Btw, this is the best explanation I've read on this topic - thanks,
John. I don't know if it's on whatever lists you guys keep for best
posts but count one more vote if it's not...unless it's no longer
accurate?
> Btw, this is the best explanation I've read on this topic - thanks,
> John. I don't know if it's on whatever lists you guys keep for best
> posts but count one more vote if it's not...unless it's no longer
> accurate?
Specifically what I'm looking for is confirmation on John's technical
reasoning for limiting links to 100 per page. For example, is the
following still accurate (you're going to have to read the whole post
to understand the question I suppose)?...
"The difference is:
100 links get 100 pages crawled / indexed
110 links get 0 pages crawled / indexed"
So to be clearer, let's take one of these pages of yours "with >100
links that are indexed and crawled just fine." Firstly, if you mean
the pages containing 100 links then great, but that's not what I'm
asking. If you mean the pages those 100 links are pointing at are all
indexed and crawled just fine (and they have no links coming from any
other pages except the page with 100 links), then that's what I'm
interested to know. In other words, is John's explanation still
accurate?
> > Btw, this is the best explanation I've read on this topic - thanks,
> > John. I don't know if it's on whatever lists you guys keep for best
> > posts but count one more vote if it's not...unless it's no longer
> > accurate?
> Btw, this is the best explanation I've read on this topic - thanks,
> John. I don't know if it's on whatever lists you guys keep for best
> posts but count one more vote if it's not...unless it's no longer
> accurate?
Is it still true? I'd rather ask if it was ever true.
Then think about it long and hard - it also applies, to an extent, to
internal links. The key thing about a link structure is that it
should be natural - each link should be a natural consequence of the
site's construction and navigation.
Every page should have a link to a home page for that site or subsite
so that real human users who find the page can satisfy their curiosity
- basically to answer the questions "who am I talking to" and "how can
I contact these people". In many jurisdictions you _need_ (by law)
company identification information, tax details, accessibility
instructions, copyright statement, etc., and these can all be hung off
the home page.
Then you might have a link to a subject matter home page for that
section of the site. Perhaps these links belong in a structured,
predictable place - a standard header or footer.
Other links really belong in the document, with anchor text useful to
the human reader.
Once you start wondering "how many links can I get away with" you're
on a slippery slope, adding unnaturalness to the site. There are
those, for instance, who claim all forms of menu are unnatural - there
are no menus on my own site,for instance.
If there's one clearly identifiable trend in Google's algorithm
changes over the last three or four years, it's back to naturalness.
It's no stretch to say that using _ANY_ of the commonly recognised on-
site SEO techniques is quite likely to bring you to Google's
attention.
This is very logical from a performance management point of view at
Google. Why check out every site for infringements of the quality
guidelines when you can build a much smaller list of sites by checking
just those that are SEOed? Perhaps something as simple as putting in
a wholly proper header structure (one <h1>, subordinate <h2>s) and
generally tidying up the meta tags - thoroughly legally - is enough to
flag a site for continued observation.
It's quite remarkable how many posters here start off with: "I did a
little bit of SEO on my site and the sky fell."
It then turns out to be something that's been on the site for years.
Ahh, it's great to have my old pre-Google posts being dug up :-).
Looking at this from a purely technical point of view, there are a few
assumptions in there which one can't really know so in the end, I'd
say "yes and no" :-).
Yes, if you have too many links (100 is purely an arbitrary number at
this point -- it could be 10 or it could be 10,000) split across to
pages which don't have any other links, then it's certainly possible
that we won't get around to crawling and indexing everything. Is one
page, which gets one link out of 1000's from only one page on the web
really important enough to merit being crawled and indexed? It's hard
to say. We have limited resources and the servers out there have
limited resources as well, so we have to draw a line somewhere.
To put it into perspective, assume we're talking about a perpetual
calendar script, which shows all days of a year with a link to every
day's events. We've crawled out to the year 2037 and we see 365 links
to individual days. Should we continue crawling? or just call it a day
and wait until we have a better reason to crawl the 23rd of November,
2037?
At any rate, having too many links on a page can certainly confuse
users, so I'd avoid it (if you can) even just because of that. 100
links per page is not a hard limit and we can handle pages that have
more, lots more... but at some point it just starts feeling strange,
especially when the only link to a new page is one of 1000's on a
single page somewhere...
There's one kind of page/site that I've run into which often has much
more than 100 links / page: these are sites that use CSS to show menus
with several levels of detail. For a user, using a menu like that can
help find a specific part of a site much faster than by navigating in
several steps. However, with that many links, it can make it hard for
search engine crawlers to recognize the relevant ones.
One novel solution a user here came up with was to use rel=nofollow on
links which were outside of a certain scope with regards to the
current page. I think that's a pretty good balance between usability
for users and usability for search engine crawlers. That said, it's
not something that would affect the average site :-). You can find the
thread at http://groups.google.com/group/Google_Webmaster_Help-Indexing/browse_...
> There's one kind of page/site that I've run into which often has much
> more than 100 links / page: these are sites that use CSS to show menus
> with several levels of detail. For a user, using a menu like that can
> help find a specific part of a site much faster than by navigating in
> several steps.
Even considering this the 100 as suggestion is very reasonable for
99.99% of the sites. If amazon can solve it's navigation with ~250
links on the homepage and ~330 on the category pages, then most of the
sites have to be able to remain far below 100 to ensure optimal
usability.
> Ahh, it's great to have my old pre-Google posts being dug up :-).
> Looking at this from a purely technical point of view, there are a few
> assumptions in there which one can't really know so in the end, I'd
> say "yes and no" :-).
> Yes, if you have too many links (100 is purely an arbitrary number at
> this point -- it could be 10 or it could be 10,000) split across to
> pages which don't have any other links, then it's certainly possible
> that we won't get around to crawling and indexing everything. Is one
> page, which gets one link out of 1000's from only one page on the web
> really important enough to merit being crawled and indexed? It's hard
> to say. We have limited resources and the servers out there have
> limited resources as well, so we have to draw a line somewhere.
> To put it into perspective, assume we're talking about a perpetual
> calendar script, which shows all days of a year with a link to every
> day's events. We've crawled out to the year 2037 and we see 365 links
> to individual days. Should we continue crawling? or just call it a day
> and wait until we have a better reason to crawl the 23rd of November,
> 2037?
> At any rate, having too many links on a page can certainly confuse
> users, so I'd avoid it (if you can) even just because of that. 100
> links per page is not a hard limit and we can handle pages that have
> more, lots more... but at some point it just starts feeling strange,
> especially when the only link to a new page is one of 1000's on a
> single page somewhere...
> There's one kind of page/site that I've run into which often has much
> more than 100 links / page: these are sites that use CSS to show menus
> with several levels of detail. For a user, using a menu like that can
> help find a specific part of a site much faster than by navigating in
> several steps. However, with that many links, it can make it hard for
> search engine crawlers to recognize the relevant ones.
> One novel solution a user here came up with was to use rel=nofollow on
> links which were outside of a certain scope with regards to the
> current page. I think that's a pretty good balance between usability
> for users and usability for search engine crawlers. That said, it's
> not something that would affect the average site :-). You can find the
> thread at http://groups.google.com/group/Google_Webmaster_Help-Indexing/browse_...
> In many jurisdictions you _need_ (by law)
> company identification information, tax details, accessibility
> instructions, copyright statement, etc., and these can all be hung off
> the home page.
Hey Phil, do you have any more information on this? Or maybe a link to
some?
metronome wrote:
> On Nov 23, 5:43 am, Phil Payne wrote:
> > In many jurisdictions you _need_ (by law)
> > company identification information, tax details, accessibility
> > instructions, copyright statement, etc., and these can all be hung off
> > the home page.
> Hey Phil, do you have any more information on this? Or maybe a link to
> some?
I am not a lawyer - you should seek professional advice for your
jurisdiction. But Googling for the UK's requirements produced, among
other things, this:
"Companies in the UK must include certain regulatory information on
their websites and in their email footers before 1st January 2007 or
they will breach the Companies Act and risk a fine.
Every company should list its company registration number, place of
registration and registered office address on its website as a result
of an update to the legislation of 1985. The information, which must
be in legible characters, should also appear on order forms and in
emails. Such information is already required on 'business letters' but
the duty is being extended to websites, order forms and electronic
documents."
Okay, thanks Phil. I'm based in the US so this doesn't apply to me,
but I'm going to do some research to see if there isn't something I
have missed regarding our regulations o'er here.
Thank you Phil and John both! As always, your responses answered my
question and then <s>some</s> a lot. I wish I could repay the favor
with more than stars. :|
> > Btw, this is the best explanation I've read on this topic - thanks,
> > John. I don't know if it's on whatever lists you guys keep for best
> > posts but count one more vote if it's not...unless it's no longer
> > accurate?
> Is it still true? I'd rather ask if it was ever true.
> Then think about it long and hard - it also applies, to an extent, to
> internal links. The key thing about a link structure is that it
> should be natural - each link should be a natural consequence of the
> site's construction and navigation.
> Every page should have a link to a home page for that site or subsite
> so that real human users who find the page can satisfy their curiosity
> - basically to answer the questions "who am I talking to" and "how can
> I contact these people". In many jurisdictions you _need_ (by law)
> company identification information, tax details, accessibility
> instructions, copyright statement, etc., and these can all be hung off
> the home page.
> Then you might have a link to a subject matter home page for that
> section of the site. Perhaps these links belong in a structured,
> predictable place - a standard header or footer.
> Other links really belong in the document, with anchor text useful to
> the human reader.
> Once you start wondering "how many links can I get away with" you're
> on a slippery slope, adding unnaturalness to the site. There are
> those, for instance, who claim all forms of menu are unnatural - there
> are no menus on my own site,for instance.
> If there's one clearly identifiable trend in Google's algorithm
> changes over the last three or four years, it's back to naturalness.
> It's no stretch to say that using _ANY_ of the commonly recognised on-
> site SEO techniques is quite likely to bring you to Google's
> attention.
> This is very logical from a performance management point of view at
> Google. Why check out every site for infringements of the quality
> guidelines when you can build a much smaller list of sites by checking
> just those that are SEOed? Perhaps something as simple as putting in
> a wholly proper header structure (one <h1>, subordinate <h2>s) and
> generally tidying up the meta tags - thoroughly legally - is enough to
> flag a site for continued observation.
> It's quite remarkable how many posters here start off with: "I did a
> little bit of SEO on my site and the sky fell."
> It then turns out to be something that's been on the site for years.
|
