| 5:02 pm on Dec 10, 2008 (gmt 0)|
|I'm sure use of the nofollow attribute was Google's way of finding most, if not all of the SEO wantabes. |
lol you're just saying this right?
I could see you saying this when it first came out ala the EFV thread debate we had over 2 years ago, but NOW?
The most well-respected, all-authority, juggernauts use rel=nofollow AT LEAST on a few pages.
|If you are using it for anything other than what it was originally intended for, I say there are some associated risks. |
I agree here. I see far too many people using it to "sculpt" who have little clue what they are doing.
We even had an "accident" story in the last update thread.
Potentially dangerous, but in the right hands. It's simply poetic.
| 5:04 pm on Dec 10, 2008 (gmt 0)|
This has become a very dangerous thread for the non experienced.
Better results can be had by interlinking your best articles together while lowering the number of links pointing to articles of lower importance.
That all depends on the site and market. Think long tail.
Keep your index pagerank high with fewer links leaving it and interlink your articles to help lower category page PR, that's my best advice.
For many sites, your 100% right.
All this talk of using nofollow to accomplish what I've just described is like bringing an elephant gun to a rabbit hunt. When you know how to use nofollow you'll find yourself barely using it because your sites structure will already be strong.
Regardless, you just described an excellent PR sculpting strategy, whether done with actual links or nofollow.
I was going to sit this thread out, but this was one of the most valuable posts on PageRank sculpting I think I've ever read. For certain types of sites, you did an excellent jobs of breaking down the complexities into a few key points.
Personal experience has shown me it works, and belongs in the toolbox of any SEO who works on high PR sites...but like a saw, it can cause severe issues if misused. The cooking with salt analogy is good.
| 5:05 pm on Dec 10, 2008 (gmt 0)|
Am i allowed to post this?
www . stonetemple.com/articles/interview-matt-cutts.shtml
Please people! Read through this interview with matt cutts and all will become clear...
edit: the interview actually has a webmasterworld link in it!
| 5:19 pm on Dec 10, 2008 (gmt 0)|
|Only those who know SEO. I'm sure use of the nofollow attribute was Google's way of finding most, if not all of the SEO wantabes. If you are using it for anything other than what it was originally intended for, I say there are some associated risks. |
These days the Wall Street Journal uses it, eBay uses it, Zappos uses it. [On internal links that is]. Like I said, it has hit the mainstream.
| 5:21 pm on Dec 10, 2008 (gmt 0)|
i used it with some success a little more than 2 years ago to distribute more page rank to the home page; the homepage pr was @ 4 the rest were @ 5; i set up no follows to have every page link back to the home page and 'no follow' every other page. the homepage pr where i had my conversion form went up to 5 and the rest of the pages dropped to 4; i believe it helped in rankings, for as soon as the pr increased, so did the page positioning move up. the pr evened out after about a year and a half across the entire site, and i don't think it mattered anymore because we performed a redesign without much loss in ranking.
i think this level of internal linking was noticed before; i'm not so sure if it works anymore.
| 5:28 pm on Dec 10, 2008 (gmt 0)|
Let's think this through and I'm going to use examples of implementations I've seen where it left me wondering how all this works.
So you use
rel="nofollow" at the link level for the granular flow of PageRank™. Let's use an About Us page since that seems to be the page that gets most treated like a step child.
We have all these internal links marked as
rel="nofollow" to the /about page. At the /about page level, there are no other mechanisms in place to keep that page from being indexed and/or the links followed from that page. You've basically "stopped" the flow of PR to the /about page internally. But, you've not stopped any external PR from hitting that page, am I correct?
So the /about page gets no internal PR but gets external PR? What is wrong with that equation?
Now, if I use
<meta name="robots" content="noindex"> at the /about page level, that keeps it out of Google's index and from what I'm seeing over multiple test pages, it keeps the page from obtaining visible PR. I know Matt Cutts states that a noindex page accumulates and passes PR but based on my real time observations at this very moment, I don't think the accumulation of that PR is shown at the TB level. Every page I have that has the noindex meta element has gray PR, every one. If they have accumulated PR, it must be an internal version. And yes, I surely want the links on those pages to be followed in most instances so I'm only going to use noindex.
Here I am debating the use of nofollow and I've never experimented with that granular level, ever. Working at the page level has been successful for our use. We'll slap
<meta name="robots" content="noindex"> and/or
<meta name="robots" content="noindex, nofollow"> on /search /checkout /terms /privacy etc. We surely don't want our /about /contact etc to be noindex, that would be counterproductive in many campaigns.
And that robots.txt file is not a mechanism to keep pages out of the index. I still don't like the fact that URI listings are shown for robots.txt entries.
|Only those who know SEO. I'm sure use of the nofollow attribute was Google's way of finding most, if not all of the SEO wantabes. |
Heh! I'm here to keep the sparks flying. Ya'll are going to have me testing this crap here soon and I'm going to join the SEO Wantabe group. :)
P.S. In doing site: searches for those pages marked as noindex, they are not to be found. Wouldn't you say that effectively removes the page from the equation and is probably the best option?
[edited by: pageoneresults at 5:42 pm (utc) on Dec. 10, 2008]
| 5:35 pm on Dec 10, 2008 (gmt 0)|
|Ya'll are going to have me testing this crap here soon and I'm going to join the SEO Wantabe group |
lol you know you wanna.
Imagine the almighty power of a Fully Operational nofollowed pageone site! mwahahahahah.
| 5:43 pm on Dec 10, 2008 (gmt 0)|
But again, there are 2 parallel discussions here.
1) PR flow- putting PR where you want it
2) STOPPING PR getting to pages you dont want indexed.
1) is best acheived through structure
2) is best acheived at page level
BradleyT, all your points about blocking are addressed in 2
In addition to 2, you can nofollow links, sure. But its not a substitute for 1. Or it shouldnt be.
| 6:03 pm on Dec 10, 2008 (gmt 0)|
Im not so sure breadcrumb navigation links within content should pass PR, this could be seen as over-optimization. I know Google recommends breadcrumb. I'm testing this at the moment.
| 6:14 pm on Dec 10, 2008 (gmt 0)|
Big egos aside, some very interesting commentary on this site. Thanks to all who have contributed.
| 6:23 pm on Dec 10, 2008 (gmt 0)|
|Matt’s answer was: nofollowing your internals can affect your ranking in Google, but it's a 2nd order effect. |
That comment from Matt was recorded by Dave Naylor back in 2008 March.
Based on some of the replies I've seen here so far, there are a few who have figured out the "ranking" part of it, bravo. :)
|Search engine robots can't sign in or register as a member on your forum, so there's no reason to invite Googlebot to follow "register here" or "sign in" links. Using nofollow on these links enables Googlebot to crawl other pages you'd prefer to see in Google's index. However, a solid information architecture - intuitive navigation, user- and search-engine-friendly URLs, and so on - is likely to be a far more productive use of resources than focusing on crawl prioritization via nofollowed links. |
Emphasis mine. I look at it from the "solid information architecture" side of things.
Maybe I should join that Wantabe Tribe and see what all the rage is about? But, if I've achieved nirvana in site architecture, do I even need to invest the time for this type of granular implementation?
That page above in Google's Webmaster Help Center doesn't give me the warm fuzzies at all. You've got instructions mixed in there with words like untrusted content and paid links, what am I to think?
Every time I get heavily involved in these discussions, I start to have second thoughts. Yes, I'm starting to think that maybe I'm a little behind the times on this damn
rel="nofollow" attribute and its use. At this point, I don't see a need for it. You still haven't convinced me, not that it matters.
And no Google, I am not joining your
rel="nofollow" cult movement just yet! I think my way is working just fine and apparently you agree too. I just don't want to mess with things that don't seem to be broken from my perspective. Maybe I need to hire one of them SEO Wantabes? I could probably do it myself. I may qualify, huh? :)
| 6:32 pm on Dec 10, 2008 (gmt 0)|
I'm with both camps on nofollow ;)
Generally speaking, nofollow shouldn't be necessarily. IMO, a site with the ideal information hierarchy, planned interlinking and intuitive structure is unlikely to need nofollow. That said, I use it as a workaround for:
- Structural problems (i.e. not linking to the right pages in the right places)
- Technical problems, i.e. linking to "bad" URLs like Google's infinite calendar example
For me, it's another tool in the toolbox. But it's a proprietary attribute for an individual HTML element - i.e. an undesirable workaround rather than a solid foundation. If I'm involved in the creation of a site from the ground up, I usually don't want stuff like nofollow in there. Just as I don't want to be ticking preferred domains and suchlike in WMT rather than permanently redirecting.
One exception to all of the above would be untrusted, user-generated links - but I'd prefer vetting those to be an editorial task rather than a technical one.
| 6:35 pm on Dec 10, 2008 (gmt 0)|
Hopefully I didn't come accross as having one of the ego's, this subject is frustrating to discuss with someone just learning the intricacies of micro-managing internal pagerank.
I just wanted to point out that Shaddows is right, two subjects are getting meshed here. Micro-managing internal pagerank with nofollow is done to help achieve better indexing for the "meat" of your site but it has nothing directly to do with blocking pages from the search engines.
#1 - If you want to block a page from being in the serps use robots.txt
#2 - If you want to block pagerank from flowing to that page AND want it blocked from the serps use both a robots'txt exclusion and nofollow every link to that page. Why? A robots.txt exclusion doesn't stop your links from passing pagerank to it.
#3 - If you want to limit pagerank from building up on a page THEN you can micro-manage internal pagerank in a variety of ways such as by reducing the number of links pointing to that page or by having that page link out to more pages or by inter-linking more important pages and doing nothing to it at all. (there are in fact many other ways thanks to nofollow, those three being the primary).
To be clear - to accomplish #3 you don't need to physicaly remove links, that's where proper use of nofollow can be used.
I think Peterdaly sums it up best with...
|If you can't run your own PR flow simulations, it's very likely you may just be spinning your wheels at best |
edit: about #2 - isn't using nofollow on every link pointing to a page enough to keep it out of the search indexes? No, another site may link to it creating a path for search engines to follow which result in it being indexed. Your site has an internal structure but that structure is still part of the web.
[edited by: JS_Harris at 6:41 pm (utc) on Dec. 10, 2008]
| 6:39 pm on Dec 10, 2008 (gmt 0)|
|P.S. In doing site: searches for those pages marked as noindex, they are not to be found. Wouldn't you say that effectively removes the page from the equation and is probably the best option? |
I have to make a correction. In doing some further searches, a "few" of my test pages are now showing for site: searches and they have a white bar. Arrrggghhh! It's almost like someone from G follows me and whenever I make statements about something, they do something to contradict me! I need to go back and review the trail and see what is going on that I maybe overlooked. It is no big deal, they were pages specifically set up to test this whole PR flow thing at the page level, not at the link level. Most are still gray and only a few have gone white.
I really hate being wrong. But, I can admit to being wrong when I am. I still don't think I am though. ;)
|I'm with both camps on nofollow. |
Man, I'm about 10 feet away from the fence. But, there's Razor Wire up there. :)
| 6:46 pm on Dec 10, 2008 (gmt 0)|
Good post Harris! I was just thinking about a few of those points. I think some of the confusion can come from the tool bar, which does go grey on pages that have been blocked only with noindex or robots.txt... this is where Google are misleading us some what!
| 6:47 pm on Dec 10, 2008 (gmt 0)|
|I have to make a correction. In doing some further searches, a "few" of my test pages are now showing for site: searches and they have a white bar. Arrrggghhh! |
pageoneresults, give me a list of your pages that are fully nofollow'd but not excluded via robots.txt or noindex meta tag and I'll see if I can get them a nice green pr by linking to a few of them from MY sites :p
nofollow does not equal noindex
| 6:56 pm on Dec 10, 2008 (gmt 0)|
I think I've responded a few too many times here, i'm passionate about this subjectm but I want to leave you with some food for thought.
A question for someone extremely comfortable on the subject: Have you ever considered using nofollow on EVERY link to your BEST page? If it has enough backlinks from other sites to push it's value to #1 in the serps you're just wasting your internal pagerank by linking to it anyway. It doesn't get better than #1, I hope that illustrates the potential power of micro-managing internal pagerank for you.
| 6:59 pm on Dec 10, 2008 (gmt 0)|
|pageoneresults, give me a list of your pages that are fully nofollow'd but not excluded via robots.txt or noindex meta tag and I'll see if I can get them a nice green pr by linking to a few of them from MY sites. |
I don't have any pages that are "just"
<meta name="robots" content="nofollow">. I only use the
<meta name="robots" content="noindex"> or
<meta name="robots" content="noindex, nofollow">. I've not had any reason to allow a page to be indexed but nofollow the links from that page.
I work with content that is strictly controlled editorially and there is no need for that
rel="nofollow" attribute. Or I should say, at least I thought there wasn't a need for it.
Also, I'm well aware that pages listed in robots.txt do obtain visible PR. I've never been fond of how the SE's treat the indexing of paths found via robots.txt files. It's also a roadmap for many and one that I try not to use for "granularity".
Just change the architecture to be more intuitive and avoid this whole micro management thing to begin with. Ya, I know, not a solution for many. Here come the bandaids... ;)
| 7:10 pm on Dec 10, 2008 (gmt 0)|
On the side noindex discussion, robots excluded pages accumulate PR as far as I'm aware - hte content is excluded, not the links to it. Which is why nofollow on links has a value of its own.
I think micromanagement is quite a good term to use related to nofollow though, with the negative connotations that word has - having to firefight attributes on individual links is not the kind of activity I like to be doing ;)
| 7:23 pm on Dec 10, 2008 (gmt 0)|
Just jesting about the big ego comment - it certainly would not apply to you anyway - but some real good posts by you - tangible and concrete rules that I agree with. I am in the midst of some internal link re-achitecting, to reduce the amount of link juice flowing to my "superfluous" pages, making judicious use of link-based "nofollow". My issue is if Google will recognize these "nofollow" directives, after the fact - these pages had been given direct link contact (so-to-speak) - and have been indexed. I really don't mind them being indexed, as there is some remote added value to searchers, but the limit of link juice flow is a selfish concern of mine, that I hope to remedy.
| 7:38 pm on Dec 10, 2008 (gmt 0)|
Good luck and let me know how it goes doughayman.
Best tread lightly and test for effect.
| 9:59 pm on Dec 10, 2008 (gmt 0)|
I wrote a pretty lengthy article on this back in Sep... my consensus is was that it's pretty much a stupid thing. PageRank is a probability value for which pages are most likely to get clicks based on visibility (# of links, # of links going to the pages that do the linking, etc.) rel=nofollow doesn't prevent PR from being "leaked", because it doesn't modify probabilities in any way (the links are still visible).
I also made the point that it perverts linking analysis because stingy Webmasters use it on all links, even the ones that otherwise make a ton of sense. e.g. forums and blogs.. sure, spam gets submitted, but the links that people provide throughout the thread are invaluable to the core topic matter. They should be given credit.
For internal links, the first point is the most crucial one... you don't "divert" or modify PR in any way, so it's superfluous to use it.
[edited by: tedster at 11:16 pm (utc) on Dec. 10, 2008]
| 10:14 pm on Dec 10, 2008 (gmt 0)|
|It's almost like someone from G follows me and whenever I make statements about something, they do something to contradict me! |
And then there's this...
Using meta tags to block access to your site
|When we see the noindex meta tag on a page, Google will completely drop the page from our search results, even if other pages link to it. Other search engines, however, may interpret this directive differently. As a result, a link to the page can still appear in their search results. |
Emphasis mine. What does the bolded part mean to you?
Isn't the end result of all this to remove a page or group of pages from the index so they are not occupying valuable space? And in that process, I thought PR was removed from the equation? Not the passing of PR, but the accumulation thereof.
I know we have two somewhat different but related topics taking place here. My feeling is that if you just remove the destination page from the index, you've solved the challenges from all aspects and don't need to worry about the micro-management of your own internal links, would I be correct in that observation?
You know, I can be really dense at times. I am trying to understand this like you are explaining it to me. Someone probably needs to give me a few Mark Harmon's (multiple slaps upside the head) before it really sinks in. :)
|Learning About PR Sculpting: internal links with rel=nofollow |
A perfect title for this split thread because I am surely learning!
| 10:33 pm on Dec 10, 2008 (gmt 0)|
|"Google will completely drop the page from our search results, even if other pages link to it" |
What does [this] part mean to you?
To me, it means: "don't include this URL in search results". Not necessarily the same as "don't include this URL in your algorithms".
Remember that a noindex robots meta element requires that search engines spider the page (which means it gets stored). It's a declaration that you want your document to be kept out of search results - but not out of spidering and indexing schedules.
I don't like the above definition myself, since I'd prefer noindex to mean "don't index". But then, I also think nofollow is a poor name; it seems to mean "don't count" rather than "don't crawl/index". I imagine programmers rather than end-users come up with these names. Don't get me on to the name of the "cache" ;)
| 10:46 pm on Dec 10, 2008 (gmt 0)|
It really boils down to being about math -
Page A has y amount of pagerank and it distributes it equally through x number of links. Nofollow reduces x. [obviously add in dampening and recursion]
So if you have 100 links on a PR3 page and you nofollow 1 of them it won't really make a difference.
| 11:27 pm on Dec 10, 2008 (gmt 0)|
|So if you have 100 links on a PR3 page and you nofollow 1 of them it won't really make a difference. |
Bigger picture, the page is PR3 with 100 links, does it rank 1st in search results? If it does you won't nofollow anything, in fact the page makes a good candidate to add more links to and still retain it's #1 rank, assuming you have other related pages to link to. (100 links is pushing it but that's a different topic)
So, that being said, do you have any other relevant pages that could use a PR3 link that also provide reader value to the topic on that donor page?
You'll be using nofollow on pages that "almost" rank first to give them that final bump and the page you mentioned wouldn't mind losing the link, match them up!
| 11:48 pm on Dec 10, 2008 (gmt 0)|
I'd agree that it's maths, but it's complicated maths, and we aren't privy to a number of the calculations.
Page A passes an amount of PageRank based on the amount it has received. This is distributed amongst the pages it links to.
There are just three variables mentioned there, and in a generalised way:
- How much PageRank a URL has received
- How much PR a URL passes on
- How this PR is distributed
I think it's pretty safe to say that the calculations determining any of the above are not known, and to some extent are unknowable - they're the results of a system, rather than part of the criteria for that system.
But while SEO is maths for sure, for me it's more about judgement calls. Is this change likely to deliver the required benefits, as compared to the other possible changes? That's maths too, but I think fundamentally an opinion is required. Otherwise, I'd spend all day doing maths ;)
| 12:23 am on Dec 11, 2008 (gmt 0)|
"Page A has y amount of pagerank and it distributes it equally through x number of links. Nofollow reduces x."
As the stonetemple link above makes clear, and many people continue to ignore and ignore, this statement is not correct. Or at least, Google has never said it is correct. What Matt says in the article suggests pagerank dies with nofollow.
| 10:51 am on Dec 11, 2008 (gmt 0)|
|As the stonetemple link above makes clear... in the article suggests pagerank dies with nofollow |
Not the way I read it. All Matt seems to say is that there is good reason to use nofollow at the link level to avoid sending PR to boilerplate pages, as noindex is more a filter from SERPs than anything else. Which is not necessarily how I do things, and may require a change. If you just noindex the page, you get the detriment of PR drain from your hierarchy, but lose one potential landing page from an odd search.
Maybe I've misunderstood you, but you seem to be saying that a page with
PR value = V
Total number of links = x
Nofollowed links = n
V/x per link if all are followed
V/x to each link that is followed regardless of n
This would be in contast to the intuitive case (given the drop-from-linkgraph statement) of:
V/(x-n) per link
If the first case is true, then there is NO VALUE in PR sculpting using NoFollow as PR simply disappaers mid-link rather than being retained for distribution though other links.
If the second is true, the value of PR sculpting goes down as n decreased as a proportion of x and/or as V increases.
| 12:12 pm on Dec 11, 2008 (gmt 0)|
We've had a number of justified warnings in this thread about knowing what you're doing before you start playing with internal rel="nofollow". I'd like to describe a pitfall in a bit of detail.
There's one factor that many people don't think about, and that's when they shoot their own boot. When you cut off PR from a url, then you also give that page no PR that it can flow on to other pages. If you get indiscriminate with your wannabe scuplting, it's something like applying tourniquets all over your body. Yes, you may not want to see this or that url getting PR, but does it help other urls rank that are "downstream"?
I was asked to look at one established site that had taken a plummet in the rankings after an attempt at sculpting. I used a browser plug-in to highlight nofollow links and their home page almost turned completely red! They were trying to channel PR into just two directories from the home Page, but then on internal pages they were trying to channel PR to two or three other directories. In addition, they had paginated results pages and were trying to send more POR to the deep pages by nofollowing links to the first 5 pages in the pagination.
The more I studied it, the more my head hurt. This was way crazy and far beyond not sending PR to your shippiing Info page. They had essentially disrupted circulation for what had been a decently structured site.
Now there are some ways to deal with pagination issues through PR scuplting that can make sense for some sites. And there are ways to help weaker directories get some juice. But you'd better know your site well - and know well how PR circulation works - before your start throwing the attributes around. You're not just affecting the PR that flows to one url, you're affecting every url that's downstream from that target, too.
| 12:20 pm on Dec 11, 2008 (gmt 0)|
|A question for someone extremely comfortable on the subject: Have you ever considered using nofollow on EVERY link to your BEST page? If it has enough backlinks from other sites to push it's value to #1 in the serps you're just wasting your internal pagerank by linking to it anyway. It doesn't get better than #1, I hope that illustrates the potential power of micro-managing internal pagerank for you. |
This is exactly what is done on a site I am involved with (a site of several million pages). We have sub-sites for several key locations in our country that have great PR on their own without getting link juice from the main super-site. In these instances instead of "over egging the mixture" we pass PR back UP through the directory structure so it can get used more sucessfully elsewhere. Lesser performing sub-sites with lower Page Rank get passed the link juice instead, this way we even have individual listing pages (which are far down the directory structure) with significant Page Rank.
| This 211 message thread spans 8 pages: < < 211 ( 1 2  4 5 6 7 8 ) > > |