• 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

High Quality Links: Help with internal link structure in SEO

Posted by Michael Martinez on July 18, 2007 in Link Building, Seo Myths

SEO Theory has received a strange burst of search referral traffic over the past few days, perhaps in response to last week’s series of articles on Backlinks and Backlink Theory. It seemed like an opportunity to boost my internal links and to say something more about links to external sites and SEO.

High Quality Links

The phrase “high quality links” is one that I helped make popular last year when I wrote on the SEOmoz blog about high quality links. SEOmoz shot to the top of search results for that expression immediately without any inbound linkage thanks to my egregious use of on-page factors (and I’ll confess that my further use of such factors to help the MOZzers rank well for other forward-thinking expressions was curtailed by uncredited editing).

The best links you can get will always be the links you give yourself. I have said that often and in many ways. Linking begins at home. Your links are treated as expressions of your trust, your opinion, and your faith in the destinations to which those links point. In that respect, you should never think in terms of any possible difference between internal links versus external links.

No TrustRank. No TrustRank. No TrustRank

Let’s talk about trust. It is a sign of ignorance or laziness for any self-proclaimed search engine optimizer to speak of “TrustRank” with respect to Google. Google did not invent TrustRank. Yahoo! developed TrustRank with Stanford University in 2004. Google did briefly, apparently, seek a trademark for “trustrank” but they subsequently removed all references to it from their network. Nonetheless, SEOs cannot seem to ever let go of misinformation, bad ideas, or nonsense, so we’ve heard nothing but “Google and TrustRank” ever since (despite the fact that the confusion over TrustRank was officially cleared up in 2005).

That said, I have often pointed out that Google uses trust filters. And that post doesn’t exactly say “Google uses trust filters”. Instead, Matt Cutts told his readers that We saw so many urls suddenly showing up on spaces.live.com that it triggered a flag in our system which requires more trust in individual urls in order for them to rank (emphasis is mine). I can see how people would be strongly tempted to put “trust” and “rank” together because they had so often already seen “TrustRank” used to refer to Google’s trust system.

As an outside observer, I feel it would be more appropriate to speak of a Google host trust threshold coupled with a Google URL trust score. The trust threshold would be the point where Google turns on a flag for a host that requires each individual URL be scored for trust. We can extrapolate that a host’s URLs will inherit its latent trust — a behavior many of us have observed over the years, and which I wrote about in On The Googleness of Being in February 2005, where I said: “…fresh content can now be shot straight to the top of targeted listings provided it’s posted on a Trusted Content Site which has little if any competition from other Trusted Content Sites.”

What the heck is a Trusted Content Site? I would say it is any host (domain or sub-domain) that Google places an inherent amount of trust in. That is, in order to improve the efficiency of its operations, Google seems to have allowed many Web sites to earn sufficient trust that every URL it finds on those domains or sub-domains is automatically treated as trusted content. But that isn’t TrustRank, which is a wholly different thing altogether. TrustRank is just a variation on PageRank calculated by crawling outward from a core set of “trusted content sites”. The only way to earn TrustRank is to get links from URLs that have TrustRank.

Google’s trust threshold seems to work differently. I can speculate on how they might score for trust in a multitude of ways, but let me suggest only one alternative to the TrustRank method to show that Google has many options for measuring trustworthiness.

Let’s say you have a domain called Xenite.org. That domain, established in 1997, has never been penalized for spam, has never been found to be excessively reciprocating, has never been found to be deceiving visitors with bait-and-switch (e.g., Content about “feed for horses” that redirects to a “buy viagra” page), and has generally provided a lot of interesting content to which many thousands of links from around the Web have been pointed. Okay, I actually have that domain.

Xenite may fall behind the times in design philosophy but I strive not to violate search engine guidelines with new content. I have worked hard to update my URL structures, I use robots.txt to exclude robots from directories I don’t want crawled, I don’t sell links, I don’t engage in manipulative reciprocation, etc. Algorithmically, Xenite has stayed pretty clean. If Google has a track record of Xenite’s cleanliness, Google could assign Xenite some level of trustworthiness simply for not having been penalized in any way.

Do they do that? I have no idea. All I know is that I don’t have to wait a year if I want a new section of content on Xenite to rank well for targeted expressions. On the other hand, if I set up a new domain (or even a new sub-domain on Xenite) I find I have to do a little work to persuade Google to let my pages rank. I need some links from Web sites that Google appears to trust. If those Web sites are willing to trust my content, Google will give me the benefit of the doubt.

If you have to call Google’s trust process anything, I think you should call it Benefit-of-the-doubt Rank. But that’s just me. You can lose trust. You can be deindexed. You can be penalized. You can be blocked from the Main Web Index and relegated only to the Supplemental Results Index. If Google decides your site is no longer trustworthy, you can suffer greatly.

But they don’t use TrustRank because I can also get new sites to rank without inbound links from Trusted Content Sites. How is that possible? It’s possible because Google seems to be willing to give new domains a fighting chance to earn some link love. Maybe it wasn’t always that way but that seems to be how it works now.

Links To External Sites And SEO

For some strange reason, a lot of people in the SEO community have really fought hard against the idea of using outbound links for search engine optimization. I’ve argued this point with some of the top names in the business: Danny Sullivan, Rand Fishkin, and others. They are convinced of what they are convinced and I doubt I shall ever change their minds. Nonetheless, I’ve been using outbound links for search engine optimization for years. I do pretty darn good, too.

Outbound links are, first and foremost, content. Now, in Google’s algorithm outbound links seem to do a fair number of things. You can say that 1 outbound link affects your rankings differently from, say, 10 outbound links. But if you assume I’m talking about PageRank, you’re wrong. Furthermore, you cannot look at a page with 10 outbound links as a “hub” or an “authority site”. A hub is a very specific type of page, as defined by John Kleinberg’s HITS algorithm. In order to be a hub, a page has to point to a lot of experts in the same topic. If you put 50 outbound links on a page to 50 experts in as many topics, you don’t have a hub.

But a hub also has to be recognized by a significant number of expert pages in its topic. That is, HITS all but requires tight reciprocation between hubs and experts. It doesn’t exactly require tight reciprocation, but in many topic communities that is almost certainly what happens. Google has never claimed to incorporate the HITS algorithm (Ask bought Teoma, which was built on HITS and CLEVER). But Google did hire Krishna Bharat, who helped develop the infamous Hilltop algorithm.

Search engine optimization has never quite recovered from the damage that Hilltop inflicted on the community, so I guess that many Googlers are feeling like that was a home run for their team. Hilltop proposes a methodology for choosing between pages on the same “host”. Its objective is to select the most authoritative document on any host. Many SEOs assumed (wrongly) that Google implemented Hilltop in late 2003 with the so-called “Florida” update. However, Bharat joined Google and helped launch Google News in late 2002, the year previous to what I sometimes call the “SEOs Blinded By Stupidity Update”.

Maybe I’m being too harsh on my fellow SEOs but presenting facts never went over well with the SEO community. Many a fact-monger like Mike Grehan, Shari Thurow, and Danny Sullivan has been trashed and/or ignored by the general community in favor of myth and nonsense. We’re all wrong on occasion about this stuff. But what is amazing about the so-called “Florida” update and the SEO community is that so many people have continued to be wrong for nearly 4 years. It’s time to let go of the Hilltop myth.

Not convinced yet? Read the story behind ‘Hilltop’. Bharat was frustrated when he tried to find unique news stories. Other sources specifically state that Hilltop powers Google News. SEO community: 0. Facts: 1. Like I said: Let it go, folks. Let it go.

Linking out to other sites benefits your site in numerous ways. It shows that your content interacts with the greater Web. Some years ago that was a major distinction between spam and real Web content. Spammers adjusted and linked out. Most SEOs were slow to adapt to the reality of outbound linkage and many saw their sites dip in and out of Google search results through the years because they only wanted to accumulate PageRank, not share it.

With your outbound links you can:

1. Make your site a valuable community resource
2. Connect with the greater Web
3. Build relevance to specific keywords
4. Earn trust

Yes, linking out helps you earn trust. Remember, in the old days many spam pages didn’t link out (although terminating pages were not automatically assumed to be spam, especially where frames were being used).

Search engines have refined their trust methodologies. Spammers have refined their trust-sidestepping techniques. The rest of us just need to earn trust and hang on to it. Let go of the PageRank and grasp the value of linking out to other resources. Many SEOs trap themselves in the Relevant Link myth, foolishly believing that their stes will be penalized if they link out to too many irrelevant sites.

Dudes, that is what the Web is all about. Link to content you like, not to content you feel compelled to link to for the sake of earning search engine trust. Trust begins with you, with your choices for link destinations. Leave it to the search engines to figure everything out. They do a pretty good job.

Google Doesn’t Index Pages With The Same Description

This particular myth may not seem like it has anything to do with internal links but there is a connection. The real issue is Google omitted results. That is, sometimes when you run a query you’ll see a short list of URLs and then Google places an annoying note at the end that tells you some results have been omitted.

Why does Google omit results? Because they are struggling to filter out unwanted or duplicate content. If you see an article on one page, do you really need to see it on 1,000 pages? Of course not. Problem is, Google doesn’t do a very good job of figuring out what is and is not duplicate content. So the Google Omitted Results issue has confused many people, including a lot of SEOs who spend their days doing site: searches for their clients.

You and I know that those pages contain unique content. Google seems to think otherwise. Why? Because Google places an inordinate amount of weight on the page title and meta description content for deciding what is duplicate content. Why do they do that? I think blogs (and possibly forums) are the culprits that led Google down the wrong path. Blogs (including this one) will, by default, repeat the same information in titles and meta descriptions. Search engines can become confused if they don’t look beyond the HEAD section of the page.

Google gets confused a lot, but they are probably cutting corners for efficiency. After all, there are many blog packages out there that generate tons of duplicate content — and many more autoblogs on Blogger that simply replicate other blogs’ content. For the most part Google does a pretty good job of filtering duplicate content. Still, a lot of people seem to have the mistaken idea that Google doesn’t index pages with the same description. They do index such pages. They just don’t always show them.

This Google omitted results issue has also led to confusion about the Supplemental Results Index. In the old days, before Bigdaddy, Gogole used to toss duplicate content into the Supplemental Results Index. But now duplciate pages can end up in the Main Web Index because they have links pointing to them (often nothing more than internal links). Hence, the duplicate content filters may block pages in both indexes. A lot of SEOs haev been thinking up until recently that the “Google Omitted Results” link meant you had pages in the Supplemental Results Index, and that’s just not so.

High Quality Links: Internal Link Structure In SEO

So if you want help with internal links you need to stop and ask what you are doing with internal links. Do you know how to make high quality links or do you just assume that internal link structures are not important? Nothing is more critical to your search engine optimization success than the need to create as much content as possible and to interlink that content as much as is feasible.

Your high quality links begin at home because the highest quality links are the links that build the most trust and visibility. You cannot stop the flow of PageRank. How PageRank flows through your site is determined solely through your internal linkage and the better you manage that flow the better your search engine results will be. PageRank flow cannot be measured by the Google Toolbar but it is what governs your crawling activity and it also determines which pages go into Google’s Main Web Index and which pages go into the Supplemental Results Index.

You have to stop obsessing over unproductive nonsense like how to preserve PageRank because you cannot preserve PageRank. You have to learn how to construct powerful pathways through your content so that people can get to each place in as few clicks as possible. You don’t want to confuse your visitors with too many links on each page. You just want to guide them through your site until they arrive where they need to be.

You don’t need to find supplemental pages you just need to link to all of your pages. Build a good internal link structure and you should do okay.

Remember the SEO method: Experiment. Evaluate. Adjust.

That will get you there eventually.