Link Structure for 30,000 Page Site

Best linking method?

[paulk]

Hi everybody, I have a question and am wondering if someone can help me out. I plan on making a website that will consist of lots of text and about 25,000-30,000 pages eventually. I am wondering if anyone here can help me out with the best linking structure to optimize the website's PR all across the website. I would like PR spread evently and ofcourse try to get the PR's value the best I possibly can. Anyone here can suggest a linking structure for me? Greatly appreciated, thank you!

[paulk]

edit: I do not plan on having any categories within the site, every article will be in 1 category

[edit_g]

A good start that still holds true:

Search Engine Theme Pyramids
[searchengineworld.com...]

[hutcheson]

If you can't figure out categories, then ... the world will be best served by you not doing any of the linking -- anything you do will be what Google calls "artificial link manipulation", and a net deficit in the sum of human knowledge.

If the content is worth someone else seeing it, then you need to find someone who thinks it's worth spending the time to categorize. It will need to be very very valuable: Speaking from firsthand experience on two such projects, 30,000 articles is a LOT of categorization!

If the content is gathered from specific offline sources: Serialization is better than no categorization at all. Invite interested parties (if you can identify any) to help populate your categories.

If the content is gathered from the net, then read the copyright laws carefully -- now is better than after the lawyers start sending you mail. Make sure you have a very large mailbox.

If you are generating the content yourself ... no, I can't imagine any human on earth knowing enough to write 30,000 pages, and yet not knowing enough to arrange them in some logical order. That you don't know how to arrange them is the strongest evidence that you really won't be able even to collect anything worth collecting.

[paulk]

the 30,000 pages aren't going to be put out right away. 30,000 is the goal for the future. I don't need categories for this sort of project.

[hutcheson]

If you have 30 pages, you need to begin thinking about a structure. If you have 300 pages, you need a structure. But "need" is no longer the operative word for 30,000 pages. A better way of putting it is: you can't have 30,000 pages without a structure. But even at 3000 pages, without a structure you're just a Laputan priest.

Now that structure may be alphabetical, or keyword-indexed, or chronological, or topical. But if you have over 3000 pages, it needs to exhibit hierarchy also.

And if you have a structure, then your links should express the structure.

[Robert Charlton]

I don't need categories for this sort of project.

How can you possibly have a "link structure" without categories? Coming up with a link structure is all about conceptualizing your categories from the beginning.

Also, you don't want to distribute your PageRank equally... you want to prioritize, and in fact you will have to, unless you decide to have 30,000 links on each page.

[SEOPTI]

30,000 sites is spam!

[Jesse_Smith]

Do you even have any idea how big the file would be if you had 30,000 links on one file?! You'll scare the visitors off in no time. No one would scroll through 30,000 links.

[ogletree]

You would need a lot of PR for that many pages. You can't have any more than 100 links on each page. For each of your pages to have one PR less than the front page you can only have 101 pages in your site. To have you pages have 2 PR less than the front page you would have to have 10,101 pages. This would assume that you don't link out and that you have no menu structure.

[Robert Charlton]

Do you even have any idea how big the file would be if you had 30,000 links on one file?! You'll scare the visitors off in no time. No one would scroll through 30,000 links.

That was precisely my point. I thought of using the winking smiley, but I was being a bit more sarcastic and wasn't quite in the winking mood.

The whole project is about hierarchy and prioritization and the management of data. My posts on the thread below about optimizing a form driven site might possibly relate to this topic too, whether or not it's form driven...

How to make a form driven site searchable
[webmasterworld.com...]

[TerrCan123]

If you want the pr to be spread out evenly then you want to only have one link to each page from another one so a chain of links combining them might do.

However the other option is to have a simple site map where the site map only has the link to the content and each page of around 150 titles.

So your main page is a major site map going down alphabetically to lower site map pages [150 links on main page by 150 links on second-tier pages = 22,500 links]

So the second level would be the links to all your content pages and under this structure it would be evenly spread out through the 22,500 pages.

So the structure is the main pages of 150 links to your second tier site map pages, then the second tier pages each have 150 links to the content pages. You could try 200 links per page to get to your 30,000 but I think you should set up the basic structure to begin and then as you add pages it will fill in over time until it is complete.

[paulk]

Ok, I kinda didn't phrase my question correctly. Here is what I want.

Website will have 5 Categories total., so lets take 1 of the 5 categories for example, something like this:

domain.com/index.html links to:

[domain.com...] (Movies - Horror Category)

Now lets say the entire Horror Category is to contain 5,000 pages. What is the best way to pass down PR to all of the 5,000 pages from the /horror category?

Would something similar to this work well:

[domain.com...]

links to:

Article 1-50 (1.html ----> 50.html) located on /horror/index.html

Article 51-100 (51.html -----> 100.html) located on /horror/page2.html

Article 101-150 (101.html ----> 150.html) located on /horror/page3.html

/horror directory will link out to 50 or so pages (page2.html, page3.html, page4.html each containing 50 articles and link back to the home page root of domain.com)

and so on......now would this be a good method to do for PR flow? would doing something like this work? Not sure if I'm confusing anyone here....thanks for everyone's help!

[trimmer80]

Article 1-50 (1.html ----> 50.html) located on /horror/index.html

Article 51-100 (51.html -----> 100.html) located on /horror/page2.html

Article 101-150 (101.html ----> 150.html) located on /horror/page3.html

Not sure if this is what you mean but it sounds like page1 1 - 50 will link to page 2 51 - 100 which will link to page 3 etc etc. I dont like this structure as it will be difficult to get crawled and not pass PR to the lower pages well.

you need /horror/index.html to give the uses links to all of the page2, 3, 4 etc. even if this is a link called... articles 1-50, articles 51-100.

[paulk]

trimmer..thanks thats what I meant,

/horror/index.html will have the following links

Articles 1-50 (Link 1)
Articles 51-100 (Link 2)
Articles 101-150 (Link 3)

and so on......

but is 50 articles per page too much to use or?

Also is it best to link back to domain.com/index.html from each one of those article pages or link back to /horror directory, or maybe do both links back to /horror as well as domain.com/index.html? thanks.

[trimmer80]

50 is a good number for robots. I try and stick between 30 - 50 for users. It is all in how you set the design of the page. But 50 should be fine.

As for link upwards. I like to use what is commonly referred to as footprints or breadcrumbs.
This means that you tell the use where they are in the site for instance in the horror page you will have a link like this

home : horror

in articles 1 - 50 you will have

home : horror : Articles 1-50

and in article 34 you will have

home : horror : Articles 1-50 : Article 34

I find this method to be the most user and search engine friendly

[hutcheson]

OK, say you're categorizing "movies".

30,000 movies dips deep into "B" grade (probably down through about "W", if your grading scale goes that far. So you probably don't want to break down by director -- or maybe just pages for a dozen or few most active directors.

But you do want a chronological breakdown, and with that many, individual years are appropriate.

Alphabetical breakdown, probably by first two letters of movie title.

Separate indexes per country, (or for each "foreign" country, if most of the movies are from the same country.

Separate indexes for "Series" == Friday the Banal part Thirteenth, etc.

Separate indexes for horror themes -- I dunno, I don't touch the stuff: maybe witchcraft, exorcism, pets gone bad, wild animals, computers, IRS audits, ... on a subject like this, the groupings are defined by the data. (Here's where you can show your knowledge of the subject.)

[notredamekid]

Movable Type or another Content management system seems like it would work perfectly in this case.