| 11:10 am on Feb 28, 2012 (gmt 0)|
Potentially the age of a link as a factor could be depreciated. I would guess that as a factor, it means less as time goes by as the amount of broken links or "editorial supervision" of old links would dirty the data set. When was the last time you went through old blog posts to check the links are all still relevant?
Given that social media properties and link building activities have escalated in recent years, links have become less about a vote for a site and more of a passing, frivolous activity in many ways.
That said, if it was in any way a significant factor, turning it off would probably have a notable impact on existing rankings for a lot of sites and seeing we haven't noticed that, it's either unlikely or not that important a factor.
| 11:54 am on Feb 28, 2012 (gmt 0)|
| 4:13 pm on Feb 28, 2012 (gmt 0)|
Position of link within a page's content. In the first 200 words our outside that towards the bottom of a page; a link should be worth the same value.
| 9:30 pm on Feb 28, 2012 (gmt 0)|
In my opinion it should be PageRank, as google has always mentioned that it is in violation of the google TOS to manipulate the PageRank in anyway.
By turning that signal off google can be assured about there is no effect on SERP of a website if he managed to get some high PR links and also there will be no benefit to the websites inflating the PR by doing anything.
In many cases a back link from a low PR page is much more important than a back link from a high PR page. So evaluating the back link with respect to PageRank becomes irrelevant.
| 9:37 pm on Feb 28, 2012 (gmt 0)|
HTML Signals with 95% surety. It's so easy to game and so many DID game that it only makes sense to discount it.
| 9:54 pm on Feb 28, 2012 (gmt 0)|
Surprised nobody has said "Anchor Text".
| 10:09 pm on Feb 28, 2012 (gmt 0)|
Do you mean internal anchor text?
| 10:10 am on Feb 29, 2012 (gmt 0)|
What about 301 redirects or some aspect thereof (i.e. just non homepage redirects, just external redirects)? Been gamed for years with people buying up old / expired sites for the link equity.
They could effectively maintain some of the functionality of redirects via rel=cannonical and site address change via WMT, without needing to use them as PR conduits.
| 10:14 am on Feb 29, 2012 (gmt 0)|
What happens to legit sites that move from one domain to another? How much collateral damage would occur from 100% depreciating the 301 redirect?
| 10:21 am on Feb 29, 2012 (gmt 0)|
I think people are frequently skipping over this bit:
'We often rearchitect or turn off parts of our scoring in order to keep our system maintainable, clean and understandable.'
What kind of evaluation would potentially muck up a system - age scoring, recip tracking, something like that? Especially when that particular method might anyway have been superseded by another change (for example, random recip link powerhouses might anyway get dinged by some quality factor while real shares between authorities might be worthwhile to score).
| 10:25 am on Feb 29, 2012 (gmt 0)|
Well, we're not talking about depreciating a 301 entirely - just it's ability to act as a signal to determine what the topic of the target page is.
The assumption would be that 99% of website owners will be smart enough to maintain the same content / similar site structure with their new site (either same domain with different URL structure or new domain entirely).
All Google needs to do is pass a top level PR value to the homepage of a domain (can even be based on total IBLs, even to deep pages) and let the site's internal structure dictate how it is distributed throughout the site. 301's don't need to play a part in that process.
At some point within the site move process (which takes months via WMT), Google runs a process that matches old page to new same page internally (same mechanism that detects dup content) and could attribute deep link PR, etc that way.
Just thinking out loud here, but it seems to me that webmaster driven 301's were an easy, low resource way for Google to manage this process, but flawed because it can be misused. The question now is Google's technology / ability / processing power sufficient to do a better job internally?
Purely from a search point of view, what value is there in relying on a 301 that I (the webmaster) can control? Is it always used appropriately? Can it be misused? What's the margin for error there? Can I achieve a similar or better effect internally?
Just speculation of course. :)
| 11:16 am on Feb 29, 2012 (gmt 0)|
|'We often rearchitect or turn off parts of our scoring in order to keep our system maintainable, clean and understandable. |
@Stever - This is the clue IMO. Something of low relevance that causes conflict with those objectives, and is done in an easier manner.
Could it have something to do within the promotion of Panda's capability? It understands context and the quality of a page sending a link, as well as the landing page and site of the one receiving it.
| 11:22 am on Feb 29, 2012 (gmt 0)|
|Could it have something to do within the promotion of Panda's capability? |
That would support the 301 theory - less need for historical information as they have an increased ability to judge live content on the fly.
| 12:29 pm on Feb 29, 2012 (gmt 0)|
|Do you mean internal anchor text? |
I was referring to anchor text in general.
| 1:29 pm on Feb 29, 2012 (gmt 0)|
I think that googlebot will be able to determine ( if not already ) if the placed link is surrounded by other outgoing links ( and if so are these links related to the topic of the page) or no.
| 6:10 pm on Feb 29, 2012 (gmt 0)|
|That would support the 301 theory |
301 redirects are depreciated already, so that they pass a reduced amount of PageRank. Web publishers change internal URLs during upgrades of technology or updating their pages to new versions. If the 301 redirect is no longer sending PageRank then that will mean a website has a disincentive to upgrade their technology or update their web pages.
I agree with what Whitey suggested, "Something of low relevance that causes conflict with those objectives, and is done in an easier manner." Does the 301 theory qualify as a low relevance signal?
One also has to wonder what was behing the depreciating PR of 301's in the first place.
| 8:52 pm on Feb 29, 2012 (gmt 0)|
In order to remove some of these potential metrics, I think it is important to think like a search engine.
In order to determine relevance, freshness and potential value of a link Google needs to determine at a minimum:
Theme of Page upon which link is placed
Location of link on page
Context of link in copy
Metrics / value / authority of said page
Age of link in relation to copy on page
Paid implications of text (editorial, sponsored etc)
Number of outbound links on the page
Nature of link (direct, 301)
Google is trying to reduce false positives.
The largest false positive right now in my opinion is anchor text. Google relies on it heavily.
My personal feeling is that all of the other metrics above are required in order to understand relevancy and context.
Anchor text is the easiest piece of the link puzzle to manipulate so it would stand to reason they would want to eventually discount that signal entirely.
However, my devil's advocate half would ask why there has not been a large shift in rankings to reflect that change if it were indeed live.
Perhaps, if this is implemented, it would be done over time, in stages.
| 9:42 pm on Feb 29, 2012 (gmt 0)|
|...my devil's advocate half would ask why there has not been a large shift in rankings to reflect that change if it were indeed live. |
That's an interesting proposal. It has always seemed to me (opinion, not fact) that a keyword rich anchor text is going to be used most often by an SEO, and less likely by someone making an innocent citation.
Is it possible that current spam-fighting parts of the algo (like detecting paid links, etc.) already knock out the sites gaming the anchor text? If so, then what's left innocently making anchor text citations might be a group so small that removing the anchor text influence might have a negligible effect on the algo.
[edited by: martinibuster at 12:07 am (utc) on Mar 1, 2012]
| 11:48 pm on Feb 29, 2012 (gmt 0)|
|and less likely by someone making an innocent citation |
Provided these come from authoratitve sources , like a recognized new article, they are going to be more credible and would likely pick up on the surrounding text and validation signals.
Over optimized anchor text [ internal and external ] is probably an irritant to Google and it's quality guidelines, since it is not natural linking.
But i doubt if either of these were touched in this update.
| 12:09 am on Mar 1, 2012 (gmt 0)|
|Is it possible that current spam-fighting parts of the algo (like detecting paid links, etc.) already knock out the sites gaming the anchor text? |
My sense is that if Google was to discount that metric, then all other metrics above would be needed in the absence of link anchors in order to determine properly, at least at this time, accurate value of a link.
|Over optimized anchor text [ internal and external ] is probably an irritant to Google and it's quality guidelines, since it is not natural linking. |
Definitely agree. We have no real idea as to whether this applies to internal or external links, or all, seeing as they are both treated differently.
From what I gather part of Google's quest in link determination over the last few years has been about handing out less penalties and more about skipping, or discounting links that look to be trying to game the system.
Removing anchor text would only leave relevance and page quality / authority, and (perhaps soon) socials signals like authorship and community (which would likely de-clutter the data Google needs to save (in order to make room for more social signal analysis?).
It would further allow Google to not penalize pages that over optimize inbound anchors, since that would no longer be a factor in rankings.
Instead, those sites would simply not rank well anymore. or gradually fall off the map.
Just a theory :)
| 9:40 pm on Mar 1, 2012 (gmt 0)|
My pure out of the blue wild guess is that it has more to do with the words in the text surrounding a link.
I think that google is probably relying more on an overall content evaluation of a page, instead of trying to evaluate the immediate text preceding and following an inline text link.
This would probably help reduce a lot of data, and would probably help to leverage whatever info is gleaned from Panda updates (google has stated that they are going to integrate panda more with the other part of the Algo).
again, this is a wild guess, so take it for what it's worth.
| 2:21 am on Mar 2, 2012 (gmt 0)|
|The title tag has been used to help identify the meaning of a linked page from the general theme of the page linking to it. |
Funny. I would interpret it the other way around: the title of the linked page can reinforce-- or contradict-- ideas about the linking page's subject.
And then... What about links that have no anchor text?
#1 Highly regarded .gov or .edu page that says "for more information try these sites" and gives a bare list.
#2 List of naked URLs in a page with no content worth mentioning.
Not the same thing at all is it?
Or, for that matter,
#3 Highly regarded .edu site put together by someone who never read The Books, so anchor text says merely "this page" or (shudder) "click here".
:: still waiting for g### to wrap their brains around the idea that some "links" simply aren't links at all except in the meaninglessly broad sense of "ways to get to your page" ::
| 6:55 am on Mar 5, 2012 (gmt 0)|
I've checked my sites this morning. I get the impression that sites with more than 25% of the inward links featuring only exact match keywords or keyword phrases as the anchor text are the ones which have moved down the search results.
| 6:26 pm on Mar 5, 2012 (gmt 0)|
Might be nit picking here but this was not part of the Panda update, but part of 40 other changes in the early March update.
| 8:37 pm on Mar 5, 2012 (gmt 0)|
why are everyone talking about 301, i dont think thats a issue at all. What I think they have changed is the value of the link text, means you have a lot of link text "blue widget" that looks non natural for google and thats bad for you, thats it. Also not always believe what google says, they know who they reach when they post such changes, things can still work like always.
| 8:42 pm on Mar 5, 2012 (gmt 0)|
Tryting to read this carefully.
It's about "characteristics" of a link to tell us "what the linked page is about"
This would indicate it's about accuracy of what the page is about rather than simply a scoring factor for search positioning. I agree that something to do with anchor text is most likely. Could be to do with anchor text which is the same as the title of the site linked to, but that anchor text may not be relevant to the specific page linked to. Just an idea.
| 9:32 pm on Mar 5, 2012 (gmt 0)|
When I originally read this point in the Google blog post, my first instinct - and hope - was for external anchor text. Something they have been using 'for years', which directly affects Google's perception of the 'topic' of the page, and which has been abused by link buyers/builders for years.
Does the fact that there hasn't been a collapse in rankings across the board, and widespread complaints, suggest that Google had done a reasonable job already of identifying all the worthless links?
| 9:42 pm on Mar 5, 2012 (gmt 0)|
I am almost sure it is the anchor text. (Could be other signals as well) This month the link building has generated almost zero results for exact match anchor text. While something already ranks well for "alpha beta" we were unable to influence the rankings for "beta alpha" on the same url.
This came as a bolt from the blue. we had seen a decline in influence from cheaper links since january but it gets tough now. This month was a killer blow. No benefits at all.
| 10:29 pm on Mar 5, 2012 (gmt 0)|
|We often rearchitect or turn off parts of our scoring in order to keep our system maintainable, clean and understandable. |
Link analysis method that is redundant?
I was thinking that they're turning off one method of link analysis because it's redundant with other methods and thus no longer needed.
Link analysis has too many moving parts?
Or is this signal making their system less understandable or maintainable? If so, does this mean it's a signal that is more or less minor and even superfluous?
What do you think is the case?
| This 66 message thread spans 3 pages: 66 (  2 3 ) > > |