homepage Welcome to WebmasterWorld Guest from 54.167.75.155
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member
Home / Forums Index / Google / Google SEO News and Discussion
Forum Library, Charter, Moderators: Robert Charlton & aakk9999 & brotherhood of lan & goodroi

Google SEO News and Discussion Forum

This 77 message thread spans 3 pages: < < 77 ( 1 [2] 3 > >     
Spam links to my site got me penalized
Steelbank




msg:3964443
 6:41 am on Aug 3, 2009 (gmt 0)

I realized today, via google webmaster tools, that one of my sites had over 2000 links to one of our landing pages. The links were run-of-site links on spammy looking websites i.e., viagra, pharma. I don't know how the links got there or why they are there - other than to think a competitor may have put them there.?

I started investigating the issues of our site because of a sudden drop from page 1 on google to page 8. The site no longer pulls up for any of our searches whereas we were between the top 5 on page one.

Any thoughts, advise on what we should do?

 

drall




msg:3983917
 10:56 am on Sep 4, 2009 (gmt 0)

Like Tedster said, the type of attack launched on us is not for the average negative seo campaign. Whoever did this was considerably versed and had a very deep understanding of Google and how to "work the system". They spent a substantial amount of time creating compelling rich content to frame us.

MrSavage




msg:3984038
 2:54 pm on Sep 4, 2009 (gmt 0)

In other words a new form of SEO company as it were. Black market or underground SEO companies who specialize in this type of thing. If people learn a playbook in this regard, there will be people signing up for this service. Obviously if there is a market for a service or product, somebody will provide it. I would just think that it's a dangerous path. Nothing will shock me when it comes to what people will do to get their top rankings. Nothing. To me, it's becoming such a thin line of the internet becoming complete anarchy and chaos. I don't think things are as stable as one might believe.

WarrenBuffett




msg:3984048
 3:07 pm on Sep 4, 2009 (gmt 0)

I've witnessed a site in my niche get the -50 penalty. A quick observation into it's backlinks, and all it took was a few hundred blog posts with no variety in anchor texts, and they were all nofollow links too! (The anchor text was '.') Trying to hide the link in a period.

Reno




msg:3984055
 3:22 pm on Sep 4, 2009 (gmt 0)

Earlier in this thread we offered the hope that perhaps Google would provide a way to "neutralize" unwanted links in the GWT. In addition, it would be helpful to have some sort of warning system that would be triggered when G begins to suspect that the linking pattern is suspicious. Something as simple as a "light" with 4 colors -- green (all appears well); yellow (you might want to take a look); orange (there's trouble in paradise); and red (10 yard penalty).

Yes, it is time consuming to have to investigate a bunch of backlinks, but the time would be minuscule compared to what it would take to rebuild a website, or worse yet, start over from scratch. And with a "warning system" in place, these link investigations could be ongoing, rather than all-at-once, so the siteowner could stay on top of it and would thus have some recourse against this sort of attack

.....................

Shaddows




msg:3984062
 3:37 pm on Sep 4, 2009 (gmt 0)

Reno, I'm afraid not.

It would make an excellent tool to start investigating the outer limits of "acceptable practice." You could scale-up your black hat practices incrementally until the warning bells ring.

On the other side, you could set up a throw-away account, and find out just what techniques send you red straight away. All the better to take down that pesky competition in an efficient way, without wasting time having to frame them.

These reasons are precisely why G, Y! and B don't discuss or elaborate on the penalty processes. And particularly don't give specific feedback on individual situations

Reno




msg:3984086
 4:14 pm on Sep 4, 2009 (gmt 0)

I see your point. I'm not aware of black hat practices (perhaps to my peril!), so it had not occured to me that an "alert" tool to help webmasters could be used against them. No good deed goes unpunished...

.........................

CainIV




msg:3985315
 3:38 am on Sep 7, 2009 (gmt 0)

Brings up, once again, the idea of Google creating a 'Do not count this link' framework inside Google webmaster tools which allows website owners to arbitrarily select links that are dodgy that the webmaster has not secured themselves, so that Google can ignore, as opposed to penalize.

maximillianos




msg:3985319
 4:17 am on Sep 7, 2009 (gmt 0)

Scary stuff indeed. It makes sense. I've seen sites kicked out of SERPS for doing such tactics to try and boost their own rank. The next logical step would be for these disgrunted sites to take their knowledge and reproduce it only this time targeting competitor sites.

Yikes.

jd01




msg:3985326
 5:01 am on Sep 7, 2009 (gmt 0)

I keep thinking it could make sense if only one link per domain, per page 'counts' and the overall value of the link 'counted' is weighted by 'relevance' between the linked pages...

2000 ROS links to a single page? Cool. I don't care who paid for them or how they got there! Only one counts and the value is based on overall relativity of the linked topics...

IOW Site A links to Site B's home page 10 times throughout the site. Only one link from Site A to Site B counts. The weight of the link 'counted' is based on the relativity of topics between the linked pages... If 100% (all ten) of the links pointing from Site A's pages to Site B's home page are from 'relatively on topic' pages, then the 'most important link' counts 'full weight'. If only 20% of the links share 'relativity of topics' then the 'most important link' counts at 20% weight.

Multiple links from one site to one page count as one link total where the value of the counted link is based on the overall relevance of the topics of the pages containing the links as compared to the topic present on the page receiving the link.

2000 ROS links to 2000 different pages? Cool! I don't care how they got there or who paid for them: 'via***' to 'buy cars now' = no relativity... Buy 'em up, they don't count!

IOW Multiple links from multiple pages on one site to multiple pages on another count at the level of perceived relatively between the linked pages.

Trucks to Trucks = Very relative.
Trucks to Cars = Fairly relative.
Trucks to SUVs = Fairly relative.
Trucks to Insurance = Somewhat relative.
Trucks to Tractors = Somewhat relative.
Trucks to Concrete = Low relativity.
Trucks to Pharma = No relativity. (Unless the receiving page contains info about a mid-life crisis and shrink recommendations! LOL)

* The converse of the preceding is also true.

** If you're a 'Googler' I think you could use 'Math 3.0' to 'score' the value and relativity, but I'm 3/4 nuts, so don't mind me too much. I just like posting craziness!

keyplyr




msg:3985337
 5:37 am on Sep 7, 2009 (gmt 0)

ukonetraining - Anyone know of good tools (preferably free) to track these links if indeed they do exist?

[webmasterworld.com...]

docbird




msg:3985349
 5:56 am on Sep 7, 2009 (gmt 0)

After seeing the mj12bot thread, I'd tried the links checker. Noticed several links from spammy/dodgy sites. I'd thought maybe related to referer spam.

Just looked at list again, only few days later; and this time tried visiting some of dodgy pages that had reportedly been linking to my site. Found several that had no links to my site, tho maybe had related keywords.
So, wonder if there's state of flux, w pages sometimes having links, and links later vanishing (speculating: as links to other pages created, w different keywords).

tedster




msg:3985351
 6:04 am on Sep 7, 2009 (gmt 0)

Yes, that's part of some spam strategies - links that shuffle around dynamically, as well as links that are cloaked from ordinary browsers, and a bajillion other tricks. A back link that is truly gone is not likely to cause you ranking problems.

ogletree




msg:3985380
 7:49 am on Sep 7, 2009 (gmt 0)

It has been known for a while that you can get somebody banned. As soon as Google started doing it a lot of people went out and got people banned. It is not easy to do and can backfire. Google thinks because there is a possibility it might help a site that people wont do it. Google thinks that nobody would get a lot of links like that to somebody else’s website. Well they are wrong. We always thought Google would never allow this but they did. They don't seem to care that there are innocent people that are getting hurt.

jd01 what you describe will get you SEO traffic eventually. That is what everybody should be doing. SEO is about making content that is useful. A good site will attract links. If you site is popular without Google it will become an authority and will get the traffic from Google because of that. If you spend your time trying to get people to come to your good content outside of Google you will get links as a side affect.

tedster




msg:3985391
 8:16 am on Sep 7, 2009 (gmt 0)

Google now certainly has many examples of false positives. They know that others can hurt your site's rankings -- if they know a whole lot about Google technically, and if they stay ahead of Google's continued defensive efforts. That's why the public statement was changed to say "almost nothing".

I think Google does care about it, but the issue is not on the top of their list. If you think that's what has happened to your rankings, then document what you've found and use the reconsideration request. But if you're in a position where you've been pushing the boundaries yourself, then you have some clean-up to do first.

true_INFP




msg:3985480
 12:00 pm on Sep 7, 2009 (gmt 0)

Sorry I don't share the excitement.

So I will have to go over each and every one of the 15,000 pages linking to my site and approve each in GWT?

And even if I do that I am not protected from anything. An attacker can build few thousands of pages with links that look legit and after few months (after "approval") change the content and the anchors while keeping it on the same approved URLs. And to prevent that kind of attack you have to monitor your links regulary, and to do that for a medium/large website you have to hire a team of link approvers ...

The only solution I see is that under no circumstances google will let incoming links hurt the website being linked to. This is the only robust solution.


I couldn't agree more.

internetheaven




msg:3985481
 12:03 pm on Sep 7, 2009 (gmt 0)

This is scary, if it's really what happened, because if this type of subversion works, any site could be vulnerable.

This isn't the first WW thread outing this theory. I thought it was common knowledge, even to newbies? The only decision webmasters have to make these days is to either be a victim, or an attacker.

I've been a victim 3 times now, I know which competitor did it on one occasion and now I really feel like being an attacker.

somebody spent alot of time and money

Then they were idiots. The black hat packages I've seen cost under $100.

Yes, I could sink most sites around me in the rankings for $100 - seriously. They don't have a stable enough base to fend off a spam link attack. Google will drop them and unless they have a WMT account they will never know why.

There must be threads on this dating back to early last year -- and the arguments are still the same. Google should only "devalue" the links, not penalise sites as all it does is give weapons to black hats.

Shaddows




msg:3985513
 1:09 pm on Sep 7, 2009 (gmt 0)

Google mitigate the problem by knowing "attack signitures" just as they know "spam signitures". For new vectors, you have reinclusion requests.

If spamming was not penalised, there would be no reason not to do it. You need to dissuade black hats from trying. Just discounting the effort is not sufficient.

internetheaven




msg:3985550
 2:42 pm on Sep 7, 2009 (gmt 0)

If spamming was not penalised, there would be no reason not to do it. You need to dissuade black hats from trying. Just discounting the effort is not sufficient.

I don't understand. Making their efforts worthless is the BEST way to do it.

They are not discouraged by penalties, if a black hat's website get's penalised they just start up a new site. This mainly is a way for black hats to damage competitors. It has has no "deterrent" factor.

reinclusion requests

Verifying just how pointless the whole thing is! Black hat get's penalised, contacts Google saying "someone else did it" - black hat gets re-included.

The only people suffering are webmasters attacked by black-hats who don't have a WMT account, don't realise they've been attacked and don't know how to undo it.

Google have provided black hats a way to damage competitors rankings and a way to undo any bad actions they may have taken on their own sites.

dazzlindonna




msg:3985551
 2:43 pm on Sep 7, 2009 (gmt 0)

If spamming was not penalised, there would be no reason not to do it. You need to dissuade black hats from trying. Just discounting the effort is not sufficient.

I don't understand that reasoning. If Google was penalizing the blackhatters who were doing the attacking, then sure, that would make sense. But Google is penalizing the victim, not the attacker. The penalty *is* the whole reason the attacker launches the attack - it's certainly not an incentive for him to stop.

And yes, the possibility certainly exists that once the attacker is discovered, then the attacker could be penalized as well. But honestly, I seriously doubt that happens. Why? Well, just take a look at the whole paid links penalty.

There are two parties involved in the buying and selling of links - the buyer and the seller (obviously). The buyer is the one who benefits from the link (in terms of Google's interest in the matter). It is the buyer who is attempting to game Google's algo, after all. So it would make the most sense to penalize the buyer. But does Google penalize the buyer? No, absolutely not. The seller is penalized instead. (The penalty is merely a TBPR drop, but they could increase that to be a rankings drop if they wanted to).

In both cases, the party actively seeking to game the algo (blackhats bowling competitors with spammy links, and buyers purchasing paid links) skate by with no penalty. In both cases, the other party is penalized instead. Now one could argue that sellers of paid links aren't as "innocent" as webmasters who've been bowled, but I'd argue that they are (with some exceptions). How many bloggers who knew nothing of SEO sold text links to monetize their blogs, and then got smacked with a penalty? Lots did.

So, I'm in the corner of believing that Google should be discounting spammy links and paid links rather than penalizing anyone. I think that would be far more effective in the long run, and much less damaging to innocent parties.

Buyers would stop buying links if the ROI was zero. Blackhats would stop bowling with spammy links, if the ROI was zero. No innocents would be penalized in the process.

Shaddows




msg:3985567
 3:09 pm on Sep 7, 2009 (gmt 0)

Ok, the reasoning is thus.

The penalty happens because the technique employed USED to give a blackhat SEO boost. G got wise, and penalised.

G believes anyone employing the technique in isolation is trying to unfairly improve their rank. From that basic premise...
Apologies for the cut-and-paste job. This is from previous page of this thread, but slightly edited for clarity:

The philosophy around penalties is exactly the same as criminal justice- retribution and deterrence.

IF you "sin" by trying black hat techiques (to IMPROVE ranking), there must be consequences. If the technique works, but then does nothing when discovered, everyone would go black-hat. There would be NO downside, just a series of short-lived upsides- leading to an arms-race as everyone struggled to stay ahead of the devaluation curve.

Throw-away domains would be launched in the thousands, soley for the purpose of propping up black-hat websites. It would be chaos.

To stop this proliferation, there has to be some way of making transgressions a risky business. So, penalties are applied. These punish the transgressors, but most importantly discourage most sites from even attempting it.

Of course, the criminal justice analogy can only go so far. The "evidence" is pattern-based, the "trial" is algorithmic, and the punishment arbitrary. The appeals process is the very opposite of transparent, and the judiciary are inscrutible, unanswerable and unimpeachable.

But for all that, I believe penalisation is a necessary ingredient in the algo, though the entire process needs improvement, IMHO.

Now, what happens today is that black hats frame you as a black hat. This is a secondary problem, and one that needs addressing. However, it is not even close to being on a scale as the first order result vis black hat SEO boosting. Spammers have the process automated. If google stopped penalising, the results would be full of spam. Lots and lots of spam.

Here's how it would work. You would release, say, 5 "main" sites. Then 1000 support sites. The 1000 would be devalued. You would launch 1000 more. Then another 1000. Your 5 "main" sites would keep their top status, because they were never penalised.

internetheaven




msg:3985572
 3:32 pm on Sep 7, 2009 (gmt 0)

If google stopped penalising, the results would be full of spam. Lots and lots of spam.

No-one wants that, obviously. But simply ignoring will work the same.

Here's how it would work. You would release, say, 5 "main" sites. Then 1000 support sites. The 1000 would be devalued. You would launch 1000 more. Then another 1000. Your 5 "main" sites would keep their top status, because they were never penalised.

That makes no sense. I really could not grasp your point or why the 5 are ranking based on 1000+ sites that ... nope, still am not getting it at all ... :(

If those 5 are not doing anything spammy then why should they not rank? If those 5 are being linked to by the 1000+ devalued sites then they will give no boost to them so they won't rank.

Shaddows




msg:3985614
 4:45 pm on Sep 7, 2009 (gmt 0)

Right, 5 sites are well-optimised. They're just normal sites, but they have a secret weapon. They have 1000 sites (an arbitrary number) containing content. The kind of content that is currently taking down the sites on this thread, just less ramped up.

Here's the thing. It works. But only for a while. Say 1 week to 3 months. Then the benefit disappears*

So, in the absence of a penalty, you launch another 1000 sites, rework the content, rinse and repeat. Once you figure how long it take to get penalised, you can actually have the next batch gaining ranking points ready to weaponise at just the right time.

All the time, you are doing "normal" SEO in the 5 sites, and simply using the rolling 1000 as a temporary-cum-permanant boost.

The point is, IT WORKS FOR A PERIOD, THEN GOES. Unless it hurts to "reload", you just keep reloading. You don't care what happens to the weaponised sites, just the site they are propping up.

This is an outline of trechnique, without specifics. I'm not going any further with the practicalities on a public site.

*In the current algo, at this point a penalty is applied, thus punishing the propped up sites, and removing any benefit from further spamming.

internetheaven




msg:3985631
 5:37 pm on Sep 7, 2009 (gmt 0)

Right, 5 sites are well-optimised. They're just normal ...

etc. etc. etc.

Still makes no sense. Whether devalued or penalised ... the effect is the same in the example you give.

mkassets




msg:3985642
 6:28 pm on Sep 7, 2009 (gmt 0)

I think Google does care about it, but the issue is not on the top of their list.

That's because the issue is not widespread enough (yet). I know you are guys trying not to reveal too much details about the cases you have seen, but seriously, the best thing that can happen is that this kind of exploits would become a common knowledge. When many authority sites will start to fall one after another Google will have to deal with it.

tedster




msg:3985660
 6:54 pm on Sep 7, 2009 (gmt 0)

Google is already dealing with it. From what I see, the types of attacks that worked earlier this summer are no longer working. In the Caffeine threads, there are more observations about undeserved penalties falling away.

I can't emphasize enough, this is not an epidemic. What we don't want to do is start some kind of panic. In most penalty cases, a penalized webmaster has some other issue, but just they put on blinders.

true_INFP




msg:3985664
 6:57 pm on Sep 7, 2009 (gmt 0)

I agree with internetheaven. The easy solution for Google is to simply disregard spam-like links (instead of using penalties).

That way the bad guys will stop doing it (1) to increase their rankings, because it will be a waste of time. They will also stop doing it (2) to hurt their competitors, because it will be a waste of time too.

Simple and effective.

true_INFP




msg:3985665
 7:00 pm on Sep 7, 2009 (gmt 0)

I can't emphasize enough, this is not an epidemic.

Hopefully. Anyhow, Google should act before it becomes one. And by acting I mean implementing a real solution (not some kind of GWT whack-a-mole games).

Shaddows




msg:3985692
 8:12 pm on Sep 7, 2009 (gmt 0)

Take an oversimplifiction based on some fictional backlink score

Non-penalisation scenario

DAY 1)
I have one site. It has 1200 sites linking to it. I control 1000 of those sites. I get credit for a full 1200 though.

DAY 2)
Same situation. I havent been caught yet.

DAY 30)
Still havent been caught

Day 60)
Ah, lost 1000 ranking points today. Bummer. Still, I have another 1000 sites ready to go.

Day 61) EXACTLY WHERE I WAS AT DAY 1. YEY.

Day 100) Still getting 1200 backlink credits. Lucky I dont get penalised

Day 300) Still where I was. Burnt through 5 sets of sites now, but still riding high on my money site. Due to launch another set of sites soon, but I'm ready.

Penalisation Scenario
DAY 1)
I have one site. It has 1200 sites linking to it. I control 1000 of those sites. I get credit for a full 1200 though.

DAY 2)
Same situation. I havent been caught yet.

DAY 30)
Still havent been caught

Day 60)
Ah, lost 1000 ranking points today. Bummer. Still, I have another 1000 sites ready to go.

DAY 61) ARGGGHHHHHH. Its not working

DAY 90) Damn, I've throw thousands of sites at this now, and nothing works. I'm stuck way down SERPs. If only I had kept within guidelines from the start, I'd be much better off. I wont try this again on my next site, leason learnt.

Whatever you wish, simple devaluation is not effective to stop manipulative practices. You need to make a lasting impression on the offender, not just take something away that is easily replaced.

This is a real solution to a real and automatable problem. It gives rise to a smaller problem that is rectifiable. Its not perfect, but it is substantially better than allowing the endless propagation of spammy sites that can never be effectively removed from SERPs.

dstiles




msg:3985702
 8:20 pm on Sep 7, 2009 (gmt 0)

"implementing a real solution"

A real solution like dropping back-link usage altogether. The whole system is based on a premise that was shaky in the beginning: that people link to their favourite sites. They don't. We have thousands of visitors a week who don't even have a web site to place a link on. How did we get those visitors? Certainly wasn't from back-links.

I noted on another thread that according to google we have 35 back-links to one of our sites. It comes top in searches for a reasonable number of keywords irrespective of that.

PR is not to do with ranking, which I recall was the original purpose of it. Whether it has anything to do with site stability I doubt, though more experienced webmasters here can probably prove it to some peoples' satisfaction.

Reno




msg:3985722
 8:47 pm on Sep 7, 2009 (gmt 0)

A real solution like dropping back-link usage altogether. The whole system is based on a premise that was shaky in the beginning: that people link to their favourite sites.

I'll grant that in the Age of Innocence the idea of basing PR on backlinks was clever, innovative, and effective. But like too many systems, it has outlived its usefulness. Should Google scrap it entirely? Perhaps not entirely, but it they followed some of the suggestions in this thread -- in particular, "under no circumstances let incoming links hurt the website being linked to" -- it could still have some positive impact, without the downside. The cat is out of the bag, to the point where that downside is considerable, and the danger seems to be getting worse, not better.

............................

callivert




msg:3985757
 12:01 am on Sep 8, 2009 (gmt 0)

They know that others can hurt your site's rankings -- if they know a whole lot about Google technically

With respect, I've got a pretty good idea how it's done just from reading this thread. The general idea is to lots of spammy links at a site, and make sure those links are obviously spammy. Spammy links have the following attributes:
* anchor text is "." or other hidden technique
* links come from pharma, gambling or other spammy-type subject sites
* lots of interlinking between the spammy sites
etc.
Beyond the details, the general strategy is to use old black hat techniques that now incur penalties.

This 77 message thread spans 3 pages: < < 77 ( 1 [2] 3 > >
Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Google / Google SEO News and Discussion
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved