A while ago I did an experiment, that I have mentioned periodically
here, where I disallowed access to a couple of pages that I thought
had the most quality inbound links.

At almost the exact same time as they were removed from the index,
allowing for a short delay, the number of pages of the site in
Google's index dropped and the remaining pages didn't do as well in
the SERPs.

Re-allowing indexing access, the site recovered, twice.  Once, over a
period of about a month and then again, over a period of another month
after the "Supplemental Result" indicator was removed which seems to
have been implemented by pushing previous data to all datacenters.  In
any event, I saw the fall and rise of the site occur twice so it is
doubtful it was a fluke.

But, I was surprised today to find that although the site is doing
well in the SERPs and everything else seems fine, the pages that I had
removed access to are NOT in Google's index currently nor have they
been for a couple of days which is longer than it took for the site to
drop its performance during the test.

If the pages, that seem to contribute so much to the site's aggregate
PageRank are not in the index at all now, where is the PageRank coming
from?

It might seem possible that all of a sudden the site received some
massively powerful links from somewhere but given the period of time
and timing, it is very unlikely.

It would seem that if you decide to remove a page, its PageRank is
tossed to the wind but if Google decides to not index a given page,
any PageRank it has would seem to be retained and its links used to
determine PageRank of other pages.

If you are curious, the pages no longer in the index are code example
pages where there is basically nothing on the pages except (X)HTML,
CSS and Javascript code displayed and virtually no other textual
content. By way of example of an example, ;-) here is one of the
pages, http://cass-hacks.com/articles/example/js_thumbs_to_full2/

Craig

Just a guess:

I assume that with "decide to remove a page" you mean the URL removal
function? I am pretty certain that this is not a real URL removal from
the "database", but rather just a filter applied before the search
results are shown.

Additionally, I feel that the link lists for pagerank calculations are
independent of the indexed pages. Pages can have pagerank when they
don't even exist (even when the domain name does not exist). I bet
that if a page has been crawled and outbound links were known, those
links would pass pagerank even if the page ceased to exist (as long as
it was in the pagerank link lists). That would be similar to the
"noindex, follow" robots meta-tag.

I wonder (never a good sign).. assuming there is a high-value page
that links to your site. Would it be better for your site if that page
was removed (404 or URL removal tool) or if the links were removed? My
guess is that a missing page would still pass value, while a page with
no known outbound links wouldn't.  But then again, that effect is
bound to be very temporary :-P.

I'm not sure how that would change anything though :-))

John

Does this mean that I can sell links on pages that no longer exist and
not have to worry about Google penalizing me for it?

On Aug 24, 2:42 am, JohnMu wrote:

Now you're getting tricky.

I have a lot of pages that no longer exist. Hmmm.

John

No fair guessing!

Just because I do that, doesn't mean you get to.  :-P

Yep.

That may be but when the pages I requested to be removed were actually
gone from the index, that is when the site took a nose dive into
oblivion.  Also, when I re-allowed them, that is when the site started
to revive but it didn't happen all at once.  It seemed to take about
the amount of time it would for the links to the previously removed
pages to be re-crawled.

That would be my guess too.

A URL can have PageRank for a page that doesn't exist but the page
that doesn't exist can't have PageRank, kinda sorta because it doesn't
exist.  ;-)

That makes sense.

Since losing a high value link is never a good thing, obviously
removing the link would be bad so the best case scenario is if the
page the link is on just goes missing, 404, and that even if Google
does remove it from the visible index, that it still retains it
somewhere.  That is a big "if" though regarding Google keeping data on
404'd pages even after removal from the visible index.

I have yet to experience losing a link of any significance through
either method so I guess I can't say which would be "better" if indeed
there would be any difference at all.

But, it is an interesting question!  Even when Google "removes" a page
from its visible index that is returning a 404, how far "removed" is
it???

Just a little bit, a lot?  :-()

Will a little dab do ya?

Is it?  Again, what does Google do with pages that are 404'd after
finally removing them from its visible index?

It would seem to make sense that Google might actually totally delete
them from its database, as opposed to just filtering them like it
seems to for owner removed or Google removed pages but are we sure?

On the other hand, set up a bunch of pages with links to your favorite
sites and then 404 the page.  That wouldn't seem to be beneficial for
search quality so my guess is that once a page that is 404'd is gone
from the visible index, it becomes so much bit dust in the bit bucket.

Well, make it your first priority to find out as soon as you start
work next month and report back to us.  :-()

:-)

No no no!!!!

You have to have the link on the page and get money for it first
before you get rid of the page.

You think I would believe you put a link on a page that no longer
exists and didn't exist when you said you put it there?

What do you take me for, an idjit??

Umm, don't answer that.  :-P

I can see this should have gone in Random Chat!  :-()

Maybe a better test would be to 301 redirect from the page you wanted
to remove PR from next time you do this test? Adam has said that 301's
pass pagerank and 'associated signals' so that might be a better way
to cut off the PR to 'daughter pages' if the page you are 301
redirecting is the only source of PR for those daughter pages?

Cheers,

M

On Aug 24, 6:22 pm, cass-hacks wrote:

The NEXT time I do this test????

Notinamillionfeckingyears!!!!!!!!!!  :-()

Actually the point of the test was to try to throw away PageRank and
see what happened so 301'ing the URLs to some other page would have
defeated the purpose of the test.

I knew, in fact, I wanted the site's performance in Google to drop
although I had no idea how far it would drop until I had lost 2/3'rds
of the site!  :-()

But, having had to go almost 2 months recollecting the PageRank not
once but twice, I'm in no really hurry to see if shooting myself in
the foot hurts, again.

I think it probably will.  :-()

Craig

On Aug 24, 7:25 pm, dockarl wrote:

So if pages that were indexed but are manually removed continue to
pass on the link love for links they contained the last time they were
indexed....isn't this just open up a new toolbox for black hatters?

Say I have this highly trusted domain on a good clean host.  I set up
a feeder page to send some linkin lovin to my new anti virus venture.
Since it's quite popular the page is indexed quite soon. I then remove
the page so it is no longer associated with trusted domain.   Now I go
fire up my blog commenting script and ping about 900,000 blogs with
links to my trusted site page.  Would this page gain in authority?
Would the trusted nature of the removed page clean the mostly spammy
links up?

I haven't tried any of this, but just wondering.

On Aug 24, 8:52 am, cass-hacks wrote:

Yep, I understand that - what I'm saying is that if you 301'd to some
other page on a different site or different 'branch' of your site,
you'd achieve the same thing - you'd (as far as PR is concerned)
orphan the daughter pages in the same way as just removing the 'parent
page', although it would take a bit longer - a 301 passes the pagerank
and associated signals of the 301'd page to the target page. Following
me?

Ciao,

M

On Aug 24, 11:52 pm, cass-hacks wrote:

Oh, that is a horse of a totally different color!

That I might give a try!

Not right now though, I haven't had much traffic to the site while all
this was going on and it is sort of nice seeing people visit it once
again.  :-()

Thanks for the clarification though, that is something I had never
thought about before!

Craig