PR change after layout change

[jetteroheller]

All my sites are created by my own CMS.

The standard navigation on each page was before my big change
December 10th 2011

Logo links to main page
Logo links to main page of subdomain

Arrow buttons to
folder before
page before
up
next page
next folder

A block with links to all the themes of the site
A block with links to all pages in the same folder

Besides this, each index page of a folder
has an index of the pages in the folder.

So when I would have understood PR right

It's the same effect when

A page receives 100 internal links from 100 pages with 100 links on each page
100 * 1 / 100 = 1

A page receives 5 internal links from 5 pages with 5 links on each page
5 * 1 / 5 = 1

Also was written, Google recognices navigation links as such,
always on the same place on each page
always the same links.

I chanted December 10th to a radical new layout.

The two navigation blocks - themes - other pages in same folder
had been removed.

This links are now behind a button created by a javascript.
onclick and onmouseover opens a big box with all this links

Google Analytics showed great improvement in engagement of visitors.
Vistiors stay longer on the site, visit more pages.
So I was very happy with my new layout.

End December 2011, less visits than usual.

I looked now on typical PR data

The sequence is always
main page
page linked from the main page
one link further

page with new layout 4 - 1 - not ranked
page with old layout 2 - 1 - 0
page with old layout 3 - 3 - 2

So I see now at all the pages changed
December 10th a harsh drop in PR when I go deeper in the site.

The other sites in my old layout bring much better
PR into sub pages.

So wants Google really force me to put again 50 navigation links on each page?

I thought reducing links in the navigation area would give the links in the content more weight.

Have links a constant value, regardless there are 10 or 100 on a page?

[tedster]

Well, your first description of how PR works is kind of a rough "rule of thumb" derived from the original PageRank patent. However, Google has made a lot of changes since then.

In particular, I think you're noticing the fact that PageRank now has a lot of interaction with the page layout itself - and the transfer of PageRank is no longer based on a "random surfer" model, but rather on what they've called a "reasonable surfer" model [webmasterworld.com].

[jetteroheller]

Okey, all the layout changes should bring a gain at the "resonable surfer" model, but I experience the opposit.

[tedster]

other pages in same folder had been removed.

How many total pages did you remove, compared to the number of pages that you kept? Is there a chance you removed any pages with good backlinks?

[jetteroheller]

I did not remove any page.

Only the layout on the page changed

There had been two columns for navigation on the left side

List of all themes of the site
List of all pages in the same folder.

This links are removed from the html,
there are now icons with a mouse over,
where at mouse over a javascript shows this links

[Sgt_Kickaxe]

I want to meet a reasonable surfer bot!

Until I do my advice is to make changes SLOWLY and monitor for effect during a longer period of time. Far too many things are interacting together to even trust the results you think you see.