Sorted by Date - Sort by Ratings
16 August 2006The rebranded Google Webmaster Central (formerly Google Sitemaps) is set to finally resolve Google’s canonical issues. For some time Google has had problems with the www.domain.com/.. and domain.com/.. versions of site URLs (detecting if they are one and the same site, distributing linking popularity properly). Many websites have incoming links pointing to both the www.domain.com and domain.com URL versions. The problem was exacerbated when the internal linking structure uses relative URLs (links like /article.htm opposed to links like www.domain.com/article.htm).
Recently Google added a new feature in their Webmaster Tools section (former Google Sitemaps), where every webmaster could set a preference for www. or non-www. URLs on every validated domain. Instead of guessing, you simply tell Google to use the www. or non-www. URL versions for every site of yours. This helps properly aggregating PageRank and showing the proper URLs within the SERPs.
In addition, Google offers a variety of other goodies like giving you crawling stats, problems, penalization info, reinclusion requests for penalized sites etc. I’ve been using these features for some time and they are getting better and better and are a must for every webmaster.
It is interesting to know what kind of stats does Google keep for every site in its index. Judging from their Webmaster Tools section, Google keeps ranking information (for which keywords does your site rank and at what positions). As I have written numerous times, ranking for as many keywords as possible (even non-competitive) gives you a ranking boost for more competitive ones (content is king).
Another great stat that Google shows is the most common words used within content on your sites and in the anchor text of links pointing to your site. I have noticed that Google has associated the anchor text keywords of some of my sites with words that do NOT appear in the anchor text, but words close to the links. So, yes, Google looks at text around links and probably uses it for scoring purposes.
The keywords associated with your content and links do not appear for all sites. I’ve noticed that when I added my sites using the www.domain.com version as the URL name of the site, I didn’t get these stats. When I set a domain preference for the www.domain.com versions, Google added the domain.com versions to the list of my sites and now shows these keyword associations when I click on the non-www. versions. I think that’s a glitch, but anyway, Google shows these vital stats when you have your sites added as non-www. domains.
I highly recommend the Google’s Webmaster Central Tools to all webmasters, especially the domain preference feature.
17 May 2006Matt Cutts has been sharing a great deal of inside information on Google’s ranking algorithm on his blog and through an interview with Mike Grehan.
You can read the full scoop on the Bigdaddy update at Matt’s blog here: http://www.mattcutts.com/blog/indexing-timeline/
The most interesting point from the post is that Google has changed the way they calculate crawling priority. They seem to factor how trusty are the incoming and outgoing links of a site when determining how much of it to crawl. Matt gives great examples. Basically, for a new site you need to have quality incoming and outgoing links. If you just swap reciprocals with low quality sites, you get both low quality incoming and outgoing links which is a no-no.
Mike Grehan has done a great interview with Matt Cutts. You can read the gist of it here: http://www.clickz.com/experts/search/results/article.php/3605961. There are two links to the full audio interview at the bottom of the article. This is a must read / hear.
Matt speaks about the importance of PageRank, reciprocal, triangular etc. links and the (non-)existence of the alleged Sandbox effect on new sites. Matt debunks the Sandbox theory that I haven’t been a fan of. He says: "I think a lot of what's perceived as the sandbox is artefacts where, in our indexing, some data may take longer to be computed than other data."
The take home message from all this information is: stop using reciprocal links, triangular links or whatever easy link exchange scheme; don’t buy links (or don’t get caught buying links); link out to great sites; focus on user experience
11 April 2006Google has just been granted a patent called "Adaptive computation of ranking", which deals with faster PageRank approximations.
Another patent called "Methods for ranking nodes in large directed graphs" deals with another variant of PageRank that I wrote about some time ago (I called it HostRank, the article is no longer available).
Both patents are follow-ups on papers by Taher Haveliwala (the author of the Topic Sensitive PageRank papers) who now works for Google. I wonder, if Google has filed for a patent on Topic Sensitive PageRanks.
I think the second patent has better real world applications and matches perfectly with Google's distributed architecture.
22 January 2006You’ve probably heard about the new SEO contest revolving about ranking for the keyphrase v7ndotcom elursrebmem. Let me ask you a question: what is the most interesting thing about this new SEO contest? No, it is not the $4000 1st place prize.
The target keyphrease v7ndotcom elursrebmem is pretty darn interesting. Why? Because it is a topical query (hot topic). Topical queries are queries that suddently start to get searched for way more than before. I bet there are more than enough people who regularly check the top rankings of the contest query, which in the eyes of Google is a sudden increase (from zero searches). That IS a topical query – hot topic. Also when the number of documents containing a query suddenly increase (as in our case) – this indicates a topical query.
So what? Google’s patent Information Retrieval Based On Historical Data has some interesting information about topical queries influencing the ranking of documents. The 2 major points are:
1. Google may score documents (sites, pages) associated with topical query higher!
2. Google’s anti-link spam detection is less sensitive to topical queries!
How many people SEO for topical queries? Very few. Webmasters are fighting mostly for non-topical queries. In my opinion, all the post-seo-contest-analysis articles that are going to pop are going to be worthless and won’t apply to the real world (mostly non-topical queries).
But that is not the end of the story. Based on the above information from Google’s paper, we can actually take advantage of topical queries - simply by regularly including content rich in topical queries! Consider this blog entry of mine. I used the topical keywords in the title. Also, all my blog and article pages list the latest blog entries (they list the title which includes the topical query). According to the patent, some or all of my pages can get higher rankings. This in turn according to the patent can further push the rankings up by getting my site ranked in the top 30 for more queries.
According to the patent, it is a very good idea to regularly include content about current hot topics. How to do it?
- write about current hot topics in your industry in articles / blogs
- include a list of the latest articles / blog entries on as many pages as you can
- include topical keywords in the titles
- at times when your site is associated with a hot topic, you can be slightly more aggressive with link building
Here are a few more examples of SEO specific topics that were hot no so long time ago – Google’s Florida Update, Google Analytics, Yahoo Publisher Network etc.