I guess it does. Funny, though, I can't see how this lines up with "we drop nofollow links from our webgraph." Probably the backlnks data is kept separate from the web graph that's used to calculate PR and so on. Hard to see this as an intentional choice - maybe it's more of an oversight?
It is possible the site is cloaking and generating same links but without nofollow param for Googlebot IPs.
I am inclined to think that Google would never drop such links from webgraph despite claims to the contrary - there is value in itself to see what proportion of backlinks to a site were marked with nofollow. Yahoo definately keeps those as backlinks (though whether they pass any ranking value is an open question, probably not passing though).
Why would they do that? If it was the other way around (only show nofollow for Google Bots) I could understand the reason.
|It is possible the site is cloaking and generating same links but without nofollow param for Googlebot IPs. |
If the link is indeed paid for, then it would be rather odd to make it (for the bot) nofollow, however it might make sense to make it marked as nofollow for reviewers and competitors who look at their site trying to figure out what's going on. Just a theory :)
Google Webmaster Tools shows nofollow links (like those from Wikipedia) so it's not too surprising that link search might as well.
"All these nofollow links are showing on a Google backlink check, does this mean that the backlinks function on the search engine does not necessarily list "algorithm qualifying links" anymore, just any site that links to to that site?"
Matt Cutts addressed this issue before on Sphinn. Just because a nofollow link shows up when you either run a link: query or look inside Webmaster Tools doesn't mean its counting.
"Drop from our webgraph" seems not to mean the link data is deleted from Google's database or is never recorded. I'd assume that nofollow link records are flagged in some way, so that Google's ranking algorithms ignore that data.
|does this mean that the backlinks function on the search engine does not necessarily list "algorithm qualifying links" anymore... |
Google has never shown "algorithm qualifying links" in it's backlink search. Never. Some of the links displayed may count, some of the links displayed may not count. And some of the links not showing in a backlink search may nonetheless be counting.
Formerly it showed PR 4 and up links, but that didn't mean those were the "algorithm qualifying links" either. It has always been a sample. ;)
|Probably the backlnks data is kept separate from the web graph that's used to calculate PR and so on. |
I've always understood that what they displayed in a backlink search was different from the data used to calculate ranking. If that's the case then showing no-followed backlinks makes sense, considering it is a results page for a query of sites that are linking to a particular web page. It's an accurate result.
Google Search has never been an accurate ranking or SEO tool because of this, including the inurl, allinanchor searches etc.
This subject was also discussed in this thread back in October....
Why is Goog giving me credit for links in WMC when they have no follow
I agree that there's no reason to assume that the links are being credited just because they're showing up the Google backlinks list.
What continues to be puzzling though, is, if you take Matt Cutts' quoted comments about nofollow links literally...
|...we don't even use such links for discovery.... the nofollow meta tag does that same thing, but at a page level." |
...then you've got to wonder how the links appear in the index at all. Not "even... for discovery" is a pretty strong statement... and, as I note in the previous discussion, "the nofollow meta tag is what you use when you want to prevent Google from indexing the link itself."
For the original discussion about Matt's comments on nofollowed links, and some very interesting tests by Miamacs, see this thread...
Matt Cutts: Further Clarification on the nofollow Attribute
The closer you look, the more inconsistent this all becomes.
Some people in #*$! forums are discussing that they have seen improvement in rankings even with nofollow tag for anchor text that was used.
>>improvement in rankings even with nofollow tag for anchor text that was used.
Which is really immaterial. All it means is that there was *some* factor that caused an improvement in rankings that could (and more than likely does) have nothing whatsoever to do with anchor text in an IBL.
|I can't see how this lines up with "we drop nofollow links from our webgraph." Probably the backlnks data is kept separate from the web graph that's used to calculate PR and so on. Hard to see this as an intentional choice - maybe it's more of an oversight? |
An oversight, or quite possibly the presence of the links on the pages may be getting used as one of the factors in evaluating the score for some metric for the pages the links are one, without even giving credit to the linked_to site in any way. In that case, there could be a different process for a different reason that may not involve webgraph use as it relates to PR or anchor text.
[edited by: Marcia at 4:03 pm (utc) on Jan. 11, 2008]
The way to check this would be to use some pretty unique anchor text in a link marked with nofollow and then see if this anchor text got associated with the target page to which that link was pointing to.
A PS to comments above. I was just discussing nofollow links on another thread, and this thought occurred...
|One thought that comes to mind is that the links may have been indexed before they were nofollowed, and perhaps hadn't been purged from the backlinks index at the time they were seen. |
Robots Meta Tag & nofollow
So I'd ask internetheaven to watch these backlinks over time to see if they disappear. The problem with Google backlinks, though, is that you only see a random sample of those that Google has indexed, so this wouldn't be a conclusive report by any means... but if all the nofollowed backlinks disappear it might indicate something.
Sorry to dredge this up, but I didn't want to start a new thread.
I have a back link showing in WMT, "What Googlebot Sees" for a link text I only used once as a test for a NO FOLLOW back link on a hightrust site.
So why, it they aren't even supposed to be used by G for discovery are they showing up in its own webmaster tools. I have checked the code and the no follow is correct.
I am not sure what in what order these links are reported in WMT, but said no follow link is in the midst, with normal links above and below. Strange eh?
|So why, if they aren't even supposed to be used by G for discovery are they showing up in its own webmaster tools. |
soxos - Several of the threads I reference above discuss this very question. In the Robots Meta Tag & nofollow thread cited, phranque refers to Matt Cutts' response to this question...
|he mentions in the interview that at one time there was a strange set of circumstances that could have caused it but they considered it a bug and therefore fixed the problem. |
don't remember if he mentions the time frame...
I have seen that before, then I look at the source code of the site and find some errors like a missing <> or a </> tag or some other html major conflict where I could see that googlebot might not be able to work through the html error.
|...for Google, nofollow'ed links are dropped out of |
our link graph; we don't even use such links for discovery...
- Matt Cutts
I think some of this has to do with a misunderstanding of the wording used by the people at Google. The quotes I have read do not say they are removed from the index. (There might be one, but I haven't personally seen it. The one above is from Google Groups.)
(I know this thread is not about noindex, but I think the situation might be similar.) If you read the information about noindex on a page, the page is 'not in the index', is 'not returned in the results', etc., but if you read some of Matt Cutts' comments on other sites a 'noindex' page can and does pass inbound link weight, as long as the page is not also tagged as 'nofollow', so even though the page cannot be found in the results (index), Google has the page and uses the information.
I think the terminology confuses people, because they think 'noindex'='dropped from the index' and interprets to 'Google does not store a copy of the page', which is reasonable, but what it really means is 'Google does not *show* the page in the results'. Period, nothing else. Nowhere do they say they do not retain a copy or use it in their calculations... It's actually the opposite. Matt Cutts clearly states a page with noindex can accumulate and pass PR.
I would guess some of the inconsistency, or what seems to be an error, has to do with the Google definition of 'dropped from the link graph' compared with what the general understanding might be.
Dropping a link with 'nofollow' in it from the crawl is probably a completely different process than dropping it from PR calculations (although nofollow accomplishes both), and the link: command probably accesses a completely different result set than the crawler and PR calculator access.
(Added: The preceding makes a little sense, because if the crawler and PR calculator accessed the same set of data either the crawler would crawl sites with inbound ROS links at a ridiculous rate, or the PR calculator would miss a large number of links. One function only needs 1 links to each page. The other needs multiple repetitions of links to each page. It could be filters, but in any case, the end result is the same. Basically, they all have different information they access.)
So, I think it is reasonable to conclude 'nofollow' does not keep a link from being returned in the index (results), but does keep it from passing weight.
Quotes in '' ('Text I typed') are paraphrase or general understanding and not specific to anyone. The initial quote is directly from Google Groups.
With regard to the above discussion, note Matt Cutts' recent comments about why nofollow links show up in the webmaster console... quoted in this thread...
Does Google honour rel="nofollow
|Matt commented on this during G's webmaster chat a few weeks ago: |
|we don't follow nofollow links, but we do include nofollow links in the webmaster console just to give webmasters a full picture of all the stuff that's linking to them. |
> we do include nofollow links in the webmaster console just to give webmasters a full picture of all the stuff that's linking to them.
|we don't follow nofollow links |
Careful choice of words? Is that the only thing they don't do with them then? Any quotes that they are not part of the algorithm at all, even slightly?
this is still the case after all this time.
one of our sites now has a site that (with nofollow) links to us in the top 10 of the backlinks. hmmm.
Kristos, see previous post by potentialgeek:
|> we do include nofollow links in the webmaster console just to give webmasters a full picture of all the stuff that's linking to them. |
|martinbuster Google has never shown "algorithm qualifying links" in it's backlink search. Never. Some of the links displayed may count, some of the links displayed may not count. And some of the links not showing in a backlink search may nonetheless be counting.[/b] |
It makes you wonder why Google provides the links: search at all. What possible use can it have for Joe User and what possible use can it have for a webmaster. Its like a little bit of misinformation mixed with a pinch of cr*#.
Edit reason: Missed a closing tag
[edited by: Hissingsid at 10:44 am (utc) on June 12, 2008]
|It makes you wonder why Google provides the links: search at all. |
It's more help to the inquisitive end user who is looking for related information, than it is for competitive webmaster research.
|It makes you wonder why Google provides the links: search at all. |
The link: search is a very effective mechanism for Google to use to mark your computer as belonging to a webmaster, differentiating it from that of a typical web user. This helps them in determining the value of your clicks and pageviews, etc.
I would imagine they will continue to support this function as long as it's an accurate signal to them, it has nothing to do with it being an accurate tool for us.
|The link: search is a very effective mechanism for Google to use to mark your computer as belonging to a webmaster, differentiating it from that of a typical web user. This helps them in determining the value of your clicks and pageviews, etc. |
Checked the search pages, no extra information is marked onto my computer. And what about the hundreds of thousands of webmasters working from home? Do clicks from all webmaster's family and friends irrelevant too? What about people with blogs looking to check who is linking to them? How many blog owners are competitive webmasters?
The only way for them to track this would be through cookies which are difficult to keep on someone's computer or by IP address which changes constantly. The toolbar maybe? But then you've got my last paragraph to contend with ...
No, can't think of a way this data could be reliable even 1% of the time let alone 99% which is what it would have to be to be part of any algorithmic shift.