• Register
• Log in

SEO Theory - SEO Theory and Analysis Blog

Algorithm analysis, Web community relationship analysis, SEO practices and techniques, industry news, etc.

• Home
• About SEO Theory
• SEO Glossary

Linking patterns: Why link-heavy pages win, lose, or draw

Posted by Michael Martinez on September 18, 2007 in Link Theory

How many links per page is okay?

John Biundo made an interesting point in Ten Lessons From SES San Jose. Let me quote him for you:

2. Who says you should only have 200 links per page?
This is something that has passed as wisdom since I’ve been following SEO, and it’s always made some sense to me. Indeed, when asked directly, representatives from Ask and Google seemed at first to agree. Their common response was effectively “gosh, if you’re designing pages with more than 200 links on them, you have to ask yourself whether you’re doing the user any favors.”

When reminded that user sitemaps often have more than 200 links, both vendors said, in effect, “yeah, well, that’s probably OK.” Both acknowledged that there’s no hard limit, but that with some notable exceptions, pages with lots and lots of links could start to look like spam. I take the usability comment to heart, and in general, will continue to advise my clients to try to design pages that users can easily understand and navigate, which will naturally tend to produce pages with fewer than a couple of hundred outbound links.

Usability is not defined by how many links you have on a page or how much scrolling the user has to endure to find a specific link. That’s more of an accessibility issue and that is why you want your site search tools to be easy to find and use (regrettably, my site search plans usually fall short of perfection).

What is the chief difference between an HTML sitemap with 200 links and a reciprocal links page with 40-60 links? The answer to that question should be obvious. But how many of you would want to say, “The sitemap links to relevant content”?

However, a coherent Web site (like an eclectic blog, for example) can have several hundred pages of content on as many topics. The HTML sitemap would not be linking to “relevant content”, would it?

Analyzing link and URL structure for meaning

How many of you would look at the URL structure of the links on the HTML sitemap and say, “They’re all pointing to the same host”? Does linking to the same host really make that big a difference to a search engine? For example, what if the host is home to 100 separate user accounts, each distinguished in its URL structure only by a tilde (~)?

There was a time when URL theorists argued that sub-directories designated by a tilde were probably treated as separate hosts by search engines. Those were the pre-Google dominance days. How do you think such URL structures work now?

If the majority of a site’s content is scattered across sub-domains, as in an information portal that uses sub-domains for each city, state, or topic category, should all those sub-domains be treated as a single host by the search engines? Do you think they do treat them that way?

So then why should search engines care if you place 300 links on a Web page? In fact, why should anyone care? There is no sure-fire algorithmic way to ensure that every URL on a domain (with or without all its sub-domains) really belongs to the same “host” or “site”. That’s a limitation of search engine technology, not of Web technology.

What are Web sites and who cares?

The Web is designed to be fluid, tolerant, and pervasive. It is intended to connect everything that can be connected. “Web sites” are afterthoughts and because they are afterthoughts they are not easy to define. If you and I cannot agree on what constitutes a Web site, then how are the search engines — who only want to document the Web content we create — to figure out what constitutes a Web site?

The link between accessible Web sites and optimization

Now let’s go back to that accessibility issue. If we blunder through our Web design, confusing search engines with blurred distinctions between Web sites, hosts, and domains, what happens to the people who are dependent upon those guides we are expected to place in our page design for their convenience?

Many SEOs know there is a connection between strong on-page optimization and good accessibility. But people have been hard pressed to explain why, other than to say that the easier it is to find what you are looking for, the easier it is to rank for what people want to find.

Nonetheless, accessibility doesn’t help us to distinguish between Web sites. Hence, accessibility is not a standard we can use to measure the boundaries of our content. Those boundaries, however, are important because they establish rules of behavior for search engines, people, and accessibility software.

What is strong linkage?

When I speak of on-site optimization I usually say you want “tight linkage”, “strong internal linkage”, “good linkage”, etc. I mean more than just putting a lot of links between your pages, however. You need to create meaningful links. Drawing connections between relevant pages is one way to create meaningful links. Putting descriptive anchor text in your links is another way to construct them.

How do URLs help Web site structure?

But if you step back and consider your naming conventions, you’ll see that embedding keywords in URLs strips you of some of that tight internal linkage. That isn’t to say that I wouldn’t want to use Grace Park’s name in a Grace Park URL. Rather, I want to make sure that the whole URL (xenite.org/features/grace-park/) tells both me and my visitors something intuitive about the content. In this case, it’s a “feature”.

Spend enough time browsing Xenite and you’ll find that we have F.A.Q. pages, directories, dedicated mini-sites, and feature article sections. Some of those feature article sections are larger than many business Web sites. Some of those feature article sections are occasionally updated with new information or content but their original design is intended to be comprehensive. They aren’t really designed for growth.

The mini-site sections, on the other hand, are indeed ready for more content. They look different, act different, feel different from the feature article sections. Xenite has become a slurry of Web content because it is host to many different types of Web sites, but those sites all promote Xenite itself and some of them heavily cross-promote other parts of Xenite.

How does linking work for superlarge sites?

Of course, not everyone wants to create a Xenite.Org. In fact, you can study superlarge content sites that are so homogenous there is no mistaking what they are: Web forums. Popular Web forums accumulate hundreds of thousands or millions of posts as the years role by. Amazingly, search engines can find some of the oldest posts on those popular forums while struggling to find newer posts on small forums.

What is the difference between a small forum and a superlarge forum?

One somewhat obvious distinction is the link profile of the superlarge forum. It’s undoubtedly pulling in a lot of links from other sites if it has hundreds of thousands of posts. People link to those discussions from other forums, from blogs, sometimes from news stories, and sometimes from directories. Social media sites will also link to forum discussions (in fact, some social media sites won’t link to root URLs — you must deep link).

A relatively inactive forum or new forum doesn’t have much of a link profile. Its internal linkage won’t help it much unless it uses static HTML pages. Do you know why? The reason may not be obvious to everyone.

The linking differences between forums and blogs

Forum software, unlike blog software, is designed to hide content. Blog software is designed to make content more visible, often by replicating that content (Blogger is an exception — they now bury old content). Forum software designers make it almost impossible to keep old discussions crawlable. They scroll off the active discussions list after a day, a week, or a month (you can usually configure how long the posts remain visible).

Some forum software offers “search engine-friendly archives” that don’t really offer any advantage over live forum posts. Old archive pages may hang around longer than the live posts but archive pages tend to strip out all formatting.

Forums tend to have distinctive link patterns. Their administrative functions often have more links than their index pages and discussion pages. Every post usually has a collection of “QUOTE”, “REPLY”, “START THREAD”, etc. links to incidental functions that really don’t need to be crawled or indexed. Blogs lack these kinds of navigational link structures.

So another distinction between a blog page and a forum page is that the forum page may be honeycombed with duplicate links to incidental functions. Some blogs can be covered with blogrolls but many blogs only show their blogrolls on their front pages. The heavy internal linkage common to modern forum packages is both annoying and distracting. But that heavy linkage actually improves usability while both improving and diminishing accessibility and it also impedes efficient crawling.

Ambiguous linkage: How to compensate for ambiguous link structures

Forum software tends to create ambiguous linkage so it should not surprise anyone that forums need a lot of external links in order to be well-indexed. It’s not that those external links qualify pages (a lot of external links may actually point to incidental function pages — do you know why?). It’s just that people will intentionally link to actual discussions more often than they’ll intentionally link to incidental functions. Statistically, you’ll get a lot of older content crawled and reindexed when there are static external links pointing to that older content.

Some forum operators will bump old discussions to bring them back up to the top. I’ve done that occasionally myself. Bumping threads helps improve their visibility for both people and search engines. But you cannot bump a blog post. Once the post is made it’s archived and stored and forgotten. The blog mostly lives through its front page. If you don’t allow your blog to create individual archive page for each post, you’ll eventually lose granularity in your visibility because people don’t have anything to link to.

Deep-linking to individual blog posts thus helps your friends and clients even though you may seem to be helping confuse the search engines. Many people choose to promote their collective archive pages (weekly or monthly) rather than individual blog posts. If you can attract links to those weekly or monthly archives that’s okay. If you don’t get links to them, you have a problem. But some older blogs now have dozens, even hundreds of weekly archive links in their margins. Unlike blogrolls, archive rolls may be carried from page to page on a blog.

Is it okay to put a lot of links on a page?

The bottom line here is that putting large numbers of links on your pages is not usually a problem. The search engines may say they don’t like that but the truth is that they have rewarded millions of Web sites with intensive crawling for having large numbers of links on their pages. There are legitimate concerns about how automated reciprocal linking programs may get some people into trouble, but it’s the fact of high volume reciprocation — where the reciprocal links are concentrated into one or only a few pages — that gets sites into trouble most often.

Some reciprocal linking programs have caused trouble for other reasons, but generally speaking the more you rely on reciprocal linkage the less likely you are to rely on better linking practices anyway. A powerful linking strategy doesn’t need to depend on external links. In fact, if external links are integral to your linking strategy, you have a relatively weak linking strategy.

Does link baiting help, hinder, or matter?

There may be people who quickly ask, “What about link baiting?” Link baiting is a perfectly valid Web marketing strategy, but it’s not search engine optimization and most Web sites are not suitable for link baiting anyway. As far as search engine optimization is concerned, link baiting is inefficient. You have to create quality content but superlarge content sites depend on volume more than they depend on quality.

Link building for volume content works on a much slower, more gradual basis, although as your community grows you accrue links at a faster rate. Getting a site to that point takes a lot of work but you can make your work more efficient by improving your internal linkage. You improve your internal linkage by creating more meaningful, more visible links. In short, you create more links as you grow your site.

Links are not the problem

Agonizing over how many links you can put on a Web page is a waste of time. Placing a lot of links on a page is not the problem. Creating a page that cries out to search engine algorithms, “SPAM! SPAMMED PAGE! I AM SPAMMING” is the problem.