Improving ranking with internal linking architecture

[Broadway]

Take the following premise (I know the example is rediculous, sorry about the length):

Say I have a website about making stuff from trees: making maple syrup, making pine cone art, collecting sassafrass roots. Everything on the site is about tree products but very few visitors looking for maple syrup information ever venture into the pine cone art section, and vice versa. (Even if they do they'll probably just exit from there. Navigating to the website home page doesn't do much for traffic or sales. The focus is keeping the person in the sub section they initially entered into - 90% of traffic is from SERP's).

The site linking architecture is as expected. The home page (hub) mentions all of these varied subtopics in passing and then links out to each one's main page.

As far as Google SERP's go, it's the main page of each of the subtopics that rank best along with the subtopic's internal pages, not the website home page. As far as traffic goes, the website wouldn't be drastically damaged if it lost all of the home page traffic (it's a long-tail site).

So, I've been looking at my internal linking architecture.

Right now I use hard coded breadcrumbs on each and every internal page (assume a pure "drill down" with no cross-silo linking at any level):
You are here > Home Page / Subtopic main page / Subtopic internal page - level 1 / Subtopic internal page - level 2

From my interpretation of the math, in terms of PageRank, this type of architecture tends to focus the clout of the shear number of pages that a website has up towards the Home Page and Subtopic main page levels. So I'm sending "juice" to my website home page, where I don't really need it (the home page's ranking isn't important to generating traffic).

I was considering instead to set up the linking architecture where:

The website home page links to the main page of each subtopic, AND vice versa. But the breadcrumb navigation on internal pages stops at the subtopic main page level.
You are here > Subtopic main page / Subtopic internal page - level 1 / Subtopic internal page - level 2
(meaning that the highest level that subtopic internal pages link to is their topic's main page)

This linking structure (once again, according to my understanding of the math) focuses more pagerank (from the clout of the shear number of pages that a website has) at the subtopic main page level (and not nearly as much gets transferred on up to the website home page). I was assuming that this would help the ranking of the subtopic main pages (at least to some degree).

As far as user experience goes I have an above-the-fold iframe that shows that I could tuck a link to the website home page in on each internal page, so as far as "expected navigation for the visitor" goes, nothing has to be significantly compromised.

So, having laid this out, clearly I must be overlooking several factors. Why isn't this a good idea? How will this linking architecture be counter productive?

[tedster]

Is there no other link to the home page anywhere except in the breadcrumb trail?

[dertyfern]

Thought navigation and footer links were somewhat depreciated. If so, this wouldn't really matter from a PR funneling point, just usability.

[Broadway]

Tedster: No, I had really intended to significantly reduce (and funnel) linking very strictly.

Per your comments dertyfern, maybe working with pagerank in this fashion is passe enough that there is no purpose to go to the effort.

[TheMadScientist]

Why not nofollow the links like you would with a 'nothing' page instead of playing iFrame games?

Seems like a simple, easy, 'no tricks' solution.

[Broadway]

Doesn't the nofollow solution leak the very PR I'm trying to conserve? It's not that I don't want PR flowing to the Home Page, it's that I don't want the Home Page boosted at the expense of pages I'm more interested in seeing rank.

[Broadway]

In regards to 'iframe games.' In terms of sculpting, it seems one of the few solutions that still exist (nofollow rules have changed, Flash can be indexed, javascript links can be followed).

Of course, related to the changes that have occurred with other sculpting techniques, possibly it's just a matter of time before the rules change and iframe content is associated with the page it appears on.

That would suggest that the only long term solution to sculpting is limiting site-wide navigation options.

[TheMadScientist]

Doesn't the nofollow solution leak the very PR I'm trying to conserve?

So does the iFrame...
The source is credited with a link AFAIK.

AFAIK, you are actually leaking more PR to the iFrame and not passing it back to any other pages than you would be to the home page. Also, the redefinition of PR and NoFollow says you will not pass more weight to other pages, so basically the same number of links = the same amount of PR being passed through each, so you would not be passing more to inner pages, but you would be removing the PR from the home page, which could then be replaced by more specific pages in the results.

You're trying to sculpt, and the best method is really to go get inbound links to the pages you want to have the PR increased on and cut down the number of links (including the iFrame) on the page...

I certainly wouldn't impair visitor navigation to try and manipulate the results or PR, because I think you're trying to over-manage the situation, but I might effectively remove the home page from the results if it does not matter, then watch to see if it's replaced by specific inner pages, which with personalization could IMO help the inner pages rank overall.

Another thought I would consider is removing the home page link from the deep pages... So, sub-sub category pages don't link to the home page, but rather link to the 'home page' of the section, then the section category pages and the sub-category pages could link to the home page... This would probably keep you more focused within a section and IMO if someone lands two sections in they're probably looking for something very specific and wouldn't be too bothered by not being able to see the main 'all categories' (home) page.

[dertyfern]

Another thought I would consider is removing the home page link from the deep pages

Do you write this assuming the home page is garnering most of the inbound links? Great suggestion, I may do it myself, but would be concerned about loosing rankings on the home page.

[TheMadScientist]

Do you write this assuming the home page is garnering most of the inbound links?

An unnecessary amount of internal link weight being passed to the home page is more specific... Broadway said the Home page rankings could be let go to increase the rankings of the inner pages, so we were talking about ways to keep the PR off the home page and I was suggesting other ways of trying to get the inner pages to have increased visibility, basically, without directly sculpting the PR to the inner pages but rather away from the home page.

The latter suggestion does sculpt the PR, but IMO it still keep a quality visitor experience and doesn't use anything which could be seen as 'hiding' or 'questionable', again, IMO. There's also no telling what Google will change in the future about the handling of iFrames, and I don't like changing things, because I think stability and consistency are important... A bunch of changes (even small) following Google's change of handling of a certain element IMO says 'manipulation' fairly clearly.

[edited by: TheMadScientist at 9:06 pm (utc) on Feb 24, 2010]

[tedster]

The source is credited with a link AFAIK.

The link only appears in the source for the framed URL, not the framing URL (the parent page). As far as I've ever seen, the parent page does not get credit for the link. The PR "vote" only flows from the URL where the anchor actually is.

[TheMadScientist]

As far as I've ever seen, the parent page does not get credit for the link. The PR "vote" only flows from the URL where the anchor actually is.

Are we saying the same thing in different ways?
'The Parent page flows PR to the iFrame URL.'

Or are you saying something different?

[dertyfern]

The latter suggestion does sculpt the PR, but IMO it still keep a quality visitor experience and doesn't use anything which could be seen as 'hiding' or 'questionable', again, IMO.

Yes, I agree. Good stuff as in my experience the home page tends to get a high proportion of PR, both because the majority of incoming links point to it, but importantly in this case, because of the sheer quantity of internal links from breadcrumbs.

I've practically ceased using nofollow. I've removed most global navigation links from pages with the exception of breadcrumbs. Kept the home page link on breadcrumbs but it really does seem unnecessary in most instances and assuming they do pass a bit of PR value it'd be nice to re-direct some to key pages in need of some juice.

[tedster]

Are we saying the same thing in different ways?
'The Parent page flows PR to the iFrame URL.'

Possibly - I may have misunderstood your earlier post. To clarify, what I have never been able to show in testing is that the src attribute actually acts as a link, in terms of Google's web graph and PR calculation. Only true anchors -- <a> elements that cause the iframe content to change to a different URL by using a target attribute -- seem to act as links and vote some PR.

The last time I tested, a url that had only an iframe src attribute pointing to it would not even stay in the Google index -- if it was discovered at all. I just checked a few such situations where I do have an iframed "utility" page, and its URL does not show up in the current Google index.

That said, I've never tried to iframe only a link to Home. I usually iframe a URL with a set of links to "utility pages" so that eaach link doesn't need to appear on every page of the site.

[trinorthlighting]

IMHO.....

This is one of those subjects where I can see that a webmaster is starting to build their website for Google (Which is a big pitfall) and not for their users which might result in you losing some ground in your search engine rankings in the long run. Bread crumbs are very useful for users and you might want to study your users before you make any changes for Google. Add to that, Page rank has very little use these days except for crawling of pages so why even worry about it?

If you really think about it, most websites deep pages rarely change once they are published, but home pages do tend to change frequently and that is where the PR should be so your page can be crawled more often to find that new content. All websites are a built like a triangle, home page on top, then second level links, third level links and so on. I know I want most page rank to go to my home page because that is where links to new content tend to be posted first and not in my sites deeper levels.

Honestly, you will find that your time would be much better spent adding new content rather than trying to figure out PR. The content will boost your rankings, playing around with internal page rank will not boost your rankings.

[TheMadScientist]

I think trinorthlighting makes a good point in general, but in this specific situation the home page is already not the focus to the best of my understanding from this portion of the OP.

As far as Google SERP's go, it's the main page of each of the subtopics that rank best along with the subtopic's internal pages, not the website home page. As far as traffic goes, the website wouldn't be drastically damaged if it lost all of the home page traffic (it's a long-tail site).

I also think good SEO has something to do with working with Search Engines, so if the changes get visitors to the right page on the site more often without having to pass through the home page, then what could appear to be 'building a site for Google' turns out to be a win-win-win and is actually a visitor and site oriented decision...

Win-Win-Win?
Visitor Win: The site is structured to rank the correct information for visitors to land on (inner pages) rather than the general overview (home page).

Site Win: Visitors find what they are specifically looking for without having to figure out how to navigate 2 or 3 clicks deep into a site, so they like the site.

Google Win: Their visitors found what they were specifically looking for in the results, which means they'll probably use Google to search again.

IMO there's a distinction between building a site specifically to rank in Google for ulterior purposes (MFA) and building a site to rank the correct pages and information for visitors, which again IMO is really what quality SEO does. The distinction I'm trying to draw is the former (even if not MFA as in my definition) only benefits the site owner and the latter benefits all 3 in the long-term.

[TheMadScientist]

Doesn't the nofollow solution leak the very PR I'm trying to conserve?

I think I may have mis-answered this earlier, but I don't feel very well and I'm not feeling like going back and re-reading to cite and correct myself, but...

Weight passed without using NoFollow:
10 links on a page, 10 points passed through each link.

The old version of NoFollow was:
10 links on a page, 5 NoFollowed, each of the other 5 pass 2 point.

The new version of NoFollow:
10 links on a page, 5 NoFollowed, each of the other 5 pass 1 point.

You cannot 'sculpt' to pass more weight into the sub pages anymore, but IMO you can focus the PR on the page containing it, which makes sense. The most linked page has the highest PR. You cannot send the PR accumulated on the page receiving the links to specific pages (or sections of the site) from the page without removing the other links, but the PR of the page doing the linking only passes 5 points, not 10... What happens to the other 5 points of PR previously passed to the other pages? It would seem to make sense to me if it stayed on the page, but like I said I'm not feeling well, so something else might make more sense tomorrow...