Forum Moderators: open
The idea came to mind looking at Google's Webmaster Tool, and checking the internal link structure. I noticed that because of the nature of these pages (i.e. each has a link from all the other pages) they could appear quite 'important'.
What are you thoughts on this?
That said, I do control linkjuice flow for ALL of my sites, and client sites too. So we use js or robot.txt to disallow extraneous content.
Matt Cutts just the other day noted something to the effect that it's good to do that in some cases, and all of the SE's at various times have alluded to it in conferences, etc.
Examples of situations where we use it:
- "email this" pages, "print this" pages, and similar pages
- images
- certain kinds of user generated content
- affiliate redirect pages
Easiest thing in the world to do. Avoids dup issues, strengthens linkjuice flow to important pages and avoids splitting PR into pages that no one wants to see rank.
We also pay attention to deployment of footer links with this in mind. Not all footer links need to be, or should be, site wide.
Matter of interest, since you do "channel" PR, and assuming your internals use KWs in their anchor text, do you think the value of the anchor text is PR weighted?
Agree completely. Also, when I re-read le_gber's post I'm not sure the comments I offered line up with the question.
We don't control linkjuice flow to footer/corporate pages with js or robots.txt. I agree that that would be dicey. So that we just do with link quantity. Not every page of a site has to link to the About Us page.
Our goal regarding comments above is to cut off PR flow to pages/elements that the SE's don't want to rank, and that are essentially useless for SERP's listings. That is the sort of thing that MC, Tim and others have endorsed too, I belive. No body wants to see "print this page" pages in the index. Yet those pages account for massive amounts of wasted PR around the Web.
Re your other question, I've always assumed that the power of the link is reflected in the way the anchor text is valued but I've never tested it. For example, if a page had only two inbound links, and assuming both links were comparable (placement on their pages, number of outbounds etc.), if the link from the PR 6 page said 'red' and the link from the PR 2 page said 'blue' I'd assume that the linked-to page would have a better chance of ranking for 'red.' Is that what you mean?
Also, when I re-read le_gber's post I'm not sure the comments I offered line up with the question.
well your comments actually line up pretty well :) It's exactly what I had in mind - PR flow through the site and PR 'wasted' on 'useless' pages.
caveman, do you always do this from the start / launch of the site? Have you ever done it after the site was live and did you notice any 'boost' coming solely from this linkjuice redistribution?
I've seen lots of pain associated with site wide changes over the past several years (nav, wholesale changes to page titles, etc.). But all I'm talking about is excluding from view pages that the SE's don't want to feature anyway. The result is just more juice behind the pages that matter. No sitewide nav changes are involved with excludiing, for example, "print this page" pages since those are invariably single links from single pages. It simply causes a recalc of the power of the remaining links. I view it as akin to sweeping the dead leaves off of the lawn. Who wants to look at that, when what's left is so much nicer? :P
Sort of, I've been mostly ignoring PR assuming that anchor text calculations don't take it into account, so a PR3 link with "widget" anchor text is as valuable as a PR6 link with "widget" anchor text.
If the anchor text value is dependent on the PR value, it would certainly make me re-appraise things :-)
There are some things about the algos that I feel I need to know, others I just make assumptions about. This has either long been more the latter, or maybe I inferred it from the papers and accepted it as an operating assumption. Been a while since I reviewed the older patents etc.
The other thing is that this can no longer be just a PR-ish sort of discussion. Context and quality measures now matter a lot IMO. Well, with G anyway. So all the more reason to view the anchor text as weighted according to page factors.
I do know, for example, that text on a linking page can influence rankings of the target page. So, would text from an important page that links to me have more influence than text from a relatively unimportant page?
Funny, every once in a while a question comes up that makes me realize I've been operating on an assumption for so long that I know longer recall precisely how I got there. Hehe.
