Need Help Poisoning Rankings, Negative SEO Only Helped

[goodroi]

With all the nightmarish claims about negative SEO, I spent a few dollars and created some websites for an imaginary word and then tried to poison one of my websites.

I headed over to Fiverrr and bought the highest rated negative SEO project and turned it on myself. It took a few days for the poison links to get discovered. Once Google found those poison links, my "poisoned" website jumped up in the serps. I checked with MajesticSEO and they showed a beautiful, highly artificial spike of the worst quality of backlinks. Kudos to Fiverr for delivering on what they promised and kudos for MajesticSEO for uncovering a huge amount of horrible links now pointing to my website.

This was back in June. I have been monitoring it through June, July, August and now September. The ranking boost eventually faded but my "poisoned" site never dropped in the serps. It did lose the ranking boost and is just back to where it started.

I failed in posioning my website. Please help me :)

Before I reveal the specific details and screen captures, what could I have done to better poison my website?



A great way to learn is by conducting hands-on experiments instead of blindly following anonymous "experts", don't you think? :)

[brotherhood of LAN]

>for an imaginary word

So were there other sites ranking for it?

I would try for a low to medium competitive term, e.g. looking at the Adwords KW tool and multiplying CPC by traffic and having something in high-4 low-5 figures.

[Clay_More]

One conclusion that could be drawn is that artificial links have a lifespan, if you are looking for a long-term benefit, you'd have to commit to long term link development.

Many SEO services, especially from low cost providers seem to change focus with changing published concepts. One month it's we can provide many beneficial low cost links, next month it's we can provide negative SEO with many low cost links.

[CaptainSalad2]

goodroi negative SEO doesn’t work until Penguin is refreshed, let us know how the rankings go for the imaginary word (that would rank anyway as the competition would also be imaginary) after the next penguin refresh!

[goodroi]

I hear you about penguin refreshes but some people claim that negative SEO can happen immediately so I figured let's test and find out what really happens and when. I was hoping for a penguin refresh which is why I hadn't announced this test that I started in June. I've become bored waiting for Google to hammer my site (doubt many people ever say that).

With penguin refreshes being so infrequent, it might open the door for a churn & burn spammer. Create a new website, pump it full of the cheapest, easiest, lowest quality links and you might rank for many months before Penguin is refreshed. When Penguin finally catches up, you just walk away and launch a new site and repeat.

[Kelowna]

tried to poison one of my websites.

What does this mean? Lots of random anchors from many sites, even low quality sites will no do it. Did you us the main anchor in the link text more than 10% of the time?

A few more details on what you did to try and poison the site is needed.

Just to note, I use these fiverr blasts quite often, but when done right they help quite a bit even for the so called tough keywords in the finance area.

[goodroi]

I used Fiverr to generate about 100% identical anchor text from over 100,000 low quality spammy sources. I would not want a single one of these links for any of my real websites.

[GreyBeard123]

Type of links?

[goodroi]

Similar to blog spam. Basically links from pages that had a submission form with a url field that an automatic bot can complete and no human oversight to stop it. Mix of dofollow and nofollow.

[creative craig]

Of course this was just for research and no money was made from affiliate or Adsense ;)

[bumpski]

Before I reveal the specific details and screen captures, what could I have done to better poison my website?
A great way to learn is by conducting hands-on experiments instead of blindly following anonymous "experts", don't you think? :)

Are the details coming?
Was this the keyword? [google.co.uk]

[Clay_More]

bumpski,

While you have really mad skills being able to determine a test phrase that's been used before, but is currently being used in another experiment, why potentially alter the experiment by posting the word/phrase?

As soon as you do that, there is an additional variable that invalidates further data.

Nice.

Ever wonder why most people don't post their experiments and results?

[philgames]

As it takes penguin to recover it also takes penguin to destroy this is a misconception with negative seo that it doesn't work after people test it out and months later see no impact.. but last penguin has been more than a few months ago... you will see next penguin update... you will see.

I also advise making blogs that contain 3 posts which look like a pbn blog (ya know standard wordpress theme and either spun content or dupe wikipedia content. with keyword rich anchor text to your site as icing on the cake.. This is also means that you are not getting links from already spammed domains that arent ignored by google (maybe) but fresh and then you can freshly make signals that maybe deemed as a way of manipulating the search results from the webmaster of the site crossing the line between a site wich has tons of auto gernated links which could have been a result of anything from user genrated (not having spam guards).. a hacked site or a negative seo attack.

[fathom]

Frankly there are a few things wrong here:

Was this really testing Negative SEO?

... or simply testing Blackhat SEO?

No one claimed Blackhat doesn't work, it is simply shorter lived these days.

I wouldn't expect Google to do anything with an imaginary term since that is like providing Google "Viagra" before "Viagra" was a term in Google.

In order to see/test any claim for Negative SEO you have to start with a legit rankable project, with real traffic, and with real natural links. Otherwise the test is invalid.

Note to escape from PENGUIN you need a PENGUIN Algorithm Re-RUN (I actually don't believe that but that is Google's claim). But all you need to be caught by PENGUIN is a data refresh and that generally occurs every couple of weeks.

For the imaginary term that can be likened to a brand name and I am sorry to say brands for even PENGUINized domains still rank, which is a good reason why this test failed to produce any declining results but not because it tested Negative SEO but what may or may not be Blackhat SEO.

[edited by: fathom at 5:21 am (utc) on Sep 25, 2014]

[fathom]

I used Fiverr to generate about 100% identical anchor text from over 100,000 low quality spammy sources.

But these above...

Need to be to these below.

I would not want a single one of these links for any of my real websites.

[McMohan]

Penalty can happen in 2 ways -

1. Anchor text abuse from low quality sites, most often free links. Penguin is launched to detect and kill such abuse and it has to refresh for you to see negative impact.

2. Getting links from network of sites that are known to be sources of paid blackhat links such as SAPE network, anchor text may or may not be abused. Penguin on its own can't uncover such links. So, Google relies on manual penalties.

This, 2nd approach is the way to go if you want to see the immediate "benefit" of negative SEO. Good luck!

[martinibuster]

Great experiment! :)

The imaginary term might be a flaw. GIGO.

[goodroi]

I agree that a real term with traffic would be interesting to experiment but to be honest, that real term would also introduce many other variables that would make it much harder to isolate what is happening.

I am a little curious why some people might think that an imaginary term would not be useful. Google applies their automated algorithm to all terms. Google does have many different subparts of the algo and different terms are treated differently by different subparts but I think there is some usefulness in using an isolated imaginary term.

I strongly encourage everyone to do their own experimenting. You can learn much more by doing things yourself instead of blindly following the advice of a self proclaimed expert who in reality has no clue.

[simonlondon]

I think fathom has nailed it on this subject.

[CaptainSalad2]

You can learn much more by doing things yourself instead of blindly following the advice of a self proclaimed expert who in reality has no clue

Frankly I’m not impressed with a single so called "SEO expert" who cant also provide a state of the art website and has a deep understanding of everything from hand coding/server optimisation and a heap of other standards every website should aspire to. In short provide a complete "product" from the ground up!

You cant polish a turd, however many "seo's" sure like to roll them around in glitter for a large sum of cash! :)

[martinibuster]

...might think that an imaginary term would not be useful.

I believe the answer is GIGO, as I stated above.

Google returns answers to queries it can categorize, identify as a concept, etcetera. Google understands concepts based on the data that is input. The data that is input are the web pages it crawls. If you introduce a query for which there is no data then Garbage In/Garbage Out (GIGO).

Otherwise all that is happening is Google is giving you a semantic needle from the Internet haystack. A word fetch. But that's not really the algo at work in the usual sense because a real search is far more than matching queries to text on a web page. That's just a word fetch.

[bumpski]

bumpski,
While you have really mad skills being able to determine a test phrase that's been used before, but is currently being used in another experiment, why potentially alter the experiment by posting the word/phrase?

As soon as you do that, there is an additional variable that invalidates further data.

Nice.

Ever wonder why most people don't post their experiments and results?

Clay_More
I'm not sure how you could know, for certain, the keyword in question? I don't think I even had the correct "keyword in question" in the link I included in my post, a link to a dynamic search results page from Google which surely cannot affect rankings. So please see below the "outtings" of this keyword (since 2012?) by the OP and many others! And finally the quotes from the site in question are certainly intriguing and immortalized in the Archive. There appear to be so many "variables" in this experiment, one would think it must have run its course.
Regardless, the OP requested help, got quite a bit of help from WW visitors, and then the thread seemed to die without the promised follow up.

[webmasterworld.com...]
10/15/2012
"Step 1 - Go to Blogger and register a site for a made up word like "the keyword in question" (don't use that example since Google just crawled this page and now knows this word) and publish it with one page of content."

[webmasterworld.com...]
07/15/2013
"I have one question for you: What have you been testing?"

[webmasterworld.com...]
09/10/2013
"I generally don't believe anything Google says or publishes unless those statements can be verified. FUD is dispelled by testing, I always have my own running and pay attention to others like 'the keyword in question'".

[webmasterworld.com...]
03/14/2014
Many mentions of the "keyword in question" from the site I won't link to:

Quotes from the site in question containing the "keyword in question":
"Personally I think it is silly and dangerous to rely on others to do the research and them hope that they will accurately share those internet marketing findings with you."
"I wonder even more why so many WebmasterWorld members don't get off their lazy a-- and do some quality research on their own."

So one would wonder about the veracity of the report anyway, wouldn't one?

A great way to learn is by conducting hands-on experiments instead of blindly following anonymous "experts", don't you think? :)

Cheers!

[goodroi]

Might be a good idea if we wait for a penguin update before jumping ahead for this specific test.

We can argue amongst each other and waste time complaining about different things but it is probably more productive and profitable if we all just do more testing and sharing of real data :)

[Clay_More]

Mid September I took two sites and set up Fiverr gigs for them. I used the same ten commercial medium high competition keywords for both.

First site received about 200k links which appear to be mostly trackbacks. That site currently shows 6,000 backlinks with a Majestic trust flow of 14 and citation flow of 8. I haven't seen the site in the SERPs for those keywords, but I'll look a little deeper mid month.

Second site got two gigs, one being just under 100k blog comments, the second being social mentions. That site shows 25,000 links still existing with a trust flow of 15 and citation flow of 31. I haven't seen this site in the SERPs either, but citation flow is still increasing and I'll probably work on the trust flow mid month.

I currently have a gig placed for a third site very similar to the other two, this one is 250k links using a two word mid level competition phrase and a made up word.

[Planet13]

@ Clay_More

Just so I understand, are you getting these links built for a NEGATIVE SEO campaign? (i.e., testing the ability of fiverr-built links to tank a site?)

Or are you trying to rank sites instead?

[Clay_More]

I'd never pointed that many links at any site before, so really it was more a test to see what would happen. The ten keyword thing was to keep anchor text diversity.

I was somewhat surprised by the results on the second site, so I'll try to find some higher quality links to see if I can boost the trust flow.

Third site is to see what differences will show up between the made up word and the more competitive phrase. They should each get very similar amounts of exposure.

I'd have to do an awful lot more testing before I'd even think to try ranking a site using just Fiverr.