Is There a new Anchor Text Strategy Because of Google Phrase Filter?

Does anchor text need to change?

[sandyeggo]

I am placing some links, all targeted and only 50 at a time. I was going to place all of the links to one of my deep pages using one "two word" phrase for all of the links, but because of this new phrase filtering tool google is using i have been a bit apprehensive. Should I use same anchor text on all 50 links or should i split them up and either:
- change anchor text among links
- not place so many links at a time for any one phrase
- target main web page only

or any other ideas? I know this has probably been discussed, but since i got caught up in that last "minus 950" penalty and have since recovered - i want to try not to get caught in up any phrase filtering algos.

[Marcia]

Excessive use of identical anchor text not being a good idea dates back to before these patent apps were even filed. It's alway best to use a percentage of variance, and it's been that way for a number of years.

I really hate to see panic setting in over something that may have no relation to any particular problem.

[sandyeggo]

i always spread links around in the past, but i was usually running about 50 at a time in a target campaign with same text - and because i got caught up in that 950 penalty and have since recovered, i wanted to know what others were doing lately and if any changes were being made to the way they plan their incoming links.
So if I were placing 60 links right now, what would your percentages be if i were splitting up anchor text/url?
25 percentish?

[Marcia]

Sandy, the "phrase" thing is not all encompassing or near as much as it's being made out to be. And I wouldn't look upon link analysis as the same thing, anyway. There are a lot of preceding IR concepts behind the "phrase" patent apps historically, and the surface hasn't even even begun to be scratched yet, except for some minor rewording of what's in the papers already, which doesn't really clarify anything.

You can't "optimize" for phrase based algo components and neither can a link building campaign be "optimized" for it beyond what's been the norm for a while. Get on topic links gradually and normally, vary anchor text to a degree, stay away from bad neightborhoods and linking schemes.

[northweb]

but since i got caught up in that last "minus 950" penalty and have since recovered - i want to try not to get caught in up any phrase filtering algos.

I would be causious adding that many links all at once. Excessive anchor text has been discussed as one of the possible problems in regards to penalities.

What did you change on your site to have the -950 penalty lifted?

[sandyeggo]

good info - thanks marcia

would you still place deep links or would you just keep it narrowed down to the main index page? In the past I have mixed it up.

[trinorthlighting]

Make your links for your users, ask yourself how they want to flow through your site instead of googlebot. Sometimes SEO's worry to much about google and not enough about users. If your users hate your link scheme, they will bounce out of your site and google will tank you eventually. If the users like your link scheme and do some serious surfing, then google will reward you.

[Marcia]

I'd mix it up to a degree, with on-topic links to primary relevant pages, homepage and otherwise. There are dozens of papers out there on link analysis, but the crux of it all is that the better the quality and relevancy of the linking page/site, and the more the linking page substantiates the topic of the page linked to, the better it is.

The word "related" is nothing new, and a link from a page that's related by topic or niche or even by interests of the target audience (i.e. Amazon's "people who bought this also bought this...") always makes good sense.

[Whitey]

Marcia - Thanks for re affirming this with such simple clarity.

[sandyeggo]

very helpful
thanks

[nonni]

I agree with the above advice, but am not sure about part of the question (phrase based). I just ran a bi-word and tri-word analysis on leading G results for one keyword, and saw no evidence of any intelligent pattern of phrase analysis. One of the top ten was an entry page - a photo, click to enter button, and a huge amount of black text on a black background. Key word stuffing from the early days of SEO - Sheesh.

Maybe the phrase analysis is being used in some places, but it is far from widespread, IMHO. Hopefully they are moving towards that, but not there yet.

[Marcia]

nonni, are you sure that doorway page wasn't cloaked? Or wasn't obfuscated behind a "noframes?" Did you check the cache and run it through a translator?

[annej]

sandyeggo, are you talking internal links? Why would you want 60 pointing at one internal page? I'm just wondering if you are diluting things with too many links every which way.

I don't think there is a problem with linking to a page with identical anchor text from related pages and in reasonable numbers. The exception would be a phrase that just happened to be what I call a problem phrase. You know it might be that because the page's search result is suddenly at 996 or some such thing. Then it's worth trying varying the wording of your anchor text to see if will help the page come back.