SEO Case Study: Link Building - natural anchor text

2009-08-04 Christoph C. Cemper

04 Aug

Home » SEO Case Study: Link Building – natural anchor text – PR0 vs. PR4 strength

SEO Case Study: Link Building – natural anchor text – PR0 vs. PR4 strength

Do you want 20 PageRank 0 links or 20 PageRank 4 links? The following contribution will help you to answer this question. Today’s topic is a case study on one of our big clients from the real estate market. It is about the variation of anchor texts, natural anchor texts and the value of PageRank 0 links. It makes clear that PageRank 0 links could be more potent than PageRank 4 links. Due to this we assume, that PageRank cannot be used to evaluate the quality of the links.

There is already an article about PageRank Myths in Linkbuilding where we differentiate between the “Toolbar PageRank” (the one we see and we write in this article about) and the internal parameter which Google uses. This internally used parameter is sometimes called PageRank but is not visible outside the system.

Even Trusted Domains don’t get away with everything

The following case study is about a long term client from the real estate market in Germany. This specific client is in Germany very big and has a ten year old domain with over 100.000 links to it. The Backlinks were organically developed over the last 10 years and they are doing really good. Last year we started a new project for a niche topic on their site related to their business.

Should be no problem,or?

A Penalty for 10 links?

A Penalty because of 10 Killerlinks to a Sub-page?

We started the project with 10 links a month. In order to kick-start it very softly we developed persona links. Those were 10 links from relevant pages, in-content links and all of them showing PageRank 0. All backlinks were follow, which bequests PageRank and counts for Google.

These few links should be okay – or?

How you can bring down a page with 10 Killerlinks

+ 30 Penalty

We noticed after 2 months, which was 20 links, we tripped the automatic filter for that keyword for that single page. The client explicitly wanted us to link with one generic keyword to the specific page. We did it because we thought due to the fact that 10 links are few, this should be no problem (after this experience we dissuade from doing so). Before we started the project the page already ranked in Google thanks to the trust of the domain. After the 2 months we hit a +30 filter for that certain keyword and for that page. It got a penalty of plus 30 in the search results of Google, although the links were PageRank 0 – why? The links were relevant but the problem was the anchor text.

How we solved the problem for the client

We took a closer look at the 20 links which were PageRank 0 but juicy, our Linkjuice tool identified them as juicy. They were counted by Google and tripped and automatic filter for that keyword. It was definitely over optimization. The solution for this problem was to:

vary with the anchor texts
change the anchor texts to natural language phrases
e.g. link the brand name
e.g. link phrases like “this page”
even place nofollow links

Nofollow links are natural

Yes, nofollow links are good because they are natural! The normal Internet User does not know the difference between follow and nofollow links. In order to see NOFOLLOW links a specific plug-in for the browser is needed and only SEOs have installed it. The probability that a page only gets FOLLOW links, with the same anchor text, on top of that with such an expensive keyword, equals zero. Nevertheless if this happens it is not natural.

If you take a look at the distribution between follow and nofollow forums in the web you can see that half of them have NOFOLLOW activated. In order to blur the tracks we placed NOFOLLOW links and varied with the anchor texts. What happened was that after 60 days the penalty or actually the automatic filter was lifted. The page went back to the search page number one in Google. It was even better than the day before it tripped the filter because it went up one position. The lesson learned is that Google obviously counts those PageRank 0 links if they are relevant, if they are in-content and if they are on relevant pages, which we find with our Serp Research tool regardless of the PageRank.

We did other tests and took 20 PageRank 4 links and gained nothing. It had neither a positive nor a negative effect. It was just a waste of time and money. If you think about it makes sense because PageRank 0 Backlinks came from relevant pages, they were developed over time and they looked natural and the other links did not, although they show PageRank. We recommend to vary your anchor text and don’t be greedy about PageRank because simply PageRank doesn’t matter anymore in link building. It is not an indicator for the quality of a link because it doesn’t say anything about topical relevance and if the link is counted.

Now you know that varying with the anchor texts is very important and besides this you know that PageRank is not an indicator for placing a link there. Of course this topic could be scientifically evaluted further but this is not our aspiration. All we want is to share our experience with you and to care for our clients.

We are aware of that some agencies, SEOs and “experts” assign essentially more value to the Google PageRank. We simply cannot agree to that and warn of a unilateral use of the Google Pagerank.

Top

Loading ...

Buffer

37 Responses to SEO Case Study: Link Building – natural anchor text – PR0 vs. PR4 strength

syn
August 4, 2009 at 11:17 pm
well most filters are lifted within 60 days. i think the issue was, that there seems to be only 10 links a month, if the links would have been spread more (with some “natural” variation) there would not have been a penality. make these 10 links, make them (and their anchor text count) but then spread juice links all over the place.

[Reply]

admin
August 4, 2009 at 11:24 pm
@syn: these 10 links / month were part of a much bigger link growth of over 130 new links per month, but all those other links went to other pages, had other anchortexts etc

hence I believe that indeed THESE 10 links caused the penalty (on that one page!)

[Reply]

SEO Company UK
August 8, 2009 at 3:10 pm
ok what i want to know is PR is tht really matters i have seen sites ranks in 1st page in google without even a PR 1.

[Reply]

Travel Affiliate
August 10, 2009 at 8:39 pm
Thanks for sharing.
Maybe Google have this trigger in place because they want to bust the typical link exchanges where the link page contains specific html code containing the same anchor text for webmasters to use for the reciprocal. Maybe ?

[Reply]

Light Rucksacks
August 14, 2009 at 9:31 am
Some interesting points raised here. I wasn’t aware that filters are lifted after 60 days, I thought they were permanent?

[Reply]

wordpress review plugin
August 19, 2009 at 4:43 pm
That is indeed a very interesting case study, what I doubt is that the persona (relevant and juicy links) that you got for the client were these paid links ? it is a possibility that Google thought these links to be paid and discounted them and you guys got hit with a penalty

[Reply]

Gia xay dung
September 21, 2009 at 3:46 am
And the end of content : anchortext in nofollow link is need to ranking with google ?.

[Reply]

Hardest Game Ever
September 21, 2009 at 12:32 pm
This is so true. Pr doesn’t matter, and it never did. People should stop wasting their time looking for high PR pages. Natural link building is the way to go.

My advice is to find keyword phrase for every page on your website, and get the links with anchor text containing that phrase. Use 60% exact match, 30% phrase match, and 10% broad match.

Thanks

[Reply]

Paul @ Hybride Auto's
September 21, 2009 at 8:22 pm
It’s interesting because this algo can be abused by competitors. Create 100 links with identical keyphrases and point them to your competitor. That’s scary…..

[Reply]

SEO-Doctor
September 25, 2009 at 6:20 pm
@Paul – this is why this theory/case study needs more testing. If point 100/1000 crap links at a competitor like that, they will just be ignored.

[Reply]

admin
October 1, 2009 at 9:27 am
@Seo-Doctor: absolutely NOT! In our case study we had 10 links to a 10 year old domain with 100,000 links! Of course it depends on a lot of other variables like on site topic/page topic…

@paul: YES, abuse to “google bowl” and harm a competitor is a very common practice in 2009 … we plan to provide more statistical data on that soon

@Hardest Game Ever: it’s not 60/30/10 – numbers vary for every industry, niche, even sub-page

best,christoph

[Reply]

Travel Holiday Links
October 16, 2009 at 12:45 pm
Google says that competitor can not harm your ranking: http://www.google.com/support/webmasters/bin/answer.py?hl=en&answer=34449

But it is oposit with your theory of vary anchor text and -50 penalty.
??????????????

[Reply]

Christoph C. Cemper
November 3, 2009 at 1:23 am
@Travel – well, you shouldn’t believe Google knows everything and tells everything

“There’s almost nothing a competitor can do to harm your ranking or have your site removed from our index. ”

Almost nothing doesn’t mean “can not”

Plus there are some ways not discussed here that are even worse.

[Reply]

lice-nse to kill
November 11, 2009 at 1:03 am
Great info. I actually printed out this whole page. I have always thought varying up the anchor text is a great thing to do, and was afraid to only put 1 main keyword for my homepage and now I know i was right all along. thanks!
Scott

[Reply]

SEO Scotland
November 24, 2009 at 2:48 pm
Most clients want (Google) pagerank, rather than page authority and this is difficult to explain unless you have a case study, perhaps the trick is to aim for both; pagerank and page authority?
Obtain the ‘natural’ links from pagerank 0 site which have authority and target the pagerank the client wants too.
More work for you, but this way you can prove the technique works.
Cheers Chris

[Reply]

Ryan
November 24, 2009 at 9:30 pm
This case study is bogus! If this were true then you could just do this to your competition. Please… you honesty think 10 links with the same anchor text would get your site filtered? Another thing that doesn’t matter is relevancy! I have tested this many times and it doesn’t matter. The anchor text is what will boost your site in the serps. Keep linking with “click here” and I will keep dominating the serps with my sites.

[Reply]

pictures of head lice
December 29, 2009 at 10:01 pm
I always thought the QUALITY was far more important than quality of links. Thats why some sites only have 10 links but are top of the game in Google, not for competitive keywords though, mind you. But it needs to be RELEVANT to your site, helps anyway, and if not relevant, on a page w/ decent PR and do follow of course is better…
Thanks for the info!
Mike

[Reply]

tim
January 4, 2010 at 7:18 pm
I think that good (long tail) keyword and backlinks do the trick
I mix them up so i get no follow and do follow to get it more natural.

@Ryan: absolutely right on the anchor text

[Reply]

Optimization Specialist
January 6, 2010 at 8:05 am
Thanks for the great Case Study about the value of pr0 vs pr4 domains and the importance of using natural anchor text for link building.

[Reply]

Shaun Rosenberg
April 7, 2010 at 6:05 am
Makes sense, I still believe a higher PR link is more powerful then a lower PR link, but they both do count, that’s for sure.

Google probably has a lot of different factors that they use to determine how strong a link is, PR, relevency, number of links on the page, etc.

[Reply]

Internet SEO
July 6, 2010 at 7:26 am
PR alone cannot help a website to get high in the serps. it takes alot more than PR.
Trying to get higher on serps is worth a while than farming for PR.

[Reply]

Dixon Jones, SEO.
August 11, 2010 at 4:34 pm
Hi Chris,
I think that it is not just the link map that can inflict the +30 – it can be a combination of SEO factors combined. In my experience (and this only happened the once) we had what appeared to be a plus 30 issue which probably did materialize as a result of anchor text optimisation – but we seemed to be able to fix it with de-optimizing the onpage context a little. Now I would not hold my hand up and say that the testing has been done to 95% confidence levels on this, because it could have beena fluke, but I suspect the correlation between anchor text and on page text is a potential factor in the plus 30.

[Reply]

SEO Scotland
August 18, 2010 at 10:48 am
So Chris, now that Google (Caffeine?) is showing more links in GWTs, does this change the way you build links, or do you find this helps, as the anchor text density is more visible?

…and now we can see there are more links, should we still use the same build ratio?

[Reply]

Wesna
October 11, 2010 at 9:18 pm
Great post! some of my clients, that know a little bit about link building are not to understand why I do nofollow links, why links from PR0 pages still count , and they refuse links like that !!! I think Im gonna make them read this artcile

[Reply]

Matthew Bailey
December 1, 2010 at 7:43 pm
The world of SEO is a crazy place. Keeping your site natural, informative, and full of changing unique content is a great way to rank. Thanks for sharing this case study, I have learned a lot.

[Reply]

james
December 3, 2010 at 2:25 pm
cant help but think the reason for the penalty was for more than simply 10 anchors; if this is all that was required to be put into the filter than if would overnight cause everyone to link ten poor anchors to your competitors above you. Google would never have an algo in place that would cause this, its too short suited and irrational. I beleive it was probably a small factor amongst many others

[Reply]

Pay on Results SEO
December 18, 2010 at 6:45 am
PR doesn’t really matter too much in this day and age. PR0 links do work and get top rankings. A variety of different links creates a natural link profile which is critical in a link profile

[Reply]

Hermes
December 22, 2010 at 12:17 am
I was having very hard time with a ripoffreport.com page ranking 2nd under a client’s site for his own brand name.

After reading your post, I pushed it to the 5th page and it took around 60 days after the links were created.

I think it work only with competitive KWs and the links created need to be CLEARLY spammed

[Reply]

Jan
December 22, 2010 at 11:11 am
Like a good Cocktail, just the right mix makes it perfect. Many SEOs forgot to fill up the backlink pool with “real organic” links like “Homepage”, a full phrase or the brand.

Thanks for this good case study!

[Reply]

Fred
January 2, 2011 at 11:23 am
This was extremely helpful. I believe we tripped an anchor text filter ourselves on one of our pages. We consistently ranked #s 3-6 for the main phrase (single and plural) – think “Red Widget” and “Red Widgets” then all of the sudden after some link building we fall to #s 17 and 22, when for more than a year we’ve been at at least #10 for the plural term.

In our case, it does not appear that this is necessarily a “-30″ filter… but it is definitely a filter. We have more back links from relevant sites than the competition.. but we have consistently used similar anchor text.

Any thoughts on whether Google just devalues the links coming in from the identical anchor-text sites? E.g. – just stops counting those links? Our link profile probably has 90% same link text, 10% varied. It appears as though the 10% might still be getting counted as we didn’t drop a normal “penalty amount” of 30, 100 or worse.

I have some questions that I’d love if you could address:

1) We have influence over some of the links we’ve gotten to this page. Would it be wise to go back to those web masters and ask them to slightly vary the link text in the existing links?

2) What specific proportion of links do you recommend for healthy link building? Should we be targeting 50% exact match? Less/more? I know this isn’t an exact science but I’m looking for more of what the rule of thumb is here.

3) Have you had any similar experiences since posting this video? I know this vid is a bit old and I’m curious if you’ve had other encounters.

Thanks,
Fred

[Reply]
leap Reply:
January 21st, 2011 at 6:16 pm
Hi Fred!

Sorry to hear you’re having troubles with your Link Building Campaign!

To address your questions:

1) If you change all your new links, you have a serious footprint problem. Google sees that after you dropped, you changed the anchor texts of some links and knows exactly which links you bought. You still might want to do this, but only for a very small number. It’s better to focus on new link building with anchor texts that compare better to your competition.

2) Which answers the second question: There is no one correct proportion of links. This is different from topic to topic, actually even from keyword to keyword. The best proportion is the one that’s being used by the top competitors for your keyword. Use a tool like our Link Research Tools to look at the anchor text and link status proportions of your competitors and try to emulate them.

Kind regards,
Christoph

[Reply]

Fireworks
July 21, 2011 at 7:34 pm
I’m not so sure of PR in this day and age but it’s still a ranking factor in terms of rankings and link building

[Reply]

Madbid
October 3, 2011 at 12:27 am
Actual real life examples like this into how page rank and anchor text works are fantastic. Far more interesting to see this rather than read rumours and heresay

[Reply]

Move Ahead Media Australia
October 7, 2011 at 4:09 am
Varying anchor texts is very important along with just having your company name or simple URL instead of an anchor. If genuine people were linking to you, would they always but your keyword? Probably not, so try varying it which will help. About PR, there are so many factors to take into consideration which are now classed as more important than PR.

[Reply]

seo consultancy
October 8, 2011 at 9:02 pm
ive suspected for a long time that PR links havnt been that important and ive found that having more links from as many different websites/domains to be more effective that high PR links

[Reply]

Move Ahead Media Bangkok
November 8, 2011 at 5:03 pm
I still think that PR, to a degree plays a part. But as always there is more than 1 way to skin a cat. Different links, blogs, press releases, social media etc. A wide array will help you rank

[Reply]

Marc
June 19, 2012 at 8:08 am
Hi Chris,

Nice article. I agree with Dixon that in some cases a -30 filter can be a combination of anchor text + on-page factors, de-optimisation on-site has definitely worked for me once before, but don’t know much more than that!

Also something not mentioned here, which goes along with some of the comments is new vs existing pages. In all likelihood a link on PR4 page has more value than on a PR0 page, but a lot of variables come into play, alongside the obvious – Domain Authority. Was the link added to an existing page which already had inbound links pointing into it, or was the link placed on a fresh, PR0 page which gained links over time after first indexation? I have not tested anything, but would like to believe this would be the best type of link. Maybe it’s not about PR at all, but getting links in newly crawled, fresh content?

[Reply]