Forum Moderators: Robert Charlton & goodroi
Matt Cutts mentioned the change at SMX Advanced last night. Excellent summary in Search Engine Roundtable [seroundtable.com] post.
I'm sure there's going to be a lot of howling over this change.
The concept of "wasting" PR implies that the PR could be put to better use. If NF detracts from the sum of PR anyway, you are not not wasting PR. As any mathmatician, logician or linguist will tell tou, two negatives make a positive; you are wasting PR.
Worse still, you are losing it completely. At least with 'wasting' it to a log-in page, you can recylce it to other pages through navigation.
As such I agree with JS_Harris that MC is either inconsistant, disingenous, wrong or a combination of all three. If an NF link consumes PR, you CANNOT waste PR by sending it to a login page. You could in fact CONSERVE it.
MC is either inconsistant, disingenous, wrong or a combination of all three
lot of G's old friends suddenly woke up in this thread .keeping it animated ..and lot of new posters with bad bing experiences zoom in and begin posting in other threads . ..interesting :)
g announcing / or not that they are playing with tags ..is them again trying the ( search engine ) tail wagging ( web) the dog ..trick again ..and trying / or not to get the naive to build sites for exclusively for g ranking and not for users and all engines in general ..
Clearly now, what part are you not getting?
You can either send PR to the page, or not. That's it.
You are wrapping yourself in a wildly illogical pretzel because your "idea" of waste makes no sense. And it's hard to even grasp how you get to that point.
Matt has been pretty darn clear from the begining, but webmaster FUD has lead people to people saying "black is white" so much that they believe it.
PR dies if link are on a meta nofollow page. PR dies if you link through a robots.txt blocked page. As Matt said, same with nofollow. Despite him saying that, repeatedly, clearly, the FUD machine ignores what he says.
You can either send PR to a page or not. If you don't want your privacy policy outranking topical pages, use no follow... or as Matt has said, repeatedly, don't link to SEO-unimportant pages so much. Use your navigation wisely.
What he has never, ever said is if you have ten soldiers of PR to distribute, and you have ten links off a page, that if you nofollow five then the other five get 2 soldiers each. That idea has NEVER been stated by Google, ever. But the exact opposite idea, as quoted above, has been stated. PR dies.
And it makes perfect sense. if you have 100 links on the page, each one can pass a 1% share of the passable PR of the page; or, if you use one of the nofollow tools, you can prevent PR going to pages. This has been what they have said, and it is pretty darn close the the least interesting thing ever said, but it gets blown up into something it is not by people who refuse to read what Google says and instead translate it into what they want to hear.
(Of course, what Google actually DOES can be different that what they SAY. I suspect they have experimented with how they look at nofollow, but their public statements have always been consistent, that is, nofollow kills pagerank on a granular level just like other tactics can kill it on a page level.)
I've re-read MC quotes. I can force myself to read it such that PR is dropped, but I can also read it that NF just makes the link "disappear". Thats the LINK that dissapears, not the PR. In fact, I just cannot understand why he talks about power-users, PR sculpting and PR flow if it never did anything of the sort.
However, on a very careful analysis, I have to say HE NEVER SAYS ANYTING DEFINITIVE EITHER WAY. All is open to interpretation.
-----------
Critter, see g1smd above for obfuscated JS links. But otherwise, yes. Using NF wastes PR, otherwise it is happily pushed to another page where it can be recycled and re-used
the natural result of this policy announcement will be to remove outbounds from sites completely. Expect a algo rebalancing so that outbounds to relevant sites give contextual ranking boosts (query-dependant).
Makes sense to me.
Another thought: Google should consider treating "nofollowed" internal links differently from nofollowed external links--especially since, for AdSense publishers, it now requires a privacy policy or a link to a privacy policy on every page. What's the harm in allowing Webmasters to "sculpt" PageRank internally since paid links, linkspam, etc. don't enter into the equation and there's no benefit to users in having mysite.com rank #1, #10, or #100 for the phrase "privacy policy" or "about us"?
Google? Can we have a "No Crawl" option along with a "Use XML File(s)" choice? That would end all this nonsense. Let us "tell you" what we want you to index. You've given many of us job security with your indexing routines. Maybe it's time to fix them?
I'm still wondering about all the others who support the rel="nofollow" attribute. We're not talking just Yahoo! and Bing, Bing, Bing. There are a plethora of tools built around that attribute. Not to mention a few who have an entire business model focusing on PR Sculpting. I wonder what their feeling right now?
Just give me the option to have you "No Crawl" our sites and instead take an XML feed of what "we want" you to have. It's only fair.
You do know that any changes Google make to it's guidelines from this point forward will of course be subject to much scrutiny. I'm sure most of us will just avoid many of the Google specific directives if there is "too much white space" between the lines. Let's bring the line-height down to about 135% instead of the current 200%. ;)
8. Nofollow to control internal PageRank flow
Search engine optimization professionals started using the nofollow attribute to control the flow of PageRank within a website. This is an entirely different use than it was intended originally. Nofollow was designed to control the flow of PageRank from one website to another. However, some SEOs have suggested that a nofollow used for an internal link should work just like nofollow used for external links.Several SEOs have suggested that pages such as "About Us", "Terms of Service", "Contact Us", and "Privacy Policy" pages are not important enough to earn PageRank, and so should have nofollow on internal links pointing to them. Google employee Matt Cutts has provided indirect responses on the subject, but has never publicly endorsed this point of view. [18]
The practice is controversial and has been challenged by some SEO professionals, including Shari Thurow [19] and Adam Audette. [20] Site search proponents have pointed out that visitors do search for these types of pages, so using nofollow on internal links pointing to them may make it difficult or impossible for visitors to find these pages in site searches powered by major search engines.
Although proponents of use of nofollow on internal links have cited an inappropriate attribution to Matt Cutts [21] (see Matt's clarifying comment, rebutting the attributed statement) [22] as support for using the technique, Cutts himself never actually endorsed the idea. Several Google employees (including Matt Cutts) have urged Webmasters not to focus on manipulating internal PageRank. Google employee Adam Lasnik [23] has advised webmasters that there are better ways (e.g. click hierarchy) than nofollow to "sculpt a bit of PageRank", but that it is available and "we're not going to frown upon it".
No reliable data has been published on the effectiveness or potential harm that use of nofollow on internal links may provide. Unsubstantiated claims have been challenged throughout the debate and some early proponents of the idea have subsequently cautioned people not to view the use of nofollow on internal links as a silver bullet or quick-success solution.[citation needed]
More general consensus seems to favor the use of nofollow on internal links pointing to user-controlled pages which may be subjected to spam link practices, including user profile pages, user comments, forum signatures and posts, calendar entries, etc.[citation needed]
Note to mods- I'm happy to go and link all those refs, but not if the post will just be deleted. Sticky me if its good to stay
[edited by: Shaddows at 2:47 pm (utc) on June 4, 2009]
<META NAME="robots" CONTENT="noindex, follow">
on the utility pages like contact, search, etc.
Since they have the same universal links to main parts of my sites as the rest the PR should go back to the site and not be dead ended on an info type page.
As far as link sculpting is concerned linking to related pages within the text makes more sense.
I thought noindex'ed pages actually accumulated pagerank?
So say I have 5 links on a page, one to my privacy page. I don't care about it ranking, but I want to push as much PR to the other 4 links a possible.
Previous consensus was to NF the Privacy page link.
Now, that would waste PR. Deleted, dead, lost.
Why would you do that when you can allow the PR to go to the privacy page, which ITSELF HOSTS INTERNAL LINKS, preserving the PR within the site, rather than losing to the Ether.
PANIC
Ignoring MC is dumb. Believing MC is dumber. But there's nothing wrong with comparing pronouncements with results. And thats more easily done with informed discussion, which I suggest this thread resembles more than "panic"
WAS IT
a) to allow webmasters to manipulate/conserve PR?
b) to allow sites that sell links (for PR) to cheat customers?
c) to allow user-generated outbound links without risk of penalty (for linking to bad guys)?
Strangely enough, the answer is c). So, nofollow now works the way it should have originally!
The big news here is that it took Google several years to get something mind-bogglingly simple right (assuming they have now got it right).
Kaled.
With this change, the entire web will now sort of "leak" PR in a way that it didn't before. Wouldn't this create problems with respect to normalization of the link matrix?
To put this another way, one classic way of describing how PR works is the "random surfer" model, so your PR is the probability that a surfer clicking on random links throughout the web would land on your page. But with this change, every now and then the random surfer will just get stuck and stop (he tries to follow a nofollow link but can't, and he's now apparently not allowed to follow the others either, so he's stuck and can't click on a link), meaning that the total probability of the random surfer going anywhere on the web will not add up to 1 (having probabilities add to one is normalization). My understanding is that there was always a little bit of decay, that is, the random surfer always had a small chance of falling off the link graph at every step, but that chance was the same for every link and every page. Now it will be totally different depending on how many nofollow links there are.
Doesn't this pretty much break the model, or at least change it radically? Does anyone with some knowledge of matrix algebra care to comment on this aspect?
G has a lot of math talent and I'm sure they could find a way to do it, but it doesn't seem like something they could just do offhand without some really deep analysis and possibly revamping their core algo. Would it be worth all that effort on their part just to stop a few SEOs from having a slightly easier technique to manipulate the flow of PR, which G surely knows they can find a way to do anyway?
This whole thing makes little sense to me.
Anyway, far from undermining the original algo, this move will guarentee its continued existance, for reasons I outlined earlier.
Like I said in my first post in this thread, I am REALLY HAPPY with this development, I'm all for it. Linking does what it's meant to. PR hoarding (my pet hate) is broken. Its just that this is an ACTUAL U-turn (IMHO) in policy.
Outbounds will be less prevalent, but this can easily be balanced by algorithmically adding some credit for downstream pages. Webmasters linking to relevant topics could then be rewarded.
The big problem is Blogs, which can now be taken down by link-spamming. Sheer volume could mean infintesimal equity to pass through internal structures.
There's a very simple solution for this. I'm almost ready to pull the trigger myself.
Find what: rel="nofollow"
Replace with:
I have not used a nofollow tag since G changed the original purpose (comment spam) that the three SE's first announced its use for.
I am not aware of any negative results. YMMV.