Latest Google Changes: Freshness, anchor text, image search

1. 78

   Greekgeekposted 12 years ago

   I keep up with Google news fairly regularly to see what's coming down the pipe. I still feel vague self-righteousness over raising the pre-Panda alarm, only to be scoffed at for suggesting that Google might start ranking sites by quality of content across a domain, not just on a page-by-page basis.
   
   Anyway, a new Google blog post is up. And I'm about to start spouting Google gobbledygook, so run away if phrases like "anchor text" and "search relevance" make your eyeballs spin around in your head.
   
   This Searchengineland.com news article does a good job of dissecting some of the key parts of Google's latest search algo tweaks.
   
   The ones that may impact us are:
   
   (a) Anchor text evaluation. Traditionally, search engines look at anchor text, the words in a clickable link, to help determine how relevant the target page is to a search query. So, for example, if a lot of links pointing to a webpage say "penguins," the target page is assumed to be relevant to the search "pengins." This is how a great many people annoyed with Rick Santorum managed to make searches for his name point to an obscene, fake definition a few years back.
   
   Google doesn't say how it's changed the way it evaluates anchor text. But since so many people game the system, trying to rank for particular phrases by putting them in anchor text, I suspect this means exact-phrase-matches are being devalued as a ranking signal.
   
   (b) Freshness evaluation. Whatever Google's been using to sort news articles so that the most recent ones bubble to the top in the news section, they've now added that algorithm tweak to regular search results as well. I have a hunch this is going to give blogs even more authority, and cause older articles to get pushed down in search rankings more often. Fooey. (I'm also sure Google is smart enough to know that searches on topical subjects like the Super Bowl are more likely to expect a recent answer than searches for non-time-sensitive subjects like "fall of the Roman Empire.")
   
   (c) Image search. Google's tweaked the image search algorithm so that images that are more relevant to the search query take precedence over images on pages that are considered "high quality." This is good for us, because it should counteract Panda a bit. In plain English: it doesn't matter if the article is poorly-written, if it has the perfect image for an image search. The only problem here is that Hubpages image files are not well-optimized to show up in image search. (Stop changing my image filenames to gibberish, Hubpages! Argh! At LEAST give me manual control of alt-tags separate from captions!)

   replyreport
   1. 60

      paradigmsearchposted 12 years agoin reply to this

      I have always thought your posts to be top quality and always read them. I have much appreciated the previous info you have given us. And I very much doubt I am alone in this regard.

      replyreport
2. 77

   Shadesbreathposted 12 years ago

   Good post.

   replyreport
3. 80

   Aficionadaposted 12 years ago

   I was hoping to see here something about why my traffic is higher today.  If these tweaks are forthcoming but haven't arrived yet, then they can't be the reason.  I don't think any of these tweaks would affect me in a positive way, in any case.

   replyreport
   1. 78

      Healthy Pursuitsposted 12 years agoin reply to this

      Mine, too. It tripled. While that's a joy to see, it doesn't seem to be based on anything I've done. So am expecting reality again tomorrow. Anybody know why it happened?

      replyreport
      1. 80

         Aficionadaposted 12 years agoin reply to this

         Today is still going well so far!  I hope the same for you.  Yesterday's earnings looked great too - my best ever.

         replyreport
         1. 78

            Healthy Pursuitsposted 12 years agoin reply to this

            Yes, me too! Well, my numbers are still pathetic in comparison to someone who's had good traffic for a while, but it's putting a smile on my face. I surely do wish I knew what happened, though, so I could continue it.
            
            Is it possible that I would see my first check this month? Oh, joy! Oh, validation! Oh...wait a minute. It may be the gods messing with our heads.

            replyreport
4. 84

   Linda Blissposted 12 years ago

   Thanks for this update - really very interesting. I agree with you about images on Hubpages. On my blog I have full control over everything from filename, captions to alt tags for all the images I upload and it really does make a difference overall.
   
   Fingers crossed all these updates will bring our traffic up )

   replyreport
5. 80

   Les Trois Chenesposted 12 years ago

   Thanks for this understandable information GreekGeek. Wonder why only some of us in HP has been affected though? Squidoo doesn't seem to have been hit.

   replyreport
   1. 78

      Greekgeekposted 12 years agoin reply to this

      Hmm. Google's blog post was dated April 3, so it may take a few weeks to detect any traffic changes from the most recent algorithm tweaks. Also, remember,  Panda may be judging subdomains on Hubpages somewhat as websites in their own right, so the fortunes of individual authors may not be tied quite as absolutely to the traffic of the whole site.
      
      Looking at Quantcast.com's traffic figures, I'm showing a drop at Hubpages around March 30th, but both sites are pulling in about the same traffic at this point (which, honestly, I think they should be; they're similar in design, and have a similar spread of conscientious, articulate, and meaningful authors lost in a sea of get-rich-quick posters writing low-quality junk that doesn't deserve high rankings).
      
      That means Hubpages got hit by the Panda 3.4 rollout. This rollout was twofold:
      
      -- Google periodically refreshes what I call the Panda rating on each domain: it checks the amount of what it defines as quality to low-quality on a website, gives that website a ranking (its "Panda rating,") and then applies that Panda rating as one factor (a strong one) in ranking all pages on that domain. This "Panda rating" remains the same until the next time the Panda algorithm is rerun. The criteria for Panda may remain unchanged during a new Panda run; it's just that the ratio of quality to poor-quality content will have changed on each website in the meantime.
      
      -- But sometimes, Google adjusts the way the Panda rating is calculated. Panda 3.4 on March 23(?) was such an adjustment, reflecting further tweaks Google alluded to back on Feb. 24:
      
      
      
      That's clearly a description of Panda (Google's code name for its 2011 algorithm that focuses on rating the quality of content on domains). 
      
      Squidoo articles on average tend to be longer, with a lot of different capsules; in fact they require at least 3 PLUS an introduction and encourage 5-6. The headers, subheaders, image alt-names, and various blank spots encourage writers to structure articles, which may help a little. Also, Squidoo articles tend to link out to sources more than hubs, since Hubpages has been stricter about links being "overly promotional." This edge should be lost now that Hubpages is being a little more flexible about links out. Squidoo's also had a recent rollout of yet more strict spam filters and duplicate content filters that are actually catching people with false positives, and as usual their HQ has been hammering home "no duplicate content, originality, credit all images, plagiarism" in a way that can be rather aggravating, vague, and saccharine, but has the effect of scaring writers into erring on the side of caution. The proof is in the pudding, though: how much duplicate, spammy, and low-quality content  do the filters on the two sites actually catch, and how quickly do staff remove  bad content flagged by members? I don't know.
      
      Back to Panda 3.4, the changes in Hubpages traffic may simply have been the usual ups and downs of a Panda reranking, although it's odd that Squidoo seldom seems to show traffic wobbles from Panda, whereas Hubpages does. (Honestly, this is why I came back to Hubpages last year; I want to know WHY! If I can understand that, then I can understand how Google works. I can't understand it if I don't get to know Hubpages very well for purposes of comparison and analysis.)
      
      One other thing that might be a factor is something Google mentioned in a list of 40 algorithm changes at the end of February: freshness of content. Squidoo lensrank drops precipitously on articles that haven't been updated in the past month or so. I hate hate HATE this: perfectly good content sinks rapidly if you don't keep fiddling with it. It's one area in which I much prefer Hubpages! But perhaps that freshness gerbil wheel gives Squidoo an advantage in Google's eyes: there's probably a lot more articles being updated over there, since they have to be.
      
      On the other hand, since traffic between Squidoo and Hubpages is comparable (according to Quantcast), it's hard to say that one is faring better than the other from Google. I haven't heard active, good members on Squidoo reporting disappointing traffic drops in quite the same way as Hubbers, but maybe Squidoo just always had lower traffic and made the best of it, whereas Hubpages members are dealing with a drop back to Squidoo levels of traffic. (Also, the lensrank algorithm rewards the most successful lenses no matter how much traffic there is, whereas Hubpages rewards traffic, impressions and clicks on a one-to-one ratio, making the drop more painful for Hubbers).

      replyreport
      1. 81

         sabrebIadeposted 12 years agoin reply to this

         
         
         
         I have noticed that Squidoo articles have been getting ranked faster and higher than HP for a while now.
         At least in my cases.

         replyreport
      2. 80

         Aficionadaposted 12 years agoin reply to this

         
         Is this something I have missed (not reading staff announcements as much in detail as I did for a while)?  I don't quite follow what you mean about HP's flexibility.
         
         Aside from that, very helpful, useful, information.  Thank you!

         replyreport
6. 87

   WryLiltposted 12 years ago

   Thanks. I never bothered with "correct keyword" anchor text. So glad to know it might not be worth much. Wondering how much the "fresh content" boost will help our older content if we keep it up.

   replyreport
7. 88

   lobobrandonposted 12 years ago

   Hmm I always check out your posts and thanks for this Off now going to read the indisearch blog post.

   replyreport
8. 74

   sandrabusbyposted 12 years ago

   Thank you so much for this information -- at least some of which I am coming to understand.  I have only been writing on the internet for less than two months and, so far, hubpages is the only site I write on.  My current interest is in building traffic to my sites.  I read from other hubbers that posting on RedGage is a legitimate way to boost traffic.  I would like your opinion on this.  Thanks again.

   replyreport
   1. 78

      Greekgeekposted 12 years agoin reply to this

      I just tried to give you an answer and got Hubpages' note that my reply was too long and that i should make a hub about it. So, well, I did.
      
      I'm not sure about my answer, but maybe my thoughts can help you make up your mind. I wish I could give you an authoritative "Yes" or "No," but I know enough to know when I don't know!
      
      And no, if a hub is unpublished, Google won't be able to crawl/see/index the content.

      replyreport
9. 80

   Les Trois Chenesposted 12 years ago

   GreekGeek, I wonder if Panda is taking into consideration unpublished hubs? I have lots as I like to strike while I feel enthusiastic and then gradually collect information and pictures to add to them. If these were counted, then about one third of my hubs would be considered substandard.

   replyreport
   1. 61

      glutenallergyposted 12 years agoin reply to this

      If your hubs are unpublished, I'm going to assume that they can't be seen by anyone but you?  I don't see how Google would be able to find unpublished hubs, since they are not ready for prime time (so to speak), which is why they remain unpublished.

      replyreport
10. 68

    Rosie2010posted 12 years ago

    Thanks, Greekgeek.  This is very helpful.  Some of my old hubs that used to do well, are getting reduced traffic.  Better go and refresh them then.  Thanks again.

    replyreport
11. 89

    jellygatorposted 12 years ago

    I don't normally keep up with Google changes because quite frankly, I don't understand a lot of the lingo.
    
    But I've noticed that despite interlinking my pages, using outbound links more often, and promoting on social media, my HP views on all hubs have tanked (1-3 a day total, sometimes as few as 8 total on a day after I've published a new one) while Google counts have risen to 15-20.
    
    When I started a few months ago, I saw 14-30 a day from HP and felt lucky if even 1 Google came in
    
    Whatever Google's doing seems to be making my articles more visible as time progresses.

    replyreport