May 2008 Google SERP Changes

[Hissingsid]

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

annej - I am very suspicious they do some link buying and wonder if Google is playing around with some algos they think might catch that.

I'm seeing similar signs this morning but I think that this is in reaction to the effect Internetheaven reported above. ie a devaluation in internal links caused an overvaluation in low value external links. I've seen a few sites rise to the very top of Serps that seem to have done it through link buying and link exchanging with off topic and/or poor geo location sites. Those now seem to be dropping.

Also one site that, out of the blue, got sitelinks for a very competitive 2 word term where there is no way it should be top of serps never mind be given sitelinks for that term. I'm pretty sure that this was an indication that external outgoing links, with the right anchor text, was given an undue boost. Today they have lost sitelinks. Lets hope they don't get them back!

What I'm seeing this morning looks like they are either tweaking down the weight of external outgoing links and incoming links or turning internal link anchor weight back up a tad.

Cheers

Sid

[edited by: tedster at 3:16 pm (utc) on May 1, 2008]

[Optimus]

Searching those two IP's from South Africa, I see identical results on all main search terms. There are some minor fluctuations on a few secondary and tertiary terms. I suspect though, that the results I see on those IP's are different to what someone in the UK or USA or Australia would see for the same terms on the same IP's. All part of google's grand geotargeting obsession, which is only serving to kill online businesses whose core clientele is often from other geographic regions/countries.

To test that hypothesis, perhaps a moderator could authorise a two word search term which could then be tested on the two IP's by users from different countries, with the results posted here?

[edited by: Optimus at 2:42 pm (utc) on May 9, 2008]

[Hissingsid]

Has Google started to use the meta description in ranking?

I thought that the description was not used in ranking but have just checked a few terms that I rank #1 for and the only place that term occurs is in the description.

Is this new or have I been sleep walking?

Cheers

Sid

[tedster]

That would be something new - or really, something old, I guess. However, it seems more likely to be off-page factors. Ranking #1 based only on meta description would be an open door for spamming.

When you click on the cached version, do you see the familiar message "the following terms only appear in pages that link to..."?

[Hissingsid]

Hi Tedster,

Yes but I can't find the phrase in anchor text.

I'm doing a phrase search in inverted commas (quote marks) like this "search term uk". I was trying to work out why I'm at #1 for that on both google.com and .co.uk but at #2 on .co.uk for "search term" and #7 on google.com.

Cheers

Sid

[tedster]

The keywords are not always directly in the anchor text on the linking page - they can be in nearby content, or in the page title, and so on. There's a bit of secret sauce in the mix here, but the key is that the keywords are somehow associated with the url that contains the backlink.

[Hissingsid]

Hi Tedster,

In that case it could still be the description that is doing it. I'm thinking of directories that scrape your url and description.

Not as interesting as a direct link between the description and Google rank. :(

Cheers

Sid

[annej]

Hey, it could be the meta description.

If Google is really experimenting with different algos like we are talking about at [webmasterworld.com...] anything is possible.

My theory has been that they are testing how often people click and stay. But if they are trying new things they may try old ones as well.

[ajay_dun]

My website traffic drooped 75% and most of keywords. I am worry so much whats going on. This problem start from past 3 days.

[bluntforce]

Totally anecdotal, but it appears to me that some sites carrying stronger PR are now doing the "serp bounce". I previously haven't seen the strongest sites, measured by top 5 results, move a whole lot in the most recent shakeups. Today I've seen some of those top 5 bouncing randomly in the top ten. Just an observation, it's too soon to attempt analysis.

In the days of the "Google dance", it always seemed to start in the higher PR sites, then work it's way down. Is it possible this is a variation starting on the lower end and working it's way up? I don't know, but it'll be an interesting week.

[tedster]

I've just seen an interesting shift in a couple different SERPs - relatively low competition keywords, but that's what made this pattern show up more prominently.

A few groups of domains at the top of these searches have the same owners. The domains within each owner network interlink very openly and in a very natural way for the visitor. There's no attempt at deception and the Whois is clearly the same.

As of Friday, it seems like the "same owner" links got devalued a bit more and the URLs with more independent backlinks got a boost.

[edited by: tedster at 5:54 am (utc) on May 12, 2008]

[annej]

Happiness is seeing the short, inadequate and not very accurate Wikipedia page on my topic is sinking in the serps!

[edit_g]

Weird. Since Friday site took a dive. Replaced by much lower quality pages on most key terms but some still hanging in there. Very strange.

[mmiller]

For the last 3 days my main site's traffic has dropped to nearly nothing - literally 4 (four) people on it a few moments ago.

I checked my search ratings at Google and what I found is that all the scraper sites that have stolen my pages (complete with my domain name on them) are now raking at the top and I've dropped down to several pages back - if that.

Looks like Google is 'fixing' things again and punishing the people that create content and rewarding the ones who steal it.

Methinks Matt Cutts is at work again...

[Hissingsid]

I'm seeing two distinct result sets on different DCs this morning. More pronounced for the most commercial terms, less so for generic, less commercial terms.

Could be just migration but might be worth looking for a pattern as this could be part of split testing.

Cheers

Sid

[QuantumEntanglement]

My website traffic drooped 75% and most of keywords. I am worry so much whats going on. This problem start from past 3 days.

Weird. Since Friday site took a dive. Replaced by much lower quality pages on most key terms but some still hanging in there. Very strange.

For the last 3 days my main site's traffic has dropped to nearly nothing

We are seeing the exact same thing for the last 3 days, as reported in this thread: [webmasterworld.com...]

[edited by: QuantumEntanglement at 7:27 am (utc) on May 12, 2008]

[QuantumEntanglement]

Just searched for a unique phrase from one of our previously well-ranking widget profile pages ... there are 3 results, the top 2 are scrapers, followed by us, the originator. The #2 result is a 0PR scraper who copied our domain name and a bunch of our content in a haphazard way, I've seen them before buried in serps but was basically LOLing at them ... now their PR0 scrape outranks our PR3 content?

Here is the kicker... I searched for the title of the homepage ... 4 results ....

scraper
scraper
scraper
italian translation of the home page added about a week ago

Something is seriously wrong here.

[AZEvil]

As of Friday, it seems like the "same owner" links got devalued a bit more and the URLs with more independent backlinks got a boost.

I'm seeing the same thing. For several days on a couple of terms I watch, the biggest player in the space had 5 sites in the top 10 and now they're down to just two on each of the SERPs...one is their main site and the others are an exact keyword match for the searches.

[c41lum]

I am still feeling the hit on my .co.uk site since Thursday. The results are all over the place for instance one of te search terms I watch now has a top 3 result with no text just a image. Very stange.

[netmeg]

For the past two weeks, I've ranked between #3 and #7 for a single word generic keyword that my (geo targeted) site never ranked higher than #55 for - and last night, fell back down to #81. My normal keywords are all still ranking #1, since my site is specific to my area, but the traffic sure was nice. I really wasn't expecting it to last; to be honest there are plenty of other sites that should probably rank above me for that single word, so I can't get too upset about it.

[crobb305]

I have observed cache dates changing, and going backwards in time. For example, my site was cached with a date of May 8 (on the 9th). Then yesterday, it said "May 7". Today "May 6". This is probably a result of datacenter variance, but it struck me as odd, in light of the reported low Googlebot activity.

[rumblepup]

I've seen a significant drop in SERP's for a position I've had on the Front Page for 2 years (and maintained it by seeking product reviews from blogs and directory listings) I killed all reciprocal links a year ago. But then about May 1st, 2nd, and 3rd, I saw myself drop from 3 and 4 to 12 and thirteen, and be replaced by sites that all belong to the same company, and all link to each other. They DROPPED PR on the last PR change (I watch my competitors)

I've worked in all white hat methods on my site since I can't remember (no I can, since I got a serp hit for something shady) and provide good content and update my blog every single day.

What get's my goat are the sites that got taken up in the SERP's. No inbound links but their own. Yup, full on triangular linking.

What give's here?

[chelseaareback]

we now have something very odd happening in that G is listing pages of ours that done really exist at all. Its sort of guessing the urls. for instance lets say we have pages like "blue-special-widgets" on blue topics and then others called "red-boring-widgets" . We are finding indexed in G 'red-special-widgets' or blue boring ones. Site set up ensures neither come up as 404 but how/why they have listed these erroneous pages is mystifying

[Optimus]

Perhaps I have been unobservant, but I have just noticed when searching from my home country and changing the URL from www.google.co.za to www.google.com that it redirected to www5.google.com/ I did some testing and there is also a www6.google.com. Both www5 and www6 are giving the same results. www7.google.com and above give a 404 error, while www1.google.com to www4.google.com all redirect to the normal www.google.com.

Has this been around and I have just failed to spot it, or is this part of the current update and geotargeting algo?

[MadeWillis]

Site set up ensures neither come up as 404 but how/why they have listed these erroneous pages is mystifying

Google has been trying to complete forms and use drop-downs and such. Could this be what is causing your new URLs? If so, try to add noindex to these pages as they are created.

[tedster]

replaced by sites that all belong to the same company, and all link to each other.

Well, that shoots down the pattern I thought I saw.

Site set up ensures neither come up as 404

That sounds like a potential problem to me. setups like that have eventually tanked many sites I know of. "Danger, Will Robinson!" - I suggest you let your server return a 404.

[chelseaareback]

thanks tedster and madewillis

will pass this on to the tech chap that does all this for me

[chelseaareback]

actually come to think of it if we allow 404 to be returned then an we will be serving an awful lot of those up. G has chosen these pages over the real ones. wdnt having an index full of 404 pages (and we are talking about 100,000+ unique ones) be even more of an issue?

[tedster]

If the server returns a 404 status, the url will be dropped from the index. If you have a Webmaster Tools account, you'll see it there for a while, and googlebot will keep requesting those urls that used to get a 200 for a long time, on a decreasing frequency. But it WILL be dropped from the index, and that's what you want to see.

[ajay_dun]

Thanks

[mfishy]

Hehe, been following a certain serp and I swear the top 10 changes dramatically all day. Very strange as this has been going on for a while. It's impossible to come to any conclusions when a page goes from #4 to #13 to #8 to #23 in a 4 hr span.....