Hi,

With the recommended limit of 100 links per page (http://www.google.com/webmasters/guidelines.html), what are your thoughts on this limit? Is it a reasonable one?

Whilst too many links per page will cause problems for bots and humans alike, could there not be a legitimate reason to have more than this limit? With the differences in interpretation of the abbr element between browsers and the abbr element is intended for user-agents and not humans, some equality should be applied. Thus, in order to guarantee the same assistance to the human or client application, some might recommend the use of hyperlinks as well.

For example:

<p>The United States of America (<abbr title="United States of America">USA</abbr>)</p>

This would become something like:

<p>The United States of America (<a href="glossary.htm#usa" title="Look up in the glossary">
<abbr title="United States of America">USA</abbr></a>)</p>


Say a site has 100-200 pages, then a proportion of those pages may have 30-50% of those as links in menus and the like on one page. Add a handful of links from within page content to the number of menu items and you could push the number of links on such pages to around about the 'recommended limit'. What happens when there are a number of abbreviations in the content? Sometimes the abbreviation is more acceptable (to the eye) than the full-length word or phrase, such as VeriSign, Inc. or Mr. and Mrs.

Which of the following would you choose to reduce the link count?
reduce number of menu items
reduce use of linked abbreviations
reduce the number of abbreviations full-stop
split pages where possible (ultimately leading to more pages to link to for some other pages)
restructure internal linking

Thoughts most welcome.

fmj

I would stick to canon Google for this one. From the same guidelines page as you cite I have;

"Another useful test is to ask, "Does this help my users? Would I do this if search engines didn't exist?"

I think this should apply to accessibility and other user enhancement techniques too. In your example you have a valid reason for adding lots of links to your pages.

You're not adding lots of external links. You're adding lots of links to the same page (the glossary) and I'm sure that's a measurement already in place. Be aware of how central you're making your glossary in your site architecture.

This is easy for me to say, though, as I'm not in a situation to preach it. I don't believe you should be penalised. I don't know whether you will. I'm confident that as search engines advance their technology that any penalisation you may suffer will be lifted.

There's been a bit of a debate recently as to whether accessibility techniques can be mistaken for spam (or are used to disguise spam). I don't think there was a satisfactory outcome and this suggests to me that no one knows. It would be a terrible shame if people had to 'de-access' their sites because of poor spidering rules. Again, this is where I don't have to practise what I preach and safely predict that if there is an overlap that the search engines will sort it out given time.

It's not the number of links on a page that matters - it's the page's filesize that's important. Google doesn't deal with pages that are over 100k, and their 100 links per page recommendation is their estimate of how big a 100 link page would be. Google happily swallows hundreds of links from a single page as long as the page's filesize isn't more than 100k.

It's not the number of links on a page that matters - it's the page's filesize that's important. Google doesn't deal with pages that are over 100k, and their 100 links per page recommendation is their estimate of how big a 100 link page would be. Google happily swallows hundreds of links from a single page as long as the page's filesize isn't more than 100k.

I would hmmhmm a bit at that. You're right in broad-strokes but I think it needs some caveatting. These forums are great because it allows us to share our own personal experiences and these are mine:

Google does deal with pages over 100k. It's just that first 100k is cached. The original Google white paper describes how the first character has the most weight and has you add characters their importance reduces. When you get to a certain limit each addition character only has a minimum of importance. In this original paper we never got to a point where additional characters did not count.

The 100 link limit has been in the guidelines for a long period of time. Google updated their guidelines at about the same time as they launched their sitemap XML program. At that time the 100 link limit was repeated. In fact, it's the only point which Google mentions twice.

I think an early use of the 100 link limit was to identify FFA sites.

In my experience, certain clients who had more than 100 links per page at that point - when sitemap XML was introduced and when the 100 link limit was repeated - began to loose pages from the index. These were large directory-style clients with hundreds of thousands and millions of web pages for a single domain. To open their large archives/directories/etc up they had hierarchical sitemaps (as google suggests) but because of the scope of these directories/archives/results there were often much more than 100 links on the top and mid level sitemaps.

To address the problem on each of these sites we recommended two things:
1) reduce the number of links below or closer to 100
2) join sitemap XML

The tactic worked. These sites which had, in some cases, lost 80% of their page inclusion returned to full strength.

This is a great example of why everything on these forums tends to be speculation. Nothing you do on the web is done in isolation and so you can never be sure what caused a change. In this case it seems likely that the sitemap XML was more important than the drop to (or near) 100 links.

Sites which had spidering problems and which introduced large sitemap XMLs took much longer for their (similar sized) pages to be added to the index.

I wasn't speculating, Wail. It's what Google said - the 100 links per page is their estimate of how big a page would be to stay within the 100k limit. Google may handle pages over 100k, but, if they do, they truncate them. It's 100k filesize that's the limit, and not 100 links. I've known pages with 500 links on them have all the links spidered, but the pages were under 100k. 100 links per page is not a limit.

Ah. I see your point of view. As I said; you're not wrong. This guidelines page (http://www.google.com/webmasters/guidelines.html) says two things:


"Offer a site map to your users with links that point to the important parts of your site. If the site map is larger than 100 or so links, you may want to break the site map into separate pages."
"Keep the links on a given page to a reasonable number (fewer than 100)."


The 100k limit - which I agree really is an issue and certainly tightly coupled with a page with too many links - isn't mentioned here.

No, it's not mentioned in the guidelines. I'm only repeating what someone from Google said, when he explained why they suggest no more than 100 links per page.

Hi Wail / Phil,

Thanks for your input guys. Sorry about the delay in response, but things have been busy lately.

So, if the specified limit is only an advisory, then the use of hyperlinked abbreviations is hopefully not an issue. As Google appear happy to use the glossary pages of indexed sites (e.g. define: foo) for their search queries, would I be correct in the presumption that their algorithm is set up to detect such content, and thus would not apply penalties as the content is deemed 'legitimate'?

I suppose that the best option would be to use abbreviations more sparingly (wherever possible), use a menu structure that keeps links on a page down to a reasonable count (whilst still providing sufficient navigational options for the user), and then implement hyperlinks for those abbreviations that are required to be used. All usability goals should then be met.

rgds

fmj

On thing i can't find a lot of discussion on is how the # of internal links on a page affects the PR that is passed along. I have a home page PR6. If i have a home page with 200 links or so, is that the reason my 2nd level pages are a PR 4? Similarly, my second level page often have over 100 links per page and the 3rd level pages usually have a PR 2. (I have a dynamic site with over 4,000,000 pages.)

You should find this helpful:-

http://www.webworkshop.net/pagerank.html

In a nutshell, a page's PageRank is divided by the number of links going from the page, and each destination page gets an equal share of it. So, if a page is linked to from a PR6 page, and it is the only link on the PR6 page, then the destination page will receive the maximum PR that the linking page can give. But if the linking page has another 99 links on it, then that destination page will only receive one hundredth of the linking page's PR.

There's a damping factor involved, but it's not important here.