Forum Moderators: Robert Charlton & goodroi
But something quite major is now going on with the Google SERPs. Members are reporting major changes in the single word search results, most particularly, but lots of other things are stirred up as well. On one of my single keywords, I've just passed wikipedia (yeah!) and jdMorgan reported the same for one of his keywords.
So it's time -- we are officially declaring Update Buffy. We'll begin with recent posts from our June SERPs Watching thread [webmasterworld.com]. What do you see going on?
< So why name it Buffy? Let's say it's in honor of someone who just
left her job but knows a whole lot about Buffy the Vampire Slayer.
We've never gone with someone's formal name for an update. Tip
of the hat to reseller, goodroi, and jdmorgan for their input >.
< further note - Matt Cutts says [mattcutts.com] that in his mind this is not an
update. OK - so we'll call this moment an "Honorary Update". >
[edited by: tedster at 7:27 pm (utc) on June 19, 2007]
I like those related searches at the bottom of the page though. My main topic is there and my site is number one on that phrase. Who knows, maybe it's better than being at 5 or 6 on the one word search.
It is strange how it the changes are so marked in one word phrases. I'm not seeing any big change in 2 or more word phrases.
So it would make some sense for Google to play with the algo for one word search, working toward greater end-user satisfaction.
I think this is mainly a semantic play, coupled with the user intention/satisfaction data that Google has been collecting for quite a while. I think they are intentionally forcing a wider variety of "information types" onto the first page.
I think they are intentionally forcing a wider variety of "information types" onto the first page
That seems obvious in one of the areas I search.
They know which secondary keywords are most often combined with a particular word. Those most searched secondary keywords are most competitive, and sites in that sector would need more backlinks.
So they devalue those sites that fit the most competitive sectors to help raise up others that are relevant for the single keyword search, but not in the super competitive sectors. That seems to be the change I've seen in single keyword search.
So it would make some sense for Google to play with the algo for one word search, working toward greater end-user satisfaction.
That would also explain the related topics shown at the bottom of the page.
Also would this give any credence to my theory that Google is tracking the most clicked on results and moving them to the top?
BYW another result I'm seeing in the top 10 is Topix. It seems to be all automated news. Would this be part of Universal Search?
[edited by: walkman at 2:42 pm (utc) on June 16, 2007]
How old is the site you talk about?
They're manipulating the algo to bring certain type informational pages up in the rankings.
Wikipedia
About
Channel? News
University site
University site
Topix
Britannica
Large Health Site
DMOZ
Answers
Amazon
Government Site
Another Large Health Site
Books. Google
Yahoo Answers
Dictionary Site
Article Site
Directory
How old is the site you talk about?
No site in particular, just for a lot of queries if the first twenty results were 80% index page returns now its 30%.
They're manipulating the algo to bring certain type informational pages up in the rankings.Wikipedia
About
Channel? News
University site
University site
Topix
Britannica
Large Health Site
DMOZ
Answers
Amazon
Government Site
Another Large Health Site
Books. Google
Yahoo Answers
Dictionary Site
Article Site
Directory
Maybe because Google regards most of above sites as "authority" sites!
The overall phenomenin with one word searches is probably related to Universal search but in this specific case it is more due to lousy search egineering.
The .net and .com both rank despite being duplicates, and despite blatantly violating Google's adsense policies, and despite having no original content, etc.
It's weird to see the one word results so bad while other results are basiccally very good. The one word results toss niche relevance out the window, serving up bad quality pages on large, poor quality domains like about, answers, topix, amazon... and most weirdly, lycos search result pages. This stuff simply does not do this well for two word searches.
One thing it does make clear, yet again, is Google and the other engines need to finally stop being so pigheaded and only rank two pages from a domain. It's five years past the time subdomains should be catered to. They are just pages on a domain -- and often set up for no reason other than to specifically rank better and independently on search engines.
Take a look at the list. Some areas are highly susceptible to a flood of any type authority site where others aren’t. When I was looking at one popular area recently none of those type authority sites could flood the area.
What Tedster is talking about took place in my areas long ago. Now they’re upping the anti even more.
I'm just doing a little whining because I can't gather 100 one way links a day like some of these sites.
In the latest results internal pages that have ranked top 3 for over a year are now ranking back on the third page.
The one thing that the current crop of top ranking sites have in common is a gazillion links.... they are not necessarily the most relevant sites.
This may be an over simplification but is it possible that link volume is unduly influencing the current SERP's whereas before that was not the case?
Is link volume drowning out relevance?
Now they've jumped into about everything related to the keywords including the long tail.
Yes, including stub pages. Google has just GOT to catch that. End users are not amused by pages that just spit back their original search terms and give no further information.
Yes, including stub pages. Google has just GOT to catch that....
Yes, the stubs are ridiculous, but let's be careful what we wish for with Google. ;)
The collateral damage accompanying such a fix could be horrendous.
Is link volume drowning out relevance?
Looks that way. Now instead of the history of the topic and articles about methods by one of the oldest magazines in the US we get free stuff, wiki stubbs, automated news like Topix, and About (it used to be good but now it's just designed to get you to click on ads)
At least no one is mentioning that spam sites, scrapers or MFA sites are at top of the serps anymore. Wouldn't that be considered as major improvement of search quality?
At least no one is mentioning that spam sites, scrapers or MFA sites are at top of the serps anymore.
You are wrong, reseller. At least in my category there are scrapers all over the first two or three pages.
That would be half an answer, but the one word serp I'm watching has two contradictory things: the about, answers, topix crap... AND completely trivial, nothing sites with under 40 links total (according to yahoo) when the normal sites ranking in those slots have 25,000+ links. So link volume is definitely not the trump card here.
(And just to repeat again, I don't see any of this phenomenon in the non-one word serps. Multiple word serps are way better than average of the last three years. )
The 10 most used word phrases in search engines on the web are:
1. Two word phrases 28.38 percent
2. Three word phrases 27.15 percent
3. Four word phrases 16.42 percent
4. One word phrase 13.48 percent
5. Five word phrases 8.03 percent
6. Six word phrases 3.67 percent
7. Seven word phrases 1.63 percent
8. Eight word phrases 0.73 percent
9. Nine word phrases 0.34 percent
10. Ten word phrases 0.16 percent
[edited by: tedster at 6:39 pm (utc) on June 17, 2007]
[edit reason] integrate the link [/edit]
Scoring high for single word queries isn't going to bring in that much traffic, but it's a good indication of a site being trusted and on-topic.
Well, at least it should be.
But I can't recall the last time when I was looking for something exact and typed in but a single word... except when I knew it's a trademarked brand name.
[edited by: tedster at 6:34 pm (utc) on June 17, 2007]
When I look at my Feb 2007 stats when the one word phrase was #5 it only brought 3.2% of my search engine visitors. Now that the one word result is bouncing between #19 and #22 the phrase has dropped to 1.2%.
Wow, I just noticed something else. In Feb 2006 the one word term brought in 5.7%. So it has gone down even though the page was at #5 both years. I have to wonder if more are finding pages on the long tail or less people are using one word searches.
BTW my traffic is still just fine. The long tail is bringing in plenty of visitors.
In order to display videos, etc, on the same SERP they did need to somehow normalize the relevance algorithms for regular SERPs with the legacy relevance scales for video search, etc -- and they may not have got it exactly right when Universal Search launched.
I'm thinking that one word searches might be an important place for them to get Universal Search right, and hence the big changes we see.
No, you need to look at what reseller posted.
One word queries are the huge, big ticket searches on the Internet, from the #1 search in the world on down.
It's not like there are a jillion genuine one-word searches in the world that people actually search for. Many words need a second word or more to make sense, "new" or "online" for example.
Also note that one word searches equal the volume of five word searches or more.
In order to display videos, etc, on the same SERP they did need to somehow normalize the relevance algorithms for regular SERPs with the legacy relevance scales for video search, etc --
You may be on to something there. I have a site I built many years ago that used DMOZ data to make a vertical portal/directory. I got hit hard back in Jan 2005 and I pretty much abandomed the site. I assume I got dinged for duplicate content.
Maybe a year ago I made the site more of a mashup and added some RSS feeds and Flickr photos. It didn't seem to help any.
My traffic has gone from about 225 visitors/day to over 5000 visitors/day.
The current toolbar is a gray bar for the site. The site was a PR5 at it's peak.
Exactly. The thing is have they ever stated what "Universal Search" really is? Other than stating they've been working on it two years I haven't seen where they really defined it.
