• March 5, 2009

    The Google Update Named Vince

    Last week, Aaron Wall published a blog post detailing a shift in Google algo results that to him appeared to favor brand names. Called one of the most important documents in recent SEO history, Aaron’s article spread like viral wildfire through the optimization community. In the article, Aaron suggested the algo shift was even more significant than the Florida update from November 2003. A couple days ago, Matt Cutts, search quality czar at Google posted a video gently disputing Aaron’s analysis.As Matt sees it, larger Brand name websites have a number of factors going for them that help them in Google rankings such as age, size of network, “trustworthiness”, time-on-site, and of course, oodles of incoming links. Earlier, Danny Sullivan reviewed Aaron’s piece and came to many of the same conclusions. He started a Sphinn discussion thread that turned into a lengthy conversation. Today, Michael Gray published a good post about the semantics of marketing a Google update to the search marketing community. “Dear Matt Cutts, What We Have is a Failure to Communicate”, in which Michael asks Matt to call a manatee a manatee as opposed to a SeaCow. He also asks Google to start issuing Weather Reports like Yahoo! does. Aaron’s Florida reference, combined with Matt’s video and Danny’s review made me think about that wonderful patent document Google filed in December 2003 and published on March 31, 2005. - "Information retrieval based on historical data" In June of 2005 I wrote a white-paper on the patent for StepForth Web Marketing Inc. titled, "Historic Data and Google Ranking...".  Any number of the points in from that patent filing, would have effects favoring the rankings of big-brand websites. Here are a few factors I noted in the white-paper drawn from that patent filing that might make big-brand websites rank stronger.  Add a newer slew of local, taxonomic and personalized factors (many of which can be found in subsequent patent filings or by scanning Bill Slawski's brain), and a closer approximation of how Google actually ranks individual documents might emerge. Google is likely considering all, some or a few of these historic record data-points in whole or in part while evaluating web documents for query rankings. On-site Elements:

    • An original copy of the document and a history of all changes to the document
    • Dates of changes made to the document
    • An evaluation of new content found on a document
    • An evaluation of the relevancy of new content to what was previously found on that document
    • Other factors indicating the change frequency and history associated with a document
    On-Site Links:
    • A copy of the original site structure and all changes made to that structure over time
    • Anchor text used to create internal links and any changes made over time to that anchor text
    • Addition of new documents to the site housed at the domain being evaluated
    • The overall relevancy of all pages linked within a domain being evaluated
    Incoming Links:
    • All historic information on documents linking to the document being evaluated
    • Changes to the content of documents linking to the document being evaluated
    • The age and number of links contained in a document linking to the document being evaluated
    • The anchor text used to phrase links contained in a document linking to the document being evaluated
    • Any changes to links contained in a document linking to a document being evaluated
    • The frequency of change to links in a document linking to a document being evaluated
    • An overall test of relevancy between documents linked to each other
    Elements found on pages or documents linking to the document being evaluated by Google:
    • The age of the document
    • The history of the domain or URL (including registration history)
    • An original copy of the document and all changes that have occurred in relation to that document over time
    • A trust/relevancy evaluation based on all documents linked to or from the linking document
    • The purpose of links found on the document
    • The frequency of changes to links found on the document
    • The anchor text used on links in the document
    • The number of links directed to the document being evaluated from the URL or domain in which the linking document is housed
    • The relevancy of anchor text associated with links found on other documents contained in the URL or domain of the linking document
    (As an aside; I might have figured out where that fellah "SEO Champion" gets his infamous “I know the 72-points in the Google Algorithm” line. The claim has boggled my mind for months now. It likely comes from that patent filing in which the SEO community was able to identify 63 unique factors covered in the document.) It stands to reason big-brand sites are getting a bump, especially if Google is working the kinks out of even a minor algo shift. When an algo shift takes place, Google does not throw out the useful or successful data-points that made up previous versions of their algos. They build on top of previous stuff. There is clearly some deep bounce happening in search results and, as we've seen in previous updates, big-brand websites are faring well. Give it a few weeks and we'll get a better idea of what's happening. Matt Cutts said he would outline more about the shift at the upcoming Pubcon South in Austin Texas.