What Did Google Turn Off in Panda 3.4?


PageRank (Photo credit: Wikipedia)

Since the activation of the Google 3.4 algorithm filter last month, SEOs all over the Internet have been speculating on the ranking factor(s) that Google claims to have turned off.

The most popular idea is that ofย  Page Rank. Google has been telling webmasters for years that Page Rank isn’t the be all and end all, and we knew that, but it had to mean something or Google wouldn’t have been so adamant about it being no big deal, right? Only Google knows for sure, but this is one of the factors that many SEOs feel strongly about having been turned off in the algorithm in regard to link analysis. Hmm…

There are other factors that some have potential, and Bill Slawski of SEObytheSea.com thinks there may be twelve, page rank being only one of them. Here are the 11 others:

  • Local Inter-connectivity: This is hard to parse, but it means that if there are a set of links ranked for search, Google also considered the relationship between links when arranging page in the SERPs (search engine results pages). Links that had some relationship to others were given higher ranking. I don’t think factor, was done in,ย  based upon my “SEO Bounce Rate” testing. Two of the pages discussing the ranking results are still showing up in the SERPs on pages one. They’re interconnected and still ranking well, even since the Panda 3.4 filter run last month.
  • Related sites:You know the links at the bottom of a SERPs page for related sites? Slawski thinks that Google found a better way of showing these and may have turned off the linking analysis for this. Not sure. See if the related links seem better to you over time.
  • Adaptive page rank: Google patented a way of speeding up the page rank calculation. Slawski wonders if the patent is still needed, and if perhaps they have a better way of doing this making the original speedy page rank calculation unnecessary.
  • Cross Language Retrieval: This means that Google used anchor text on a page in English, for example, to understand what a page was about in German. They probably don’t need to do this anymore. Let’s figure that Google speaks a Ghodly number of languages.
  • Link-Based Clustering: Google used to (or maybe still does, this is all speculation) consider a cluster of links to determine rankings. So, for example, you had a bunch of pages that all pointed to one another, and this helped figure out which ones were the most important. Now, it’s more likely that Google takes the content of the page into consideration when ranking rather than the linking clusters. This makes sense, especially since the LSI (latent semantic indexing) is getting better and better all the time.
  • Anchor text indexing: Though Google used to use anchor text to establish reputation for a page, this isn’t the only method, so though this core idea may still be used, Google has gotten much more sophisticated than to use this method alone.
  • Historical Linking Data: While Google probably still analyzes the age, consistency and other link factors of links they’re analyzing over time, it’s probably doing this much better, using a different method
  • Linking Weight Based on Page Placement:Depending on where a link is on a page had and probably still has something to do with how Google views its importance. Prominence has traditionally played a big role in SEO and probably still does to a certain extent. Google has just gotten much better at figuring out what’s important and what’s not.
  • Reasonable Surfer Link Values: This “Reasonable Surfer” patent made all links on a page equal in value. This may no longer be the case.
  • Links Between Affiliated Sites: This is a big one and almost certainly changed. If you have one site that points to another, the link value is low, and Google knows. Even if you put your blog on a different “C” class server, you probably won’t be hiding the fact that the site on the other server is related to you and these links just won’t count very much, if at all.
  • Relevance Between Linked Pages: Let’s face it, some pages are more important than others. So, a link to a privacy policy, for example, wouldn’t be as important as a link to another content page on your site. This may have gone away — again, because now Google has a better way.

So, listen, Google is constantly changing and it’s getting smarter and smarter all the time. We know that Panda is a self-improving filter, too, so your SEO strategy will continue to change over time. That’s why it makes it so important to stay informed, if you want to use SEO as part of your overall traffic strategy. (You’ll notice I said “part” because you should NEVER depend on one source of traffic at a time… ever.)

And I’m not an engineer and I don’t study patents like Bill Slawski does. He’s a very smart guy, but his writing is sometimes cryptic to the un-trained, like me for example. ๐Ÿ™‚ I mean, I think I have my head around this stuff, but NOBODY, not Slawski, not me, or any other SEO on the planet knows what Google does exactly. All we can do is study what happens and make plans based on effect.

If you’re an engineer, you’d be smart to hop over to Slawski’s blog and read: “12 Google Link Analysis Methods That Might Have Changed,” and keep me honest. ๐Ÿ™‚

The big takeaways are these:

  1. Page rank doesn’t matter. Quit basing your importance on that and look at your traffic. That’s what matters.
  2. Linking to yourself is OK, but just don’t over do.
  3. You’ll never be 100% certain that what you think is happening at Google is really happening.

Take what you hear, see, and feel and use that to your best advantage. After all, this SEO world we live in is just a big game. Some players are just better than others.


Enhanced by Zemanta