Can inbound links really hurt you?

[inertia]

The company that i used to work for out-sourced all their link campaigns from a company on the other side of the world. The links that they would establish for us were all PR3+, they were on pages with plenty of other links and the pages were totally irrelevant. We used to make sure that they built the links at a steady rate which wasn't to quick and if the PR dropped we asked them to replace the link.

Did it work? Yep! Worked every time! Generally speaking, for the level of site we used to work with, 100 links would usually see the site in the top 10 SERPs for the anchor text used.

Now, i know this is bad practice. I knew it wasn't something that the SEO world would agree with and i knew it was a linking practice that was judging link quality just on tool bar page rank. But, the boss told me to start the campaigns and within 3 months they were (usually) successful!?

I saw this happen with 100s of sites and id say the success rate of these link campaigns was 95%. From what people are saying here that would be a great way to attack a competitors site???

Thoughts please...

[kgun]

Originally Posted by crankydave

Precisely the point. They can't. They can't "know" whether or not a link is paid. They "guess" and act upon that "guess", based upon a pattern.

As I said earlier, algorithms, bots, filters, etc. are not intuitive. They can't "know". They don't consider what "makes sense". They don't "think" (not yet anyway). They can't assign "no value" to a paid link, they need webmasters to do that, what makes you, or anyone else for that matter, think they can assign "no value" to any other link "automatically"?

A pattern is a pattern. When that criteria is met, it is acted upon automatically. Doesn't matter why the pattern exists. "Why" cannot be automatically discerned.

Have your ever programmed? Some patterns are identified better by statistical / mathematical algorithms (Bots) than by human observation. But identifying bad IBL's is not an exact science. There will never be a "perfect" world.

Originally Posted by Peter (IMC)

I completely agree with Dave. It is possible to get your competitor's site drop in rankings by doing really bad link building for their website. Everybody agrees that doing link building the wrong way can hurt your rankings, and I think everybody would agree also that Google can not know who's responsible for the backlinks that a site has.

Natural question:
Drop in rankings because a prior positive effect is zeroed out (neutralized) for one reason or another?

What would the implication be if you could destroy a competitor's business by establishing bad IBL's links to that site? Wouldn't that be the start of the end of the internet (some companies with deep pockets taking over the internet). IMO that would also be the start of the end for that Se. Zero (neutral) weight of IBL's (votes) is IMO the only viable option.

[crankydave]

Originally Posted by kgun

Have your ever programmed? Some patterns are identified better by statistical / mathematical algorithms (Bots) than by human observation. But identifying bad IBL's is not an exact science. There will never be a "perfect" world.

No kgun. I am not a programmer, nor would I claim to be. I don't dispute that bots can do a better job identifying patterns. That's not what's at issue here. What they cannot identify is "why" the pattern exists. No, it's certainly not a "perfect world". Mistakes get made.

Originally Posted by kgun

Natural question:
Drop in rankings because a prior positive effect is zeroed out (neutralized) for one reason or another?

First, one would have to "assume" there was some positive influence to begin with. Then, one would have to speculate as to precisely which rankings were influenced by that assumption. All of the rankings? Some of them?

Hypothetical kgun...

A reasonably established site with a reasonably established small link profile. Ranks well for several terms. Receives an influx of a large number of sitewide links from elsewhere. All the rankings fall. Links are removed, rankings return. Repeat.

Loss of "prior positive effect"?

Dave

[crankydave]

Originally Posted by kgun

Zero (neutral) weight of IBL's (votes) is IMO the only viable option.

Then why can't they do this with paid links they can identify? Why do they need webmasters to "assign zero (neutral) weight" to them?

Dave

[kgun]

1. Do you see an argument for paid links not passing pagerank?
2. Do you see an argument for links from social networks getting reduced pagerank passing power?
3. Do you think a computer program (bot) can identify patterns related to 1 and 2 and neutralize the effect?

Originally Posted by crankydave

Then why can't they do this with paid links they can identify? Why do they need webmasters to "assign zero (neutral) weight" to them?

Do you build that statment on old statements by Matt Cutts or Google resource pages?

Marketing is a type of drama.

My experience goes back to the date of the punched cards (FORTRAN programs). Microsoft is the computer company that IMO has understood offline marketing and branding best. They love to get focus around their new technologies. Is there any difference with the best online marketing and branding company, Google?

Nofollow on OBL's (IBL's to another site) give an exact solution. Pattern reckognition algorithms give an approximate solution. If you prefer an exact solution, put nofollow on links that shall not pass pagerank.

That will make the world more perfect.

[crankydave]

Originally Posted by kgun

Do you build that statment on old statements by Matt Cutts or Google resource pages?

It's not a statement, it's a question.

Why does Google "require" webmasters to "reduce the value" of any links if they can do it themselves?

Dave

[kgun]

Originally Posted by crankydave

It's not a statement, it's a question.

Why does Google "require" webmasters to "reduce the value" of any links if they can do it themselves?

Dave

Agree

[galide]

Sorry to interrupt the conversation between you two guys, but still didn't see anywhere the proof that inbound links can get your site penalised?

.. and in the absence of proof, I personally tend to base my judgement on what Google has to say about it officially, which is that links cannot have a "negative impact" on ranking.

How to report paid links
"We’ve always tried very hard to prevent site A from hurting site B. That’s why these reports aren’t being fed directly into algorithms, and are being used as the starting point rather than being used directly. You might also want to review the policy mentioned in my 2005 post (individual links can be discounted and sellers can lose their ability to pass on PageRank/anchortext/etc., which doesn’t allow site A to hurt site B)."

And I will stand by it until someone can prove otherwise, or at least provide some clues about how to set up a benchmark to prove that inbound links can damage a site.

Manu
<please add your link to your signature CD>

[activeco]

Originally Posted by inertia

I saw this happen with 100s of sites and id say the success rate of these link campaigns was 95%.

Now that is what I would call a valid statement for this forum, without a lot of philosophy.

Originally Posted by galide

.. and in the absence of proof, I personally tend to base my judgement on what Google has to say about it officially, which is that links cannot have a "negative impact" on ranking.

How to report paid links
"We’ve always tried very hard to prevent site A from hurting site B. That’s why these reports aren’t being fed directly into algorithms, and are being used as the starting point rather than being used directly. You might also want to review the policy mentioned in my 2005 post (individual links can be discounted and sellers can lose their ability to pass on PageRank/anchortext/etc., which doesn’t allow site A to hurt site B)."

Good one, too.

Now, I am ready to sacrifice one domain for testing purposes around this question. It has PR2, bringing some $100 a month, which translates into pretty good extra spending holiday money.
If anyone wants to set up testing conditions he can send a PM.
Once the conditions set, anyone can do anything he wants to the domain.
If enough interest, I will open a separate thread.

[crankydave]

Originally Posted by galide

"We’ve always tried very hard to prevent site A from hurting site B."

I don't think anyone doubts that. Doesn't say "can't" though.

Originally Posted by galide

That’s why these reports aren’t being fed directly into algorithms, and are being used as the starting point rather than being used directly."

Pretty good indication that if they were fed directly into the algo they could "hurt site B" don't you think?

Now, what about the things that are fed directly into the algo?

Dave

[crankydave]

Just a couple of other things to add to the discussion...

Two Search Interviews

Andy, yes: we can manually remove or otherwise penalize sites on incidents that we consider to be webspam.

Official Google Webmaster Central Blog: An update on spam reporting

So, the next time you can't help thinking that the ranking of a search result was not earned by virtue of its content and legitimate SEO, then it is the perfect moment for a spam report.

A complete exclusion from Google´s SERPs is only one possible consequence of a spam report. Google might also choose to give a site a "yellow card" so that the site can not be found in the index for a short time. However, if a webmaster ignores this signal, then a "red card" with a longer-lasting effect might follow.

Can a particular IBL pattern be considered "webspam"?

Can it be detected and penalized "automatically" as well as manually?

Dave

[inertia]

Originally Posted by activeco

Now, I am ready to sacrifice one domain for testing purposes around this question. It has PR2, bringing some $100 a month, which translates into pretty good extra spending holiday money.
If anyone wants to set up testing conditions he can send a PM.
Once the conditions set, anyone can do anything he wants to the domain.
If enough interest, I will open a separate thread.

This is an interesting idea. But the only way we could really conduct this experiment is by spending a lot of time and money building links for your site - which i cant afford. By my reckoning these will only have a neutral or positive effect on the sites performance!

From what i have seen with my previous employer. A few hundred links on dodgy link selling sites with high PR will continue to improve your rankings regardless of the perceived quality.

[galide]

You raise good points Dave, and I think that all the questions you are asking are valid ones. One can speculate a lot on what is happening on a daily basis in the SERP, on the official posts google is releasing or on the patents they are filing. And speculations are healthy, when they lead to actions though:

1) trying to formulate a theory about what impact onsite and offsite triggers have in the Google ranking algorithm.
2) set up a benchmark in order to demonstrate the theory.
3) provide solid data proving that the theory is sound, and that particular conditions lead to consistent and predictable results.

Once step1-3 are complete, one can then put the theory forward with confidence and advise other people to follow it. But without a proven theory, I believe that it is always safer to play by Google's rules, as described on their (un)official blogs.

One thing is for sure, people such as Activeco can benefit greatly from this type of speculations - my blog is also open for testing: http://galide.jazar.co.uk. Please throw as many links as you want at it and see if we can to take it down

Manu

[deepsand]

Originally Posted by subhzash

Inbound links are really danger when those are from irrelevant sites or irrelevant pages.

IBLs from an "irrelevant" site are not only never of positive value, but are "dangerous?"

Is not the purpose of the PR algorithm to incorporate the preferences of web participants as to what is of interest and value?

[deepsand]

Originally Posted by activeco

Tell it to all those sites went to Hell.

Might it not be that said sites went to hell because, absent meaningless IBLs, they'd nothing of import to the SEs in question going for them?