military priority defense targets in the case of war.) And as a public service company (providing most people with access to world’s information), they do have responsibility to present users with good and diverse information, which was happening less due to use of nofollow.

So is the change good or bad? I think it is very good, as now all those big sites will not just absorb all the pagerank from small sites that are linking to them (think of Wikipedia and most bookmarking sites), but will also lose large chunk of pagerank through the external links (which BTW make those sites better and more information rich and on which some of them are 100% based, like social bookmarking sites). This change will result in big sites going down in rankings, and small sites going up again. You will see again your homepage when you google for it, and not a stumbleupon page that just links to your site and has no useful content.

Note however that above change was only told by Google, while webmasters need yet to experiment and see how exactly will things change, and what exactly was the nature of the change.

Again, here are few links where you can read extensive background on the nofollow, pagerank distribution, and its effects.

What should I do with my website(s)?

If you’ve played with nofollow on internal links on your site in order to do some pagerank sculpting and to prioritize important pages, than this change will affect negatively on your site. I yet have to think about and decide what to do with few sites where I used nofollow on internal links. I am almost sure, I should do the navigation restructuring, and push important pages up in the navigation, and accessible from homepage and from many other pages, while pushing unimportant pages down into deeper levels of the navigation. Also, in pages where I have ‘login’, ‘register’, and similar, I think I am going to remove nofollow, and maybe add few useful links to those pages which don’t have many links, and have these non-important links. In other words, I am going to do on all sites what I was doing already most of the time, as that was, after all, the preferred way of sculpting the pagerank, or should I be more accurate, crawl prioritization:

http://www.google.com/support/webmasters/bin/answer.py?hl=en&answer=96569

Alexa.com toolbar is a web traffic analysis program integrated into your browser. I always knew that services like Alexa keep track of what websites users visit, time they spend there, and similar general statistics. What I didn’t know until today is that they collect very personal information about you, and they also publish some of it. For example, private pages and links that you visit may appear on Google as Alexa may place them in their site analysis section (sites linking to… and similar). Imagine you write a private blog post where you link to some site you want only your friend to know about, and then few days later you find these links on Alexa site and indexed by Google thanks to Alexa not respecting your privacy! Also, did you know that:

… When the Alexa Toolbar is turned on, Alexa collects and stores
information about the Web pages that you view, the products you purchase
online, and the data you enter in online forms and search fields.

However, please note: Alexa does not collect information entered into
forms while you are on a secure (https://) site, so we do not record
your credit card information when you are making a purchase online. You
should never enter your credit card information on a non-secure site…

Just think about this the next time you browse the net with the toolbar turned on.

So today I typed a very common word into the search engine — word and, and I wanted to see who ranks the first for such a common word, and to see why would someone be the first for the word and.

I was very pleasantly surprised as I discovered what seems to be a really very good site, and useful site, that among other things also talks about word and. What site would talk about word and? A site about grammar!

Not being native english speaker, I appreciate such a good online resources.

Anyhow, you will notice that word ‘and’ in its title doesn’t have … how would I say it… a fundamental meaning. It is there simply to connect two words, and title may as well be as good without it:

Guide to Grammar. Guide to Writing.

Guide to Grammar in Writing.

or some other variation. Words grammar and writing often occur near each other without word and. For example, ‘proper use of grammar in writing…’ (btw. I am aware there may be a few grammar mistakes inmy posts — I didn’t practice English grammar since high school).

this is the top site.

However, you will notice that next few results fundamentally require word ‘and’ in their titles.
Continue reading »

If you get a comment that sounds meaningful when isolated, but when placed in context of your blog post doesn’t seem quite right or informative, do a search for exact phrase from that comment. Many automated bots post ‘meaningful’ general comments. If the same comment is already indexed by search engine somewhere else, its a SPAM. Delete it.

This simplified model is a nice introduction: Vector Space Model.

More food for thought: similarity measures and original Google architecture.