Menu on top and bottom total more than 100 links

[silverbytes]

I've read "keep pages under 100 links" as an absolute truth. In the other hand long pages with horizontal menu on top, justify to have another horizontal menu on bottom page. Specially if top menu is hiearchical and bottom menu is different (ie: plan text). Since users scroll page down, I find usefull they may continue browsing site by clickin one of links at bottom page, wihtout having to scroll all way to "heaven" top page to go on.

But, that gives me some 150 links in page, what is "antigoogle". I had no problems this far but I understand the advice is to stay below 100 links.

Anyway I feel like if I remove bottom navigation I'm doing it for search engines and not users (and general concept is otherwise, do pages for humans not search engines).

One last thought: how do you keep your links under 100 in large sites? I use to put same navigation bar all over, but what are other alternatives?

[tedster]

One hundred links not an absolute truth - it's a guideline for getting optimum results. The more links you throw onto the page, the harder it is for any search engine to sort everything out and rank your pages in the best way.

Still, recent Google communications have made it clear that they do NOT stop at 100 links and they will do their best to handle whatever you throw at them. But as I have commented elsewhere, for the sites I work with the ones that are doing best on Google are in the area of 40 links total on the Home Page. It can take a lot of work to craft an information architecture that allows that small a number, but it serves you well, both your visitors and the search engines have an easier time of it.

[silverbytes]

Ok. Does those sites have 40 pages or how do you handle links and contents in order to keep only 40 in home?

[tedster]

They have thousands and thousands of page, silverbytes. The key is information architecture - an essential and often ignored discipline that comes from Library Science, just as HTML itself does.

The best resource I know of on this is O'Reilly's book, now into its third edition. We also have a few good threads here:

Information Architecture for the small site [webmasterworld.com]
Information Architecture for the small site [webmasterworld.com] - part 2
Putting Information Architecture into practice [webmasterworld.com]

[silverbytes]

Thanks. What is the effect of having different menus on same pages linking to same internal pages?

I mean if you have a top menu linking to 50 pages and again another version of that menu on bottom linking to same 50 pages does that negatively affects your page rank?
Is that page rank divided by 100 instead 50?

I've heard of positive effect on footer nav bars using usefull anchors in past. But is that effective nowadays?

[tedster]

So how does Google calculate PR when there are multiple links to the same target url? There are some strong differences of opinion on that one! One person I know feels that only the first link is counted, even to the point of saying that the second link's anchor text is disregarded. Other people feel that each link gets exactly its share.

Now if you have 50 links at the top and an identical 50 links at the bottom, it would be the same mathematically if the PR vote is 2/100 or 1/50. It only becomes an issue when there ae links to other urls also on the page.

This is probably one area where the original PageRank forumla has been modified. My experience tells me that the second link's anchor text does count. And footer links that are truly navigational are also still helpful as far as I can see. What is problematic is a footer with lots of external and off-topic links, rather than user-friendly extra navigation.

[silverbytes]

second link's anchor text does count.

I agree.

[Marcia]

Even if the second anchor does count (how about 3/4 anchors to the same page), does that mean that they both give a portion of the PR pie to the linked_to page?

Two-fold question (though it may be redundant):

a) If there 100 links on a page, 50 on top and 50 on botttom, going to the same 50 pages, does the amount of distributable PageRank get divided by 50 or by 100?

b) Leaving anchor text aside, will each of those 50 pages get:
-1/50th of the PR love from the homepage? or
- 1/100th of the PR love passed from the homepage?

I once asked an engineer about this in relation to shopping cart pages that at the time could not be spidered or indexed. I asked whether if there were, for example, 100 links on the homepage and if half the pages linked to were not spidered or indexed, would the PR from the homepage get divvied up by just the indexable pages (say half or 50), or by all outbound links from the page.

The answer I got: the amount of PR for each link is divided by the number of ALL the links on the page - even the ones to the non-indexable cart pages.

I asked, "you mean divided by 100 instead of 50?"
Answer: "Yes, by 100."

Things may have changed since then in some respects, but if that were no longer valid, wouldn't that be more than just a minor modification of the original PR formula on how PR is calculated?

Also:
What if there are 150 links on the page, 100 top and bottom to the same 50 pages, and another 50 going to different pages? Do those 50 pages get 2/3 of the aggregate passable PR, and the 50 that have only 1 link each get 1/3 of the aggregate passable PR?

And: As a rhetorical question, are we absolutely certain that global navigation links - like 1px sitewide footer links - pass PR in the same way as a link in the body text of, for example, above the fold of the homepage?

[edited by: Marcia at 1:47 am (utc) on Aug. 11, 2008]