Google Updates and SERP Changes - July 2010

[scottsonline]

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

@john1975 the search term I posted weeks ago showed how bad Google was messed. For a year or two that search returned a real product, for a few weeks/month it return gibberish. It started to return real results a few days ago and is once again accurate or close to it. My guess is millions of other semi-long terms are now also snapping back into line.

The traffic pattern is evolving back towards normal. Google was up again yesterday but only 8% from a week ago. Bing was up 17% but Yahoo was down 35% which only confirms what the studies show - most people will go to Yahoo 2nd if Google fails. What's so clear to us is that it's that 5%-8% of total clicks that Google had filtered off that were the ones converting at probably 80%. Highly targeted searches by informed consumers that didn't need the magic intervention of Google trying to divine what it was they really wanted.

Thanks to MC and Google for slowly figuring it out. Like everyone else I was a bit harsh at times but it's tough to spend years reading the WMGL, reading everything the people from Google wrote and following it within the rules and then viola a major change happens that devastates our business and pretty much everyone else we knew in favor of spammers, scammers, mirrors, mashups etc. The filters still aren't weeding out those results and I'll say it again instead of trying to tinker with the end result why not hire/invest/spend more time looking at actual spam reports when legitimate sites report the ones running 45 sites selling the same item under different keywords which actually really DOES degrade user experience and user choice?

[edited by: tedster at 5:22 pm (utc) on Jul 1, 2010]

[gouri]

In GWT, for the Average Position of a keyword, is that number determined from the positions of the keyword from the actual clicks to a website, or is this the average position all over the world (not just the position of the keyword from clicks but just rankings overall in Google USA, Google Canada, Google UK, etc. also)?

[ohno]

Is anyone else seeing identical figures for impressions of various keywords in WMT?

[anand84]

I see the same. There are a bunch of keywords showing the same number of impressions - a dozen showing 880 impressions, another ten or so with 720,480,390,320 etc..

[tomapple]

I too have quite a few keywords showing the same number of impressions in addition to the odd impressions/clicks to queries ratios I mentioned in another thread.

[mrez74]

The last day or so I'm again seeing two different set of SERPs for several of my keywords. It's about 50% one set and 50% the other set. One set has my keywords in better higher position compared to the other set.

It's weird that for the last couple of months it has been stable with just one set of SERPs. Now it's back and forth like it was back around March and April of this year.

Anyone else noticing this as well with their sites?

[curioustoddler]

I am seeing backlinks from one or our owns sites in webmaster account in external links but they are not showing it from site.com/../ , rather these backlinks are appearing as IP/../.. and this IP is our site's IP. Why these URLs are appearing as external backlinks?
It haad happened in 2008 when we switched hosting shifted to VPS for this site and it went away within few weeks. Now google is showing it again.

[ohno]

@ mrez74-yep saw that yesterday, two completely different sets of results being served at different times of the day, may explain why we seem to have a switch flicked on at certain times and sales flood in!

[ohno]

Just checked our logs, soemthing strange happened last night at around 2.30am, no visitors until past 6am this morning and lots of foreign traffic with junk search terms (on more than one site). Looks like today will be "fun".

[Martin Ice Web]

mrez74 - yes, since last friday theres something like two or more sets. But holiday season started, so i am not surprised. Now the students have taken over the gog complex while the engineers are on vacation from their hard work the last 6 month....

[FranticFish]

When it comes to the long tail, what I'm seeing so far seems to gel with comments made in last months thread by Victor - [webmasterworld.com...] and also message #4152370

One site I've taken on recently always ranked well on the home page for various terms like 'brand of widgets', but nothing long tail because the actual product pages were only crawlable via paginated search results and were unoptimised and uncategorised.

We made adjustments to the site architecture so that we introduced about a hundred new category pages optimised for 'Buy used / secondhand <type> <widget>'. This introduced these pages into the top 20/30, and from there we planned to build links to those pages.

However what I'm seeing now is that all these pages have fallen dramatically.

What's interesting to me is that if I search for 'Buy <type> <widget>' the rankings are HIGHER than for either 'Buy used <type> <widget>' or 'Buy secondhand <type> <widget>'.

In other words it appears that the anchor boost the home page enjoys on <widget> helps inner pages rank for terms more closely related to that GENERAL phrase, and because the home page does not have a link profile containing the RELATED phrases, the inner pages rank worse for the more specific queries that they are optimised for.

Therefore, by building links to the home page containing these generic terms ('buy' 'used' 'secondhand') we could help the inner pages.

Is anyone else seeing anything like this?

[aristotle]

Is anyone else seeing anything like this?

FranticFish-
I think it helps to have the keywords and keyphrase variations scattered throughout the site on most of the pages. In this way all of the pages tend to support each other and boost each others' rankings. I can't prove that this technique works in every case, but it seems to have worked out quite well on my own sites.

[FranticFish]

Thanks, I took note of Hissingsid's remarks on this too and it matched what I saw on other sites so we've started a blog for that reason.

The products themselves are in a feed which is exported to at least four big resellers so I have to focus on the category and subcategory pages as these I have text control over and can make unique.

[CainIV]

It's about 50% one set and 50% the other set.

I see the same thing as well

[sandboxsam]

I am seeing one domain showing up in search results for about two weeks now that it has been abandoned for at least 6 months.
Only one page is available and that is a place holder for hosting:
“Welcome to Your new Web Hosting Account!
We would like to welcome you to your new Web Hosting Account…“
This site ranks well for many keywords and is moving up in the search results. There are no key words on the webpage.
However, there are still links pointing to this domain even though G shows none when site:domain is used.
Content is king; I think not! Content does not seem to matter here.

[Andylew]

Things are still in a bizare state here.

12th July we saw a doubling of our traffic which was nice this has stayed stable however this seems to still be based on 10% of our pages in the index.

On the 12th for around 20 minutes I saw site: return a 'correct' number of pages (matching wmt) and indeed a peak is visible at that time in our stats. This was returned around 50% f the time so clearly two datasets in play.

As it stands now we are seeing improved rankings for our listed pages - the number of listed pages has not increased since 90% disapeared on 1st may. The 10% which are listed however have outdated descriptions although the caches are recent and some of the titles are also months old.

It looks like google are still tweaking based on a now rapidly aging longtail dataset then possibly once that looks good swap onto the more recent dataset.

[webdevfv]

In the last month and more I've noticed one particular 'company' doing very well in Google.

They've got 3 or so main websites and then tons of doorway pages all linking to those main sites that sell the same products but with slightly different emphasis.

Hmmm, so much for duplication of content, doorway spamming etc. Does Google want us all to go down this route because if these sites remain in the top 10 for much longer others are bound to follow... Nudge nudge, wink wink

[ohno]

Dire week here this week, not one sale today either. Traffic has been around normal but lots of foreign junk traffic.

[Allen Graves]

webdevfv,

Are you sure that they are not doing any other type of SEO that is giving them the rankings they are currently being given?

While those tactics you mention are taboo, maybe they are being overshadowed or cancelled out by some other white-hat methods.

Just a thought - I mean, in all reality, Google would be nuts to want everyone to start doing that crap again.

Allen

[seocolin]

@Bastilar
I managed to get a high quality PR5 (home page) link, it has been indexed and my site is back in the top 10.
Could be a coincidence but there may also be something in what you said. Thanks again for the advice.

[tedster]

@seocolin - happy to hear that good news. In the past I have often seen one good new backlink fix a world of troubles. Good to know that it still seems to happen.

[nethead]

@seocolin
did it improve on only one keyword or did you improve across the board including long-tails?

[seocolin]

@nethead - site has improved accross the board, long-tails and some new terms I tweaked for over the last few days.

[backdraft7]

@ohno - again we are seeing the identical pattern you are experiencing - things have been pretty much back to normal, then yesterday the switch went off again - same traffic levels, but no conversions - aka Zombie Traffic.

[ohno]

Weird isn't it? After typed that we had some sales, what I do not get is we have never had patterns like this since 2002. There is either a big change brewing or this is the new G, massive changes from day to day....

[member22]

< moved from another location >

i checked one of my keywords 10 minutes ago and in the US I now rank 2nd.

However in Europe my ranking hasn't moved ... why is that? is that a test with some filters that google is applying ( which happened to me a few months back ) or is it because they update the US 1st before Europe ?

By the way the PR hasn't changed...

Thanks,

[edited by: Robert_Charlton at 4:20 pm (utc) on Jul 17, 2010]

[Fraggler]

My site which has been affected since the May change has repeated a spike I got 31 days earlier.

The rankings for my main keywords across many pages disappeared (each can be found in the last few pages of results) at the start of May - traffic dried up pretty significantly.

On the 15th of June all rankings returned to their previous highs (several at new highs) and traffic flowed rapidly. 12 hours later and the Google traffic disappeared again and so did the results.

Thursday/Friday (16th of July) all rankings returned again and disappear 12 hours later.

I don't know if this is any use to anyone but it does seem strange to have such a short boost and to be 1 month apart as well...

[fred9989]

My sales reveal clearly when what I have termed the "old" and "new" results are running - because the "new" results don't produce any sales...and that's because my sites (health sites) aren't directly featured.
What does appear in the results are what I'd term feeder sites, in other words, satellite or peripheral sites which link into the main sales sites. All of them are substantive sites with useful content, so it's not an issue of substance. This seems to be much more about the identification by Google of networks of sites interlinked in a way that allows them to spot self-promotion. There's also something going on about links, which seems to amount to a downgrading of the value of anchor text on links into a site. The same may be true of onpage factors.
For example, the search term "causes of xx", where xx is a medical condition, produces inexplicably similar results to "treatment of xx" and "cures for xx" - or indeed, "xx treatment" and "xx cure". This is a very new development, and the results presented seem rather generic, not specific to a cure or treatment of cause for condition xx; more a sort of general information resource. I suspect that could be to do with quality and authority signals being rated more important right now than content and link factors like anchor text. So there may be some kind of evolution or trial going on, but once again the results don't seem to offer the searcher exactly what the search terms imply.
Fred

[Planet13]

For example, the search term "causes of xx", where xx is a medical condition, produces inexplicably similar results to "treatment of xx" and "cures for xx" - or indeed, "xx treatment" and "xx cure".

Instead of looking at it as google simply placing less value on anchor text, is it possible to look at it that caffeine has allowed for better implementation of semantic variations, that is to say if the search was for the words "treatment disease X" that google now sees "cure disease X" and "remedy disease X" and "disease X treatment" having more or less the same value in terms of relevance?

That would explain why so many people have done a search for a term like "used titanium widget" with all sorts of sites returned in the SERPs, and then at the bottom of the google search page, seen a link that says, Search for "used titanium widget" (the exact same phrase they originally typed in)?

So the original query returns results with more semantic variations, while the "refined" results have less semantic variations?

[fred9989]

Planet 13
I agree with you and I think that is what has happened, however, the interesting thing is that the pages that turn up are not actually relevant to the original query.

[savweb]

I see unusual(higher) google bot activity for about 4 days, site:website.com has increased by about 80% and growing. Webmaster toots external links updated more frequently now, looks like something is happening, but results are still full of garbage.