Forum Moderators: Robert Charlton & goodroi

Message Too Old, No Replies

Google Changes Treatment of PR 'Saved' by rel=nofollow Sculpting

         

Robert Charlton

9:13 pm on Jun 3, 2009 (gmt 0)

WebmasterWorld Administrator 10+ Year Member Top Contributors Of The Month



It's being widely reported that the Google crawling team (as best as I can assign responsibility) has changed how it's treating PageRank previously "saved" by the use of the rel="nofollow" attribute.

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.

signor_john

1:38 pm on Jun 10, 2009 (gmt 0)



As I've argued elsewhere, allowing proliferation of NF would break the google algo, at all levels. Google needed to check this trend, and now they have.

I think you're overstating things a bit. PageRank is only one element in Google's algorithm, the use of nofollow for PR sculpting is a long way from being universal, and--if anything--the excessive use of nofollow could actually be useful to Google as a "signal of SEO."

They will not lay out the effects of NoFollow, as it will depend on circumstances.

Sure, just like other things may depend on circumstances. Determining search rankings is like creating a perfume: One or two components of even the finest perfume may smell unpleasant in isolation, but it's the overall blend that counts.

Shaddows

2:34 pm on Jun 10, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



I think you're overstating things a bit.

Maybe a slight overstatement- but not that much.

Concepts such as "trust" and "authority" rely on working links. Then there's the "semantic" (for lack of a better word) association of pages through anchor text. Most of the value and innovation Google has brought to search is based on data garnered through link analysis.

They will not lay out the effects of NoFollow, as it will depend on circumstances.
Sure, just like other things may depend on circumstances. Determining search rankings is like creating a perfume: One or two components of even the finest perfume may smell unpleasant in isolation, but it's the overall blend that counts.

I don't disagree. All I'm saying is Google doesn't ever clarify anything as soon as it depends on deployment, or is context specific. Especially as they have no reason to stick to the same contextual rules in perpetuity. They prefer grandiose Kant-like imperitives that they have no need or intention of ever revising.

[edited by: Shaddows at 2:35 pm (utc) on June 10, 2009]

freejung

9:56 pm on Jun 10, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



JS Harris, that's excellent info, exactly the kind of empirical data we need to evaluate this. However, I find your "grandfather" theory hard to believe -- that would mean that the ranking algo used to rank a page would have to depend on when the page was first crawled and last changed, and G would have to keep track of exactly which algo elements were in place when each page in their index was created. Not that it would be impossible, but it seems like a lot of pain and expense for questionable benefit.

From G's standpoint, what would be the benefit of this? It might be more "fair" to webmasters who created their sites under a particular set of assumptions about G's rules, but it would actually reduce the quality of the search results (assuming G believes that their recent changes improve quality, which presumably they do) by applying outdated algos to certain pages and not others. That sounds pretty implausible to me.

It seems much more likely that this whole thing is, as Shaddows suggested, FUD.

I hope you update us with your results.

Giacomo

10:13 pm on Jun 10, 2009 (gmt 0)

10+ Year Member Top Contributors Of The Month



Hi! Since there's so much confusion (and lack of official info) on this subject, I just wanted to let you know we've had a first unofficial confirmation from a trusted source of what Matt Cutts said at SMX Advanced. Here's what an Italian Quality Search Team member said during the Q&A session of a Google-hosted live webmaster chat event today:

Q.: will the PageRank of page C (linked from page A without rel=nofollow) increase as a consequence of adding the rel=nofollow to the link from A->B?

A: Hi, there is no consequence for C due to adding a nofollow to the link from A to B, hope that is clearer now

[edited by: tedster at 10:41 pm (utc) on June 10, 2009]

Critter

10:20 pm on Jun 10, 2009 (gmt 0)

10+ Year Member



That's Gold, Jerry, Gold!

Hissingsid

9:10 am on Jun 11, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I've read all the way through this thread and to be honest I'm still left wondering what I should actually do. I have nofollow links to utility pages on each page of my site.

The two suggestions that I'm considering are:

Tedster

One common need in sculpting is to keep certain "utility pages" from soaking up link juice, since they aren't much good in the search results. You know the ones - "log-in", "my account", "shipping rates", "privacy" and all that.

Most sites want to show those options on every page, for their user's convenience. But that means those pages get a lot of internal link connectivity - they easily can end up with very high PR. So I often create a single page containing all those utility links and display that URL through an iframe. Now there's only one URL with links to those utility pages.

annej

I just put

<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.

But which is the best option?

Cheers

Sid

syed

10:34 pm on Jun 12, 2009 (gmt 0)

10+ Year Member



If we use nofollow for some links, then other links will not gain any value? Meaning, no extra juice for dofollow link if I use nofollow for the others?

Ok, in that case, the only way increase PR transfer power for other links is reduce the number of useless links itself as otherwise we are just wasting PR - which basically as you mention going into a 'blackhole' or no where.

Let me give you an example. I have a widget site and I have 4 pages. Out of these 4 I have only one page that I want ranked high - say "top widgets" page and the rest three pages I dont care about in terms of rankings and those are the pages that I'd rather not pass the PR link juice to. Lets say those are "About us" "Support" and "FAQ" pages.

So in order to give more link boost to important pages like "top widgets", I should just remove those three link pages and integrate into one and say call it "About us/Support".

In my opinion, instead of wasting 3 links PR value (3 link pages that are useless content wise anywa), now we have only 1 page with PR wasted..so other useful page gets a lot more PR juice.

What do you guys think..will this work?

[edited by: tedster at 7:36 pm (utc) on June 13, 2009]
[edit reason] remove specifics - use "widgets" instead [/edit]

JS_Harris

12:09 am on Jun 14, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Update from my previous posts in this thread - replacing the nofollow tags did not result in restored rankings. The old nofollow links appear to have been grandfathered and by removing them I lost any sculpting benefit they provided permanently.

As far as I'm concerned this means nofollow is officially on the scrap heap at Google, it's completely ignored. I suspect it was inevitable, Google needs full control of rankings and tools like nofollow gave webmasters too much power.

I'm done wasting time with nofollow, linking freely seems to get the best results on new articles. If you have links with nofollow i'd suggest never touching them because you may not be able to put them back on with any effect again for sculpting purposes.

[edited by: JS_Harris at 12:13 am (utc) on June 14, 2009]

signor_john

2:57 am on Jun 14, 2009 (gmt 0)



As far as I'm concerned this means nofollow is officially on the scrap heap at Google, it's completely ignored.

But is nofollow on the scrap heap as far as Google is concerned? Google certainly hasn't given us any indication that such is the case.

tedster

3:03 am on Jun 14, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I'd say it depends on what you mean by "the scrap heap". The rel="nofollow" attribute still performs it's original purpose -- the link sends no juice to its target, and it protects the location where it appears from a penalty for linking to a bad neighborhood.

Giacomo

9:37 am on Jun 14, 2009 (gmt 0)

10+ Year Member Top Contributors Of The Month



Yep, nofollow is still pretty useful in a number of cases. Many have made the mistake to think that if it can't be employed for PageRank sculpting anymore, then it's not worth using it at all.

Anyway, yesterday Matt Cutts twittered that he's working on a "PR sculpting blog post", so we'll have more official info soon about how Google handles nofollow, and hopefully some advice on how/when (not) to use it.

savweb

1:59 pm on Jun 15, 2009 (gmt 0)

10+ Year Member



protects the location where it appears from a penalty for linking to a bad neighborhood

I don't think so, because user doesn't know whether it is nofollow or not, he just clicks.

tedster

4:55 pm on Jun 15, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



The user isn't protected - you're right. I mean that the site owner is protected against a Google penalty.

Hissingsid

5:24 pm on Jun 15, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Hi Tedster,

I've removed nofollow from 3 of my sites today using your suggestion of a page in an iframe.

I think I will do the same with my main site tomorrow. My only issue is if it will validate as it uses xhtml1-strict.dtd. The ones I've done today are transitional. I could always change it to transitional I suppose.

Cheers

Sid

tedster

5:36 pm on Jun 15, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



It would be good to hear from you about any results, Sid. Threee months after deploying this on one site I've seen a very real improvement in which Sitelinks are chosen. Search traffic to those newly chosen sitelink pages is up significantly, even on queries that do not trigger a sitelinks display.

savweb

11:44 am on Jun 17, 2009 (gmt 0)

10+ Year Member



The user isn't protected - you're right. I mean that the site owner is protected against a Google penalty.

Not exactly i think, because if user isn't protected, why webmaster should be ? Google is for users, not for webmasters. If a user can go to bad website via nofollow link, why should google ignore this fact? I mean that one of the possible dark sides of "nofollow" could be invented as something like a secret weapon against seo's. Didn't anybody ever thought about it this way? I don't think that google is revealing all the truth.
This is just my opinion.

Shaddows

12:02 pm on Jun 17, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



The point of nofollow is so that webmasters can withold their PR vote from user generated content (UGC), or from a site mentioned in a context such as
"There is a horrible scraped MFA site over at EvilScraper.com"

SERP manipulators spam UGC for backlinks. NoFollow means the blog/UGC/whistle-blower content owner doesn't get blown out the water for linking to poor neighbourhoods.

This IS how NF works, its what G invented it for.

whitenight

12:16 pm on Jun 17, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I don't think that google is revealing all the truth.

This is smartest comment in this thread. Thank you!

--------------

Come on, people. THINK!
It's already "been around" for a whole year, according to MC. [webmasterworld.com]...

<stares at watch to see if anyone makes a logical deduction based on this fact>

Oh wait, MC already said "blah blah blah"
"Yay!"
That means I don't have to THINK as a webmaster.
I can just blankly believe what he said.
Thank you Matt.

see 50+ other earlier rants about MC "pronouncements", testing, and logical deduction for more info...

raschidt

8:41 pm on Jun 18, 2009 (gmt 0)

10+ Year Member



< moved from another location >

Internet tends to copy the experience in the real word, where any body can suggest (by mouth to mouth) any good source, this is like a link, but if we use the “nofollow” attribute, this is like say:

I like coffee from “the corner`s shop” but don`t listen to me.

Or we refer to something or no.

In internet we have the option to refer somebody (good source) but say something like (don`t believe or not follow my vote or not listen to me) … That`s not good, that`s breaking the normal rules.

I understand in Internet any search engine can place his own rules, but one of the goals of search engines is offer a real (or natural) experience to users, if that is the case, the best way to offer a real experience is “reading what people is writing” in other words, following the links like a real person will do.

What do you think?

[edited by: tedster at 9:34 pm (utc) on June 18, 2009]

graeme_p

11:58 am on Jun 19, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



The point of nofollow is so that webmasters can withold their PR vote from user generated content (UGC), or from a site mentioned in a context such as
"There is a horrible scraped MFA site over at EvilScraper.com"

UGC, definitely. I am not so sure about withholding votes from bad sites. For example, suppose I am writing a blog post to rebut another blog post on a political site I disagree with strong, and therefore want to give a ratings boost to. Now from my point of view, it is tempting to nofollow the link to the post I am responding to. Form Google's point of view, if it is worth responding to, I should be giving it some link juice.

Pico_Train

3:07 pm on Jun 19, 2009 (gmt 0)

10+ Year Member



Why would you link to a horrible MFA site at evilscraper.com in the first place? Waste of time, code, bandwidth and everything else.

signor_john

8:40 pm on Jun 19, 2009 (gmt 0)



Why would you link to a horrible MFA site at evilscraper.com in the first place?

It isn't what you'd do, it's what someone writing a blog comment or a forum post might do.

johnhh

9:14 pm on Jun 19, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



whitenight has hit the nail on the head

Once you get into the "Google is everything" mindset it’s dangerous. Very dangerous.
So take a step back.

Major search engines agree on "nofollow ="

Webmasters use it.

Goggle realises "hey hold on we can't have this, webmaster deciding what sites are trusted and what sites are OK, lets do something about it - it may alter the SERPS and we may miss something”.

Next you know Mr Google says, " we have changed this " effectively diminishing the effect of "nofollow" , and by the way "we did this some time ago"

WebmasterTools = benefits Google , Adsense = benefits Google, AdWord = benefits Google. People blindly following Mr Cutts = benefits Google.

Stand back - develop sites people want and need, do the graft, do the content, continually improve for the user, do the time. The influence of Mr Google suddenly comes less.

signor_john

11:27 pm on Jun 19, 2009 (gmt 0)



Stand back - develop sites people want and need, do the graft, do the content..

You must be talking about Yahoo, where the graft is collected by the "Search Submit" or PFI team. :-)

Hissingsid

9:50 am on Jun 20, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I have to agree that Matt Cutts has two main tools at his disposal.

1. The algorithm

2. What he says

I think that he is the sort of guy who would think it was really cool for him to have such powerful control that what he says leads to a massive reaction around the web. It is much easier for him to say that they are tracking something and the algo is "filtering" it out than to actually change the algorithm. That way 90% of the "problem" goes away and Google engineers don't have to do anything.

Tell people that they are just p*$$ing their precious page rank up the wall and they are going to react quickly.

If this was ever proved then half of his armoury goes so I think that there always has to be a germ of truth in what he says.

Cheers

Sid

whitenight

5:37 pm on Jun 20, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



If this was ever proved then half of his armoury goes so I think that there always has to be a germ of truth in what he says.

... if only someone or some people had done some testing with rel=nofollow within the past few months we could prove the truth of this situation..

<stares blankly again at his watch>

Wait! There IS a thread where people tested results with rel=nofollow within the past few months...
on this forum no less.

<following rant pre-censored for PG-13 audiences>

I wonder if that highly public thread and its conclusions and this announcement have ANYTHING to do with one another?

Naw, Google would never use the Law of Confusion to allow webmasters to doubt their OWN TESTING AND CONCLUSIONS.

lol how many licks does it take to get to center of a Tootsie Roll Pop?
The world may never know...
cause no one could ever test this without biting into them after 1...2...3...licks, could they?

robzilla

7:52 pm on Jun 23, 2009 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Pardon my wild and slightly uneducated guess, but could it be that links with "nofollow" are treated in a way that is similar to how Google handles, or, I should probably say, used to handle, dangling links, as outlined in the original PageRank document? To minimize the impact of links carrying the "nofollow" attribute on the website, and perhaps even the web as a whole, they could be temporarily removed from PageRank calculations. The document states:

Because dangling links do not affect the ranking of any other page directly, we simply remove them from the system until all the PageRanks are calculated. After all the PageRanks are calculated they can be added back in without affecting things significantly.

The difference being that the PageRank calculated to be passed through those links carrying "nofollow" is dropped.

This 147 message thread spans 5 pages: 147