Google.com SERP Changes - December 2008

[potentialgeek]

< continued from [webmasterworld.com...] >

Does Google refresh pages linked from Sitelinks more often? I think it already does, but if it doesn't it should.

p/g

[edited by: tedster at 3:26 am (utc) on Dec. 3, 2008]

[tedster]

That might be a chicken or the egg question - since Google already ups the spidering for any url it measures as important. Given how widely awarded sitelinks are these days, i'm pretty sure it's still dependent on the domain, and whether the sitelinks are being displayed for something more than navigational queries.

---

It seems to me that the SERPs have been almost uncommonly quiet for the past few days. I guess with the US holiday and so on, that's to be expected. I'm not sure I could handle any heavy drama until after the New Year - which is when I think we'll see more signifiant churn again.

[JS_Harris]

Not directly, no.

A site that has earned sitelink status is old enough and doing the right things to merit the extra attention which indirectly leads to new links being crawled from those pages BUT the sitelink itself is not a factor.

[Shaddows]

I'm with Tedster. If G thinks a page is worthy of a sitelink, it is worthy of extra spidering attention. Similarly, a page getting extra attention might be considered a candidate for sitelinks.

The lack or SERP activity (especially after such a hectic period)- do you think G has stopped twisting knobs, or webmasters have stopped tweaking sites? Or both?

[rjkdesign]

14 of my sites between 0-2 months old all similar in design are getting the following odd search results.

3 days not in top 1000 for main site subject keywords.

Then 12 hours top 10 for main site subject keywords.

Followed by 12 hours not in top 1000 for main site subject keywords.

Then 12 hours top 10 for main site subject keywords.

Followed by 3 1/2 days not in the top 1000 for main site subject keywords.

It's crazy

[confuscius]

"It's crazy"

Nope, it's normal.

[Shaddows]

Have a look at last months discussion of partitions. I would suggest yours is getting passed between a 'normal' partition and a sub-prime one. Possibly the sandbox is now a partition?

14 sites though? Hmmm. Are they on similar topics? Share a backlink profile overlap? Same Whois? Interlinked?

12 hour period (in the sense of repetitive time block)? Are they business focused, home focused or otherwise specialised so time-of-day is important?

[rjkdesign]

The 14 sites are each promoting individual internet models and the domains contain the name of the model.

Time of day is not important. Each site has incoming links from the same 6 multi model sites which have pr3-5 and are hosted on separate servers.

[Shaddows]

First thing I would do is ensure there is not near-duplicate content on the multi- and single- model sites (I assume the singles are taken from the multis, but with special treatment).

Secondly, I would seek strong independant backlinks for the sites so they stand on their own.

Then I would stop fretting for 3 months and see where I am after that time.

[rjkdesign]

thanks for the advice shaddows :)

[northweb]

My site is 6 months old and is starting to show up in the top 20.
When I check the google dance tool in most data centers it shows top 10 for 2 keywords and bouncing around this month. But on a few data centers it is not showing up one of which is google.ca.

I'm targeting the Canadian market. The site is hosted in Canada but I noticed it uses U.S. ip addresses. Should I consider moving the site or wait for the December serp changes to finish.

thanks

[robzilla]

Should I consider moving the site or wait for the December serp changes to finish.

December SERP changes will be followed by January changes, which will be followed by February changes, etc. etc. There's constant flux. If the holiday season is a busy one for your site, it might be better to put any major changes on hold until it's over.

[wheel]

My thoughts would be to change it now.

[Shaddows]

I'm targeting the Canadian market. The site is hosted in Canada but I noticed it uses U.S. ip addresses. Should I consider moving the site or wait for the December serp changes to finish

I've never done that, but what I've read suggests the longer the site is geotargetted to the wrong region, the harder it is to move (or get G to realise its wrong). Theres no such thing as SERP changes finishing, and your site is young, and only just breaking into the top of SERPs.

In summary, I would move now.

[kevsta]

anyone see them tweaking the dials again in the last 24 hours?

looks like they're playing with the strength of the bond between index and inner pages again to me.

last night i was getting 2 different serps again for one query and one was the usual 3 sites all with indented listings #1 & 2, #3 & 4 #5 & 6, and the other SERP had the same sites but with "broken" connections, so #1 was still at 1, but it's second page was at #7

so we had #1 & 7, #2 & 6, #3 & 5, with #4 stationary on both serps , (actually the 4th site, but in 7th position on first SERP because of the dropouts)

does this make sense? anyone else see it yet?

[edited by: tedster at 7:40 pm (utc) on Dec. 3, 2008]

[kevsta]

lol, I see Tedster.. ;)

the serp is doing the same thing right now here. massive disconnect between index page and inner pages.

[gomer]

I have a site hosted in the US that serves a local Canadaian GEO market. The site has done great for years in the SERPS. I think the issue is not where your sites is hosted or the IP address but the content and incoming links for your site.

[tedster]

disconnect between index page and inner pages

I'd love to hear more about what kind of disconnects you see. One thing I've seen in the past days was not so much about ranking changes, but the second results from the same domain were not always indented. Now today, they seem to be indented twice as far! But those are just display changes, and I don't watch them very closely.

On the SERPs that I'm watching, I still see pretty strong connections between the Home Page and indented results showing up.

[wheel]

Does anyone have a summary of what happened in the last month or so? Or a summary of the speculation?

I've been pinned to #1 for my niche for a long time. Recently I yo-yo'ed but then settled down to #2 on my main term but was still number 1 on the secondary stuff. Now I'm seeing my slipping to #2 on secondary terms as well.

I really don't test things, I'm more of a 'throw enough backlinks at it to fix the problem' type but I'm not so sure that this is even my problem...since I'm seeing sites outrank me that I'd like to moan about as having crappy backlinks :). I can't imagine I've tripped a filter since I don't tend to do much of anything specifically, and I think a penalty wouldn't leave me at #2 :). So I'm a bit inclined to think that there's some factor in my backlinks that has been devalued substantially in the last month or so.

(the only other alternative, one of my larger backlinks nofollowed my link recently. I called and they fixed it, but I'm wondering if effectively removing a link for a month or two then putting it back could hurt me that much).

thoughts?

[Receptional Andy]

Most of the sitelinks I've seen, even on larger sites, seems to follow a relatively simple process. Perhaps something like the below:

The site meets sitelinks criteria =>
# Show the strongest-linked pages
- Retrieve the best candidates for sitelinks (most linked to with repetitive anchor-text, and ignoring certain files (like legalese))
- Remove certain (frequently-occurring) patterns from the best link text for this URL (brand or repetitive keywords)
- Display sitelinks

IMO, sitelinks are useful information, because they reflect qualified data from Google. But, I believe sitelinks are generated from ranking-data, and so are a symptom rather than a cause.

[newborn]

Total shake up in the serps AGAIN, sites that were nowhere around are now #3 and #4 and Im on page 2 what the heck happened....

[hughie]

i'm seeing the new batch as well, we went from no.1 to totally gone for a major term, now back at no2.

lets hope this new set sticks!

[jaffstar]

One of our sites today totally dissapeared from G for all keywords, still found if I search of domain.com, it has pr and has pages indexed.

Should I panick or is this part of the update thats going on? We have been going for years.

[wheel]

It almost always makes sense to panic. There's little else can be done :).

[SEOPTI]

jaffstar: please check if your site can be found for domain without .com, if not you've got a filter or penalty.

[directwheels]

One of our sites today totally dissapeared from G for all keywords, still found if I search of domain.com, it has pr and has pages indexed.

Should I panick or is this part of the update thats going on? We have been going for years.

did you check to see if you are in top 1000? you could have been hit by a 950

[tedster]

Here's an interesting tidbit from a rare Matt Cutts post on Google Groups. It seems that there was a bug in the algo for spotting spammy use of iframes. It got spotted an 'hopefully' fixed around December 1.

...our search algorithm saw a large area on the blog that was due to an IFRAME included from another site and that looked spammy to our automatic classifier. I believe that this bug has been fixed now. We also added additional safety checks to the relevant system that would escalate to an engineer if this site had the same issue in the future.

Google Groups discussion [groups.google.com]

Thanks to SE Roundtable [seroundtable.com] for spotting this little peak behind the curtain.

[gford]

As of today I am seeing the site: command really returning bizarre results. For some of my sites it is only returning the PDF of our privacy policy (we have non-pdf also) but the info: command does show many cached pages when checking page-by-page.

I am seeing this on multiple sites, quite bizarre.

[tedster]

That is very odd, gford. I just checked a few sites I'm familiar with and don't see that kind of buggy site: data, but it certainly is good to know that something went whacko again.

I thought I'd highlight some parts of the Matt Cutts comment that jumped out at me:

our search algorithm saw a large area

That sounds "something like" actually rendering the page, rather than just examining the source code. In case anyone has been wondering whether those patents are really in use ;)

looked spammy to our automatic classifier

Does this mean that large gaps on the page can throw an automatic spam penalty? I'm thinking here of the somewhat common practice of throwing a chunk of content at the bottom of a page, only visible after scrolling beyond a false "end of page".

escalate to an engineer if this site had the same issue in the future

An automatic trigger for human inspection on a PER SITE basis - now that's interesting.

[JS_Harris]

Another consideration - Age of content

I've had top spot for a specific vehicle model since I wrote an in depth review on it over two years ago. The article is worthy of top spot on its own content wise but in general the official site will rank 1st for (one of) it's own products.

I knew one of the engineers working on the model long before it was oficially released and (with permission) I covered it eagerly. The article is a full 90 days older than ANY other mention of this model on the internet.

Does it pay to be first (with dictionary like attention to detail of course) ?

edit: Tedster - they've been using an automatic trigger for human inspection on a PER SITE basis for a very long time. They rely on the filters to tell them where to look and they then use that data to improve the filters. Chicken and egg stuff.

[edited by: JS_Harris at 2:55 pm (utc) on Dec. 8, 2008]