Over 90% of anchor text tends to be brand-name on OSE link profiles. Why?

[ZakGottlieb71]

I reported this as a bug in OSE, because often I explore these links and find that the pages include both a brand-name link AND a regular keyword link, but for some reason OSE was only reporting the brand-name link...

This led me to wonder how many links this occurred for, and therefore whether or not to trust the fact that the majority of the sites I ran OSE on returned at least (in most cases, more) than 90% brand-name links.

I understand that brand-name links are amongst the most important to obtain, but that it's also important to get anchor text for keywords to build a varied profile. Given this apparent flaw in OSE, is it wrong - in the case of very successful sites - to take this ~90% as being anywhere near the correct percentage of brand-name links that I should be aiming for as a proportion of the total profile?

Extra Credit :)... And this may help potentially help resolve the issue: does "Inbound Links" tab in OSE just report links to the Root Domain, or to that and every other page on the site?

[ZakGottlieb71]

Thanks Cyrus, you've been incredibly helpful.

[Cyrus-Shepard]

Manism is one word and will most likely be interrupted as such by the Search Engines. Since it is thematically related to "man" it might still pass some benefit if contained in anchor text.

Whereas "askmen" is interpreted as 2 separate words. This might pass more value, but the effect is most likely minimal as Google becomes more sophisticated.

[ZakGottlieb71]

Thank you, I'll be sure to check out your video. Very final question then: If we're saying "askmen" is a keyword rich domain, could the same be said for "manism," for a site dealing with "men's issues"? i.e. Google recognizes the "man" in there?

[ZakGottlieb71]

Thank you, I'll be sure to check out your video. Very final question then: If we're saying "askmen" is a keyword rich domain, could the same be said for "manism," for a site dealing with "men's issues"? i.e. Google recognizes the "man" in there?

[Cyrus-Shepard]

1. Ideally, but not always the case with smaller sites, local sites and folks who can't get their own branded domain. But you are correct, there's a great deal of difference between a keyword rich domain and a non-keyword domain.

Unfortunately, there's no "correct" answer to the ideal ratio, although 90% branded seems incredibly high to me. In fact, I wish fewer of SEOmoz's inbound links were branded, as we could benefit from more variety.

Here's a video I made on anchor text

Along with some related resources which address the topic:

• Are We Over Engineering the Link Graph?
• Backing Up The Brand – Are Over Optimised Link Profiles A Barrier To Top 5 Rankings?
• How to Unnaturally Naturally Vary Your Anchor Text

[ZakGottlieb71]

Thanks, Cyrus. Let me just ensure I understand correctly:

1. "The Domains are exact match for their brands." Isn't that nearly always the case? Or are you referring specifically to a highly relevant keyword being present in the brand name (in which case, we're saying Google can detect the word "men" within "askmen" and give "credit" for it?)

2. Understood.

Beyond both of these, I'm still curious why 90%+ of inbound links should be brand name or close variants. You wrote:

"In the end, you want anchors closely associated with the keyword topics you are trying to rank for."

...Yet again, it appears that a 90%+ proportion of SEOMoz's inbound links are ALSO brand name, or close variants. The takeaway I'm getting is that this ~90/10 split is what I should be focused on in order to achieve ranking success. So I guess my final question is, is it wrong to think that?

[Cyrus-Shepard]

Hi Zachary,

I apologize because I misunderstood your question the first time.

Going back to your original question....

often I explore these links and find that the pages include both a brand-name link AND a regular keyword link, but for some reason OSE was only reporting the brand-name link...

OSE will report the first link found on a page's HTML, and only the first link. This is consistent with testing done on how Google counts anchor text. (If you find something different, it would indeed be a bug)

As for the prevalence of branded anchor text, I think there are a couple of factors at work here:

1. The Domains are exact match for their brand, which accounts for the high proportion of branded anchors. See the stats for SEOmoz
2. Large online publishers like the examples you cited tend to belong to publishing groups that own a large network of sites. The sites tend to interlink to each other using the same anchors.

Having a keyword rich brand name can really help you in the ranking department. So can developing search volume for your branded terms.

In a vaccuum, I'm not sure branded anchors pass greater or lesser value than others. In the end, you want anchors closely associated with the keyword topics you are trying to rank for. If these are your brand keywords, all the better.

[ZakGottlieb71]

Thanks for the great answer. I am still a bit perplexed, however, and here are those examples to clarify:

www.askmen.com

www.artofmanliness.com

Just taking the first example - If you plug that in, you'll notice that under the "Anchor Text" tab, almost all of the keywords with an even vaguely significant number of links to them are brand-name or a close variation of this. There is the occasional "cars on askmen," or "fashion on askmen," but even these hardly seem to be very carefully selected.

Askmen is an enormous site with a PR of 7. Can this really not be taken to imply that brand-name links are significantly more (or at least, no less) effective than targeted keyword links, and one's campaigns should thus be heavily weighted in their favor?

Thanks again for your help, and thanks Cyrus.

[Cyrus-Shepard]

In addition to Robert's answer, I'd like to add that although OSE doesn't have a built in bias towards "name-brand" links, it does favor links with high Page and Domain Authority, and URL that have good inbound link profiles themselves, because these are the most useful to crawl and index.

Of course, OSE doesn't report all links, especially those buried beneath layers of navigation and those on pages with very few inbound links themselves.

Part of the issue may be temporary as well. Each index changes from the last. The most recent index was an improvement in that it reported far more domains, but significantly fewer URLs. This is a trend we may see reversed in future updates.

More about the last update here.

[RobertFisher]

Zachary,

It is a bit difficult to assist without knowing at least a url you are looking at. I just used OSE on a couple of sites we managed to see if there was some type of bug, but it returned to them money what it should have.

When you are looking at anchor text on OSE, you can select a term or phrase and it will make a difference.

When you have clicked on the anchor text tab, just under the url will be two drop down choices:

Show anchor text TERMS/PHRASES for link to THIS PAGE/ALL PAGES ON THIS ROOT DOMAIN/ALL PAGES ON THIS SUB DOMAIN then Filter. After you select what you want, hit Filter.

On your extra credit, When you have clicked on the Inbound Links tab, Look below the URL box and you will see another set of choices. You can see a choice for type of link, from all or internal or external pages, to any page or all pages on the root, sub, or page. You may also group the result by domain.

Hope this helps you out.