Over the past several months, the battleground for my name, “Robert Elwell,” as a search term has been like trench warfare. While getting it from the top 30 to the top five in months prior, my personal website often suffered in terms of optimization due to a lack of external links and somewhat low keyword density for my name. Being on a .edu site, it had a slight advantage in some senses. I began joining a number of social networking services that were search engine visible so that I could do more with respect to giving my homepage external links that were relevant to “Robert Elwell.” Over the summer, I had at last achieved what I was aiming for—my site at number one for my name.

And then Google changed its PageRank algorithm, and it all went downhill from there. I lost two spots in the rankings to personal genealogy websites that seemed to have an edge only from a bag-of-words perspective—lots of “Robert, Robert, Robert” and “Elwell, Elwell, Elwell”, but not quite a website about Robert Elwell. This made me assume that Google was becoming more sensitive to textual content rather than external links. It seemed that the current state of Google’s ranking algorithm left something to be desired if you could beat out textually pertinent websites by simply repeating a keyword in body text with little to no markup. It had to be some kind of stepping stone to something else, because this seemed like an older Google, a less sustainable search strategy in the current Internet.

But this week, what I hoped would happen came true. Google’s affinity towards text has moved up another level. My personal website is at the top, with my contact page as a listed subpage. Where did my contact page come from? It was never the case that it showed up in the results—if anything, a funny, old paper I wrote on an African language would sometimes show up despite the fact that it was just a PDF document that mentioned my name somewhere at the top. Some time ago I turned my “contact” page into an hCard. An hCard is what is known as a type of microformat, which is an extension on top of HTML that keeps the markup legible to older applications but allows for the extra inclusion of semantic content within the markup. An hCard is a way of telling the machine, “This part is my name. Here is my address. That’s my phone number. And this here is my personal website.” The first part and last part are the most important. Google must now support this format. This is important, because microformats of all kinds are the most valid intermediary to the semantic web of the future. They are the easiest to implement for both the user and the programmer.

From this anecdote, if it applies elsewhere, we can make three assumptions about how Google has changed recently:

• It has better microformat support, and believes it to be an important and informative markup strategy.
• Google has included some reasoning about these microformats (enough to know that this person and this website should be associated due to an hCard).
• The biggest search engine is implementing semantic search at its own pace, meaning that it does not run the risk of being displaced by a young Turk with better semantic inferences any time soon.

If you are trying to optimize a page for a name, consider using an hCard. Furthermore, microformats in general have a promise in the future for informing Google in the same way title and headline tags do, and it will be important to optimize with this in mind.

This entry was posted on Wednesday, October 22nd, 2008 at 2:05 pm by relwell and is filed under Apogee Labs, Natural Search (SEO), PageRank, Website Design. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.