The relevant link myth
Posted by Michael Martinez on June 18, 2007 in Seo Myths
Relevance only counts in search engine results.
Unfortunately, somewhere along the way the search engine optimization community came to the misguided conclusion that relevance is important for links. Many SEO experts now routinely advise people to get only “relevant links from relevant pages”, and wherever possible those links should only come from pages that themselves have “relevant links pointing to them”.
If you want to optimize for Ask.com, that’s probably a good philosophy. If you want to optimize for Google, Live, or Yahoo!, that’s about as inefficient as you can get with your link building. And link building is already the most inefficient means of optimizing for search engine results.
By definition, all links (that pass anchor text) are relevant. If someone with a poker site links to your “Care and handling of Asian camels” page, where the link passes anchor text, the link is relevant to your page.
The anchor text does not have to have anything to do with the topic of your page. Remember, Google’s system is built upon the faulty conclusion that inbound link anchor text can be trusted to accurately describe the contents of a Web page. Any search engine that allows links to pass anchor text also subscribes to this enormous falsehood.
So it is virtually impossible to acquire an irrelevant link. The search engines don’t recognize the concept. Even so, search engine optimizers sagely advise all readers to only get “relevant links from relevant pages”.
In essence, what the SEO collective wisdom is telling us to do is acquire links only from competitive pages, and yet linking to your competitors with anchor text relevant to their topics will only help them more than it helps you.
Relevant anchor text is a subtle metaphor for “manipulate search engine results by spamming your page to the top with repetitious link anchor text”. After all, the search engines treat the anchor text as if it is part of your on-page content. Hence, the more you repeat that anchor text, the more relevant your page will seem in the scoring process to those keywords.
The idea that a link may carry greater weight from a page that is devoted to the same topic as the destination page is another bit of nonsense that SEOs have adopted. How many of you would turn down a link from the front page of Yahoo!? I see no hands. Apparently, relevance isn’t all that it has been cooked up to be after all.
If CNN started linking to your car insurance affiliate site from their front page, would you contact their Webmaster and say, “Please stop linking to my site — CNN has nothing to do with car insurance”? Convince me that the relevant link from only relevant topics is really what you are looking for.
If you get 1,000 DIGGs tomorrow and as many links from a broad selection of Web sites, will you ask Google to let you specify a “NO LINK CREDIT” list in your robots.txt file so that it doesn’t pass anchor text from unrelated sites? I see no hands. Apparently, relevance isn’t nearly as important as natural linkage.
I can only guess how much longer SEOs will continue to babble on about “relevant links from relevant pages”. If you want to maintain a competitive advantage over the majority of SEOs, then you should hope they continue to babble like that forever. But in time common sense will prevail over collective wisdom. After all, eventually people will put 2 and 2 together to see that “22″ is a much larger number than “4″.
This is not a semantic issue. It’s an issue of perception. The perception has been shaped by a distorted interpretation of why some sites fail to achieve good results with their linkage. Those sites that engage in “excessive” reciprocal linking, or that only exchange links with sites that really have no value in themselves (no original, interesting content, no significant linkage), have no hope of being recognized as valuable sites by Google.
Just because people have engaged in cheap link building strategies with low-value sites doesn’t mean that all links from unrelated pages are harmful or unhelpful. Quite the contrary: most such links do help, and in many cases they help better than so-called “relevant links from relevant pages”.
When you’re studying a Web site’s backlink profile, don’t ask yourself how relevant the links are. Ask yourself how useful the sites linking to the destination really are. You’ll most likely find that the “useful content” test is a far better indicator of quality than the “relevant content” test.
10 Comments on The relevant link myth
By MattUK on June 18, 2007 at 8:34 am
Hi Michael,
I agree that link relevancy isn’t algorithmically scored in the SERPS, it’s debatable if this is actually possible beyond looking at the link relationships between sites listed in the same ‘neighbourhoods’ and their authority/trust status.
Aside from that where I believe that relevancy is worth chasing is when it’s a case of flying under the radar in terms of a clean looking link profile. There was a pretty high profile case over here in the UK a little while ago where they got penalised for having massive amounts of .edu pages linking to them that were pretty obviously out of place.
In terms of an algorithmic ban than you’re right, relevancy isn’t worth chasing, but in terms of avoiding a manual flag and removal then it’s always worth keeping your link profile as natural looking as possible.
By Halfdeck on June 18, 2007 at 2:27 pm
Irrelevant links from an irrelevant site will pass juice and anchor text as long as Google doesn’t suspect that the intent of the link is search engine manipulation. In short, lack of relevancy might make Google’s detective work a little easier, but there’s no reason for Google to devalue an editorial link just because it points to an off-topic site.
It’s interesting to me that adult webmasters, for example, are obsessed with relevancy to the point where an owner of a “lesbian porn” site will refuse to trade links with a “milf porn” site on the far-fetched assumption that Google can tell the difference between over 70+ different niches in porn.
By Sergei on June 18, 2007 at 3:37 pm
Hi, Michael
I agree, that “relevance” of the link is not only based on “relevance” of page with the link. The factor of “relevant links from relevant pages†is not fundamental. For example, a link from Washington Post is much better that “featured” link to your site from niche site with duplicate content.
But I think, that gaining of “relevant links from relevant pages†is looks like nature of linkbait, so it’s stays on. It’s not a myth.
Search engines bots are stupid, and they can’t to measure exactly the relevance of page without analysing of paths to it.
Just my 2 cents.
By chelovision on June 24, 2007 at 3:04 am
@Halfdeck
Its probably true that some adult webmasters will not trade links for the reason you state, but I think that is a case of too much undisciplined acquisition of inaccurate information. Those webmasters would be conflating Google myth with their industry’s best practices, which is pretty common in non-adult SEO as well.
Adult webmasters do have a very good reason not to trade links with sites outside their niche–conversion. Bandwidth is still an issue and your granny lovers arent going to be too interested in your cheerleader offerings (and if you have granny cheerleaders, well you have a new niche…)
Mr Martinez - Finally! It’s high time you stopped wandering hither and fro agitating the forum owners of the SEO world…
By Cumbrowski.com on July 1, 2007 at 12:22 pm
Michael,
The relevant link babble is a valid advice to companies who have a brand to protect and a public image to look out for. Linking to a competitor is of course nonsense. There are plenty of other relevant pages out there that do not belong to a competitor.
As Tetsuto pointed already out, links on relevant site make it hard to impossible for Google to suspect fishy business and an attempt to game the system. Google has no real grounds to penalize anybody for placing a link to a page that is relevant to the subject of the page the link resides on and benefits the user who visits the page. Even if the link is paid for (in one form or another).
You are right that it does not matter TODAY if the page that links to you is relevant to your site or not, as long as the anchor text is relevant and the site that links to you is still able to pass on link juice. This is very likely to change in the future and make people regret some of the links they have out there, which were not detected by Google before that. They will come back to bite you and the whining will be great.
If you are not in it for the long run and want to make a quick buck here and now and are willing to gamble and take some risks to get stopped right in the tracks before you are able to cash in, by my guest and get links from any sucker out there who is willing to link to any site regardless of content as long as the money is right. Make as much money as quickly as you can and when detected (manually or algorithmically) run and start from scratch again.
But you know that already. You got some links out of it (relevant links btw.), which was the whole idea of the exercise. Good Job!
By Eric Itzkowitz on July 9, 2007 at 5:18 pm
My thoughts:
1) Linking strategy depends solely on your business business model and its goals. Do you want your website to stand long and true, or do you want to take the money and run (a few times a year).
2) Obtaining a relevant link, as opposed to just any old link, gives you the best shot at converting a visitor into a buyer/subscriber (qualified traffic).
3) A #1 Rank is worthless if it’s not going to mean new customers/subscribers.
4) A well-respected, long-standing brand has to be very careful of where it proactively acquires links and traffic from.
By leondz on July 18, 2007 at 4:09 am
Can we get any authoritative sources or the results of any vaguely well conducted experiments in to back this up?
By Michael Martinez on July 18, 2007 at 4:14 pm
Every outbound link on this blog is carefully executed for an SEO reason. As Mr. Delrio pointed out last week, I am executing SEO theory with nearly every blog post. Watch what I do carefully and check the search results.
That is, I am afraid, as “authoritative” as my examples can get. And, of course, as people watch those SERPs and see value in them, the SERPs will change due to other people’s optimization.
This blog IS an SEO experiment (many of them, actually).
By Justin-Goldberg on September 6, 2007 at 9:35 pm
Any chance of giving out the results of these tests?
By Michael Martinez on September 6, 2007 at 10:28 pm
Justin, SEO Theory does not experiment with link anchor text.
Comment
Log in or Register to post a comment.