Google - Ten recent algorithm changes

[kidder]

This is well worth a read if you have not seen it yet.

Ten recent algorithm changes
Official Google Search Blog
11/14/11
[insidesearch.blogspot.com...]

Today we’re continuing our long-standing series of blog posts to share the methodology and process behind our search ranking, evaluation and algorithmic changes.... [W]e’re always looking for ways to give you... insight into the over 500 changes we make to search in a given year. In that spirit, here’s a list of ten improvements from the past couple weeks....

[edited by: Robert_Charlton at 5:53 am (utc) on Nov 15, 2011]
[edit reason] added source info and quotes [/edit]

[Robert Charlton]

This one in particular grabbed my eye....

Better page titles in search results by de-duplicating boilerplate anchors: We look at a number of signals when generating a page’s title. One signal is the anchor text in links pointing to the page. We found that boilerplate links with duplicated anchor text are not as relevant, so we are putting less emphasis on these. The result is more relevant titles that are specific to the page’s content.

I've been seeing an ever increasing number of pages recently with "augmented titles", as I call them. I'll venture a guess that this signals eventual reduced reliance on title elements, most likely evolving as Google tests user satisfaction with the changes it makes. That said, I suspect the title will remain a strong signal for a while.

PS: The reduced reliance on boilerplate anchor text is also to be watched if it's a ranking factor as well as a title factor.

[kidder]

Extending application rich snippets: We recently announced rich snippets for applications. This enables people who are searching for software applications to see details, like cost and user reviews, within their search results. This change extends the coverage of application rich snippets, so they will be available more often.

Pretty soon they will start framing entire websites...

[tedster]

Am I missing something here? All ten "algorithm changes" seem to be about the visual and textual appearance of the SERP, but there's nothing about actual ranking changes.

[chrism]

To me this highlights a misconception I had about the 500 changes a year to Google search. It sounds as if they would include an extra pixel of padding above the search box as one of the changes, whereas I thought (like tedster it appears) this meant changes to algorithms ranking sites in the SERPS.

[AnkitMaheshwari]

Pretty soon they will start framing entire websites...

+1 :)

[onebuyone]

Refining official page detection: We try hard to give our users the most relevant and authoritative results. With this change, we adjusted how we attempt to determine which pages are official. This will tend to rank official websites even higher in our ranking.

This is clearly a ranking change, but I'm not sure if I understand this correctly.

[deadsea]

Even when Google actually tells us about some of the changes they are making, they word them so vaguely to be next to useless.

The image ranking one sounds like it could affects some folks here. I've heard people say "my images are hot linked all over the place and they rank great in image search". It sounds like that hot linking won't help you now. It may be time to reign in your bandwidth costs and prevent image hot linking if you had done it for image search.

[indyank]

Refining official page detection: We try hard to give our users the most relevant and authoritative results. With this change, we adjusted how we attempt to determine which pages are official. This will tend to rank official websites even higher in our ranking.

This is clearly a ranking change, but I'm not sure if I understand this correctly.

They have asked their quality testers to rate pages as "vital", if they are the official pages. This is well explained in the guideline doc.

For example, if a user searches for Windows 7, the official pages will be from Microsoft and they have adjusted their algo to always return a set of relevant pages from Microsoft's domain or sub-domain on top, for "Windows 7".

[Pfui]

Pretty soon they will start framing entire websites...

Methinks they've basically been there, done that... "Google cache raises copyright concerns" (2003) [news.cnet.com...]

[Content_ed]

Pretty soon they will start framing entire websites

That program has existed for years, it's called "Blogger."

They just leverage the greater web comunity to do the framing for them so they can hide behind the DMCA.

[Bewenched]

Pretty soon they will start framing entire websites..

That's exactly what their image search does, just use a bit of javascript to bust out of frames.

[zeus]

tedster - yes there is nothing in that blog that really would help with SEO and if, it would not be worth doing. We always have to remember that google will never be 100% clear about anything, so some words from Google can also just be stuff they would like us to do to our sites, but with no influence on ranking.

[lfgoal]

Better page titles in search results by de-duplicating boilerplate anchors: We look at a number of signals when generating a page’s title. One signal is the anchor text in links pointing to the page. We found that boilerplate links with duplicated anchor text are not as relevant, so we are putting less emphasis on these. The result is more relevant titles that are specific to the page’s content.

This grabbed my eye as well. But what do they mean by this? Do they mean that the anchor text you use for internal links to one of your pages may be subbed by google for one that their algorithm "thinks" is more relevant to the content on the page? Does that also mean that the anchor text selected by some outside site linking to your page may be subbed by google for the same reason.

...does this also mean that internal linking may offer less bang? And if so could we differentiate between boiler plate anchor text that is used for internal linking and contextual anchor text (appropriate in-body phrases used to link from one internal page to another).

[tedster]

That's the way I read it, Ifgoal. It's a change of degree for what they use when they change the title that is displayed in a particular SERP.

[bwnbwn]

I assume tedester this does not include the navigation on the site but internal links I point at this page. Dublicating the same text for the link all through the site is what we all consider part of being over optimized anyway correct.

[indyank]

I guess they are referring to anchor text used on site-wide sidebar or footer or navigation links (like those on the menu bar) as boiler plate anchor text.This could either be on site links or links coming from sidebars and footers of external sites.

[Panthro]

I really dislike that Google takes the liberty of rewriting Titles to appear in the SERPs as what they believe is more accurate or preferable.

[indyank]

I really dislike that Google takes the liberty of rewriting Titles to appear in the SERPs as what they believe is more accurate or preferable.

No, they do it for their users as they know what their users wanted to see :)

[lfgoal]

I assume tedester this does not include the navigation on the site but internal links I point at this page. Dublicating the same text for the link all through the site is what we all consider part of being over optimized anyway correct.

Makes you wonder if the algorithmic preference would be to push pages of a site toward a wikipedia layout-design approach.

[lfgoal]

...but in essence, I don't see how anything has really changed from a pragmatic stance. The answer seems to be:

1. Produce more pages reflecting what users are looking for (phrases) with content that is good enough to keep them on the page and site.

2. Get more links, the best being those that are given based on quality of content since those typically a lower degree of manipulation.

3. Drive user behavior of your site (get them to the site, keep them on the site) and use this plus site reputation to gain trust.

Wikipedia manages to keep users on the site for quite a while. I can't tell you how many times a visit has turned into seven different bounce-arounds just by following the various in-text internal links.

[tedster]

rewriting Titles to appear in the SERPs as what they believe is more accurate or preferable

That's done on a query by query basis.

In theory it's a good idea, one that should improve the information signal for each page in that particular SERP and help draw clicks. In practice, there certainly have been some spectacular misses while the title rewriting algorithm gets developed.

[smithaa02]

Refining official page detection: ... With this change, we adjusted how we attempt to determine which pages are official. This will tend to rank official websites even higher in our ranking.

Super bad news. Google 'vital' system is fundamentally flawed...vital rankings are being handed out willy nilly without proper regards for what should be truly vital.

In the EWOQ guide they give the example of Derek Jeter's stat page on the Yankees website (if memory serves me correctly) as counting as vital for the keyword Derek Jeter. This is silly...so what if he gets traded? Then they have the example of the Chicago government counting themselves a vital for the search term Chicago (again if I remember this right). This is absurd...the chicago city government should absolutely count as vital for themselves (aka searches relating to chicago + government) but they are not Chicago themselves. I'm worried this could be a huge coup for government/political website even if people aren't doing political type searches. The worrisome thing is if this is controlled by testers, what happens when they assign chocolate to Hershey's? Books to amazon.com? I just don't see the vital program as being properly supervised to prevent this type of mess.

What truly is vital...would be ownership. If an artist has their own website...that should be vital...even if it is a facebook site. This needs to be much more literal than it is now.

[Pfui]

What truly is vital...would be ownership. If an artist has their own website...that should be vital...even if it is a facebook site. This needs to be much more literal than it is now.

Agreed!

I tend to a prominent film artist's site. He owns it, he's active with it, he's been at it for 14-plus years (as have I; bless him:) It's content-rich and as vital/authoritative as you can get on his works, on how he works, on him. Size-wise, it's the equivalent of a 400-page book.

And it'll be a cold day before his site rises above Amazon-owned IMDb's single, solitary page for his name.

(At least we're above WIkipedia's single page. Finally.)

[smithaa02]

Maybe in the future we'll see something like google local for vital terms... A public registration where you fill in a form for a certain key phrase which you purport to be your own. In the same way google local businesses are checked, 'vital owners' could be checked. What would be nice would be if the community could see the applicants and voice counter-concerns if vital rankings were assigned errantly (which they are being now).

Google does have to be careful though... I love the idea of an artist ranking over an Amazon page...but what should all terms work like this? If you do a search for Jerry Sandusky should his (assume he has one) personal facebook page come up first? If you search for a dictator, do you really want their official home page?

Part of this problem can be solved by being very specific. If you do a search for China, the main Chinese government site shouldn't be considered vital (if the search term was Chinese government that is a different matter). Same with say Saudia Arabia...if amnesty international can compete on that term from the democratic 'link community' then I don't see it wise for google to override this with an official website just because it's official.

Vital should only really be for people and companies that can answer accurately, "who am I" and not "what do I want to represent" or "what do I want to be about". Other associations should be vital for their own sake and not for what they associate with.

[Sgt_Kickaxe]

Pfui, thanks for sharing that. When Google makes a change that has you saying the opposite then we'll know that Google's changes are indeed impressive.

[MikeNoLastName]

--"my images are hot linked all over the place and they rank great in image search". It sounds like that hot linking won't help you now. It may be time to reign in your bandwidth costs and prevent image hot linking if you had done it for image search. --

We have the same problem. 70% of our "page visits" from Google, are actually "accidental" image hits on things as distantly related to our site content as mudpies are to gourmet dining.

--I really dislike that Google takes the liberty of rewriting Titles to appear in the SERPs as what they believe is more accurate or preferable. --

In our experience lately, if you go to webmaster tools and resubmit the page, the rewritten titles go away, at least temporarily... but I wonder how this comtributes to duplicated content penalties. I'm assuming (from some experience) that these "rewrites" are what show up when you come up at the top of search results and they show your top sub-section pages as a group, but we've also found that most of them are ALSO indexed under the correct title.

--The worrisome thing is if this is controlled by testers, what happens when they assign chocolate to Hershey's? Books to amazon.com? I just don't see the vital program as being properly supervised to prevent this type of mess.--

Heh, I saw a perfect example of this. A site which TRIES to pass itself off as the "OFFICIAL TOURISM BUREAU" for a city (but defintiely is NOT), was apparently manually reviewed and fooled the reviewers and now occupies the top spot for that city search.

(edited to add another response)

[tedster]

Mike, that's not exactly the way Google uses quality raters. Even if they weren't fooled by the site, the quality rater feedback would just go to the algo team for future changes. It would not cause an immediate manual tweak to that SERP.

I have been noticing one recent shift in the SERPs that I do like. I do a lot of math research, and Wikipedia often has been dominant. So I use the footnotes on the Wikipedia page to find more in-depth references. Recently, I'm seeing Wolfram Alpha outrank Wikipedia more often. I like it - they are truly expert in math AND IN COMMUNICATING about math.

Certainly math professionals have written the Wikipedia content, but they aren't doing much except throwing out their information. Wolfram Alpha is a lot better for researching a particular math field.

I do like seeing something improve about Google Search - I wish I saw more ;(

[smithaa02]

"Heh, I saw a perfect example of this. A site which TRIES to pass itself off as the "OFFICIAL TOURISM BUREAU" for a city (but defintiely is NOT), was apparently manually reviewed and fooled the reviewers and now occupies the top spot for that city search. "

Yeah, Mike this is a great example of how the vital system is messed up. A visitor bureau or chamber of commerce shouldn't even be considered vital to start with (unless somebody is doing a search for that organization). Frequently, these types of organizations are nothing more than cabals that take tax payer money and really only help member businesses (and especially members who happen to be on the board of directories)...they're not about regions or cities, but selfish self-promotion. There are a number of quasi-official institutions out there that can cause problems for reviewers. Take say Fannie Mae and the keyword mortgage...imagine some clueless reviewer thinking they were a government institution and giving them the keyword 'mortgage' as being vital...

[netmeg]

See now, I'm heavily into the local govt sector, and a lot of users and publishers would strongly disagree with you about government sites and chamber of commerce/visitor bureau sites not being vital to a search for the city or location. By no means is that a commonly held opinion.

For every result we think is ridiculous, there are plenty of people who think it isn't.