Links, do they sitll work? - Test Case

It is really simple question, do links still work?

I found out some interesting things from this ranking test case and I will share it with you.
I will write all the info about this ranking test and its results, so you can draw your own conclusions.

Little intro:
This test case is mainly about influencing Google SERPs with links and various anchor texts.
The purpose of this test is to discover keyword rankings in these situations:
- rankings of pages/keyword with no links
- rankings of pages/keyword with links, same anchors used(main keyword)
- rankings of pages/keyword with links, various anchors used(main keyword, generic keywords and its various %)
(And accidentally we will see some meta title/description situations as well)

I started with this ranking test case/experiment on 02.04.2013., before this last penguin 2.0 update,
so luckily we will be able to see results before and after penguin 2.0 update.


How this test was conducted:
- For this test case I've used non existing keyword "prekintrebalings",
so there where no search results/competition on Google prior to this test took place,
so this way we can really test what is working and what not in controlled environment and without any outside influences.

- I've made 4 same wordpress blogs with same templates using same title("prekintrebalings") and without any content on them, so they where identical.
Later I found out that meta description on 2 of them(out of 4) was different and because of small meta title,
google used this meta description as "addon" title to fill in the short title.
Later almost in the end I made those 2 meta descriptions same as others, to have "google titles" the same as well and results changed,
I wasn't surprised, but for some time now title is treated as non important factor anymore by some people in SEO community.
You will see more in the pictures.

- I've build 30 backlinks/bookmarks with IMT Bookmarks tool(Private Network) to 3 out of 4 blogs in 10 day time period
and on 1 blog no links where built what so ever.

Details about blogs links and anchors used in this test case:
1) prekintrebalings1.wordpress.com
100% of anchors used is main keyword "prekintrebalings"

2) prekintrebalings2.wordpress.com
30% of anchors used is main keyword "prekintrebalings"
70% of anchors used are generic keywords(click here, visit this site, etc..)

3) prekintrebalings3.wordpress.com
No links built

4) prekintrebalings4.wordpress.com
65% of anchors used is main keyword "prekintrebalings"
25% of anchors used are generic keywords(click here, visit this site, etc..)


- I never accessed these blogs trough the Google SERPs, in order not to influence on visitors behavioral rankings factors like CTR, bounce rate,etc...
Blogs are accessed once to set them up and once(only on two of them) near the end of this testing to change meta description to match other two.

- test case was conducted on google.com/US



The results, "prekintrebalings" SERPs in chronological order:

http://www.imtalk.org/images/stuff/t...ebalings/1.gif

http://www.imtalk.org/images/stuff/t...ebalings/2.gif

http://www.imtalk.org/images/stuff/t...ebalings/3.gif

http://www.imtalk.org/images/stuff/t...ebalings/4.gif

http://www.imtalk.org/images/stuff/t...ebalings/5.gif

http://www.imtalk.org/images/stuff/t...ebalings/6.gif

http://www.imtalk.org/images/stuff/t...ebalings/7.gif

http://www.imtalk.org/images/stuff/t...ebalings/8.gif

UPDATE 28.10.2013
http://www.imtalk.org/images/stuff/t...ebalings/9.gif



Notes:
Because this test case was conducted in controlled environment and wasn't public, so there where no outside influences or manipulations,
once this test case is presented to the public it can be compromised(and will be) influenced an manipulated by others for many reasons and therefore not accurate anymore.
It can be influenced not intentionally and only by for example clicking trough the SERP links and browsing/exiting pages on blogs, etc...



Conclusion:
It is obvious that blog without any links was always in the last place from day one and until the end of this test case.
So we can conclude that links work!
Even after penguin 2.0, who knew, right? ;)

The next logical question is what type anchors work?
As we could see the best performing anchors for this test case where mix of 30% main keyword/s and 70% of generic keywords
(in real life we need to use much more keyword variations as well as brand names, URLs, etc. as anchors).
It is also natural and secure way to do linkbuilding in terms of anchor usage, why, because it mimic natural distribution of anchors used in real life
(with other things as well which are not the scope of this test case and are well covered and explained in my tutorials and posts).


UPADTED Conclusion on 22.01.2014
All still stands from above conclusion expect that % of anchors needs to be changed to this:
1) Main Keywords, Additional Keywords/Keywords Variations = around 30%
2) Brand name(if you don't have brand name put this: "yourwebsite.com" without quotes) = around 20%
3) Exact URL you are promoting = around 25%
4) Generic Keywords = around 25%
This is explained in post #19...


In the end any of us will draw our own conclusion from this.
I always say don't fully trust anyone, especial not big multinational companies as Google, they don't care nor work for you, they are only after money for their shareholders!
I'm not saying you need to trust me, no, instead do your own homework.
You are more than welcome to make your own test cases(it is not hard at all as you can see) to check what works and what doesn't.
And I will be very happy if you do it and publish data here from your tests for everyone to see as well.

I think it is great time for us IMs and SEOers because most people are scared to do SEO right now, so it is very easy to rank sites and earn $ again,
because there are literally no competition anymore.
I'm not saying to do SEO in any way especially not to SPAM by any means, but to do SEO smart and proper way.

I hope you enjoyed this, that it was useful and that you will share this with others.

Take care :)

Wow great case study, you should def go more indepth with this, like you said with the anchors. Nonetheless good post, definitely put my wonders to rest. :)

Very nice.
I've also seen a similar test recorded on video by a SeNuke Guy if I remember well... There was a mix of wordpress and blogspot blogs.
Suggestion in the future: it would be great if (in order not to compromise the results of a test) if there were two sets of identical test cases. I mean in parallel, along with the 4 blogs with "prekintrebalings" keyword, another 4 blogs with the keyword ... "madoxxingrosej". And any action is applied the same on any blog, but on the time of making the first public (showing the results here - that of course as you guessed, will generate traffic on the blogs, keyword searches...), the second would go forward until we can make more accurate conclusions.

What do you think ?

You are correct, that was similar but very different test case.
These type of tests are done for many years now to identify what works and how in google algo(and other SEs) by many serious and top SEOers.

By these methods I'm testing google algos for long time and draw my conclusions, that is why I always say do your own testings, it is the best and only way to know what works.
Now, because of so many testings in the past I've done I sometimes don't even need in depth test cases like this to see what is going on with some ranking factors,
simpler tests can give me the info I need to make my conclusions as well, I guess it comes with experience.
With this test case I just wanted to show how it is done properly and how others can do it too, to see and find out answers on questions of their interest by themselves and very easily.

Yes I agree that we should have more tests like this, but unfortunately they need time and dedication and currently I don't have much time for many more, maybe few depends on my time...
That is why I asked each and every of you to make similar tests and post results here for others, same as I did with this test case.

Hi there, great case study.

I think that no matter what, Google will always rely on links. They cannot do it any other way.

second, "natural" is the key. You see, I was submitting one site (promoting it really) for the past few days (testing this strategy for one client). Well I am promoting content. Depends on the site, I change the title of the post, but if I was to submit it with the same title everywhere it would look like a spam.

Another thing is to spread the promotion and add in variety to it. Articles marketing, social media, video marketing, blogs and so forth.

Mixture will work today and in years to come, because it is natural.

One of the best posts i have read in some time. Great to see real tests being published.

Thank you for this great post and very useful SEO-Test! It shows what I´m thinking anyway, because I think that links will still work in the future. Google want´s to bring up websites to the TOP that provide useful content to users and therefore are recommended by them. And these recommendations just work by linking. Therefore, for Google in future backlinks still will be an important characteristic for a "useful" website. And this is confirmed by your study. The important thing about backlinks is the way HOW to build them. In this regard your study shows a good formula: 30% to 70%! Thank you for this post!

Great job Nemanja...

Thanks for sharing the results and taking the time to prove the reasoning behind the method.

Great testings Nemanja

Definitely we are going to share this with others. Thank you for this post

Great Job...

Thanks for sharing

Great test you conducted. At least we can be assured that links still work. Some of my keywords have been hardly hit by the recent update. But they are now getting back to their place. Your research has given me new ideas and I will share the results after I put them to test.

cleared my all doubt i google made it all clear with your research 30% is perfect ratio but Nemanja what about remaining 70% can we use variant related keywords over there for example my main keyword is web design melbourne so can i use website design melbourne australia as 70% variant keyword or should i only uses here or click now and all .. :)

Good test! Nice to see you eliminating the variables (as much as is possible), again to agree with most other posters links will be important for a long time, it's how Google made their search work better than others and they would be back to square one by taking power away from link in general. Key is to be natural, links are still the most difficult factor to manipulate if google can see past unnatural backlink mixes. There is still plenty to do with links, so long as you spend a bit of time to do it properly

Great test to prove that SEO is still here but the only thing is that it need be done in a good way and not spam. You very well said "do your own homework", this quote is really catchy and those who think will keep thinking about what you want to say.

Links do work and this should be done in a good manner.

Quote:

---

Originally Posted by andrewflower12

cleared my all doubt i google made it all clear with your research 30% is perfect ratio but Nemanja what about remaining 70% can we use variant related keywords over there for example my main keyword is web design melbourne so can i use website design melbourne australia as 70% variant keyword or should i only uses here or click now and all .. :)

---

As I already said in first post this is not definite %, it worked best for this test case, although I'm using it in my link building currently and for some time now,
but each of you can tweak it a bit to suit your needs, nothing in SEO is final.

I'm using it as I explained in wiki tutorial, so will copy paste % here.
Here are some percentage of anchors I think will work these days:
1) Main Keywords, Additional Keywords/Keywords Variations = around 40%
2) Brand name(if you don't have brand name put this: "yourwebsite.com" without quotes) = around 25%
3) Exact URL you are promoting = around 15%
4) Generic Keywords = around 20%

So to answer your question I would use variant keywords only in that 30% maybe you can stretch it to 40% if you have much variations.

This was the same or equal conclusion many people have drived from this Google updates. Google will not be able to go around without LINKS. Because that will be the end of Internet and that will be end of Google itself.

What they did is by changing their algo they made sure you cannot dominate your niche and keywords anymore as that an easy thing to do.

This way 30% anchor text and balance something else would produce different results. Once that change was made your site rankings dropped, because those who build their links based 100% on anchor texts alone - got sent down the pages to 100 or -400 th position in SERPS. Those sites now became targets for their ADWORDS marketing machine..

Google made sure that algo changes will sell more ADWORDS and hence more clicks...We keep getting $100 coupons in the mailbox and you are getting them too. They have been relentless in making sure now your rankings are killed you would pay us to get traffic. I will personally never give in to this tactic though.

I f you read GOOG stocks earnings reports on Yahoo.com etc you will see they were able to increase their clicks nicely and hence get more revenues. Their stock is skyrocketing, while your fortunes in this recession got even worse. :mad:

Thank you Nemanja for a great post.

Hello, first post here and many thanks for the case study! :cool:

UPDATE

Few months after this test case this is the situation.

http://www.imtalk.org/images/stuff/t...ebalings/9.gif

As you can see blog num 2 is lost now, which was the best blog in the end of our test case(30% main keyword).
blog 1 with all same anchors is no1 and blog 4 with about 65% same anchors is no2.
This is very interesting, even more interesting if you consider that this is happening after last penguin.
Same things are happening with my other sites as well.
So based on these new information, anchor text percentage needs to be changed,
to play on safe side we will go with blog no4, because having same anchors is just not good, like it is in no1 blog.
I suspect it works for this keyword because it is not well known/affiliate/competitive search term.
Blog no4 was always in he middle, like a silent performer, as I said based on my other submissions and the fact that it is not competitive search term.

I think this would be the right % of anchors for your links:
1) Main Keywords, Additional Keywords/Keywords Variations = around 40%
2) Brand name(if you don't have brand name put this: "yourwebsite.com" without quotes) = around 25%
3) Exact URL you are promoting = around 15%
4) Generic Keywords = around 20%

I will update original post as well with this new information about anchors.

Hi
its a good case study helping us understand what WORKS!
Fred