Parasitic Hosting Lite?

[1script]

Hi all,

What do you guys think about this scenario and its effect on Google rankings:

A forums site (or any other type that accepts user-generated content for that matter) has spam posts with spam keywords and spam URLs mentioned in text but not converted into actual HTML links. Despite there being no actual links back to them, spammers still link to the pages from a few junky places. Not sure what they are trying to do - probably the whole process is automated and no one actually knows that there are no links to them or there is another angle I just don't know about.

So, it's a classic parasitic hosting scenario except the perps don't gather backlinks this way hence "Lite"

Are there people in this great group that think this could be reason enough for a real harsh Google penalty? I am investigating a case that involves 10+ sites that lost 95% of their traffic because most (thou not all) ranking keywords were pushed back -200, -300 or removed from SERPs altogether. This sort of parasitic hosting is one of the very few things that is similar about these sites.

I should mention that there might probably be 20 spam pages like that on a 50,000+ posts forum - to give you a sense of scale of the spam infestation I'm talking about.

Also, anyone has a success story to tell about submitting a recon request for a problem like that? The external pages that link to me are exceedingly hard to document - some show up in WMT, some in Y! links search but I doubt it covers all. What sort of info would you submit to Google in a reconsideration request like that?

[tedster]

Parasite content usually refers to content inserted on a domain without the owners knowledge or intention. But I assume that this content entered your site through a normal user-generated content mechanism, correct?

If that's the case, I am not aware of anyone getting a major penalty simply from "citations" rather than actual links. Not saying it can be so - just that this would be the first such report I've run into.

[martinibuster]

This might not be the case for you, but it's worth mentioning because it might be useful for someone else. There is a variant of forum spam that is sometimes referred to as Profile Building Spam. They make about a dozen posts then disappear. The goal is to build backlinks to their profile which is where the link drop appears. Forum software typically creates a link to the member's profile in every post they make. The spam link is in the profile.

...think this could be reason enough for a real harsh Google penalty?

I know of a site that had a software vulnerability exploited for over a year. They were hosting around a million web pages linking to the worst imaginable sites. I discovered it in the course of researching the SERPs and seeing what's ranking and what the quality level was of what was ranking. Did it affect their rankings? No.

I observed this site's rankings for four months to see what happened but nothing negative happened to their SERPs for the phrases I was tracking them on. I estimate they were hosting these pages for over a year based on the date the software indicated for the time the spam was posted and I believe those dates were correct.

[tangor]

All the above granted, but why would one not moderate the UGC and eliminate the spam? Automation is fun---that's google (and they can kill you)---when taking charge is so easy? Websites are not "set it and forget" doohickies.

[hugh]

All user generated content should use rel=nofollow tag and that includes member profile user "homepage" links. Unfortunately this won't stop the spam. I recently picked up on another post on this forum which suggested adding a check box to the registration process of a forum stating the above but it's made zero difference...

[1script]

@tangor:

All the above granted, but why would one not moderate the UGC and eliminate the spam? Automation is fun---that's google (and they can kill you)---when taking charge is so easy? Websites are not "set it and forget" doohickies.

I honestly had to look this up first:

doohickey: A thing (used in a vague way to refer to something whose name one cannot recall)

But, yes, I sympathize with your frustration and I do remember names of my sites. But since I also want to remember the names of my children, I cannot claim that I review every single post being made. And, as you correctly point out, automation often fails. So, in most cases the problem with this spam is that Googlebot always manages to snap the cache copy of the post before I delete it. I do RSS pings and I also do Twitter posts (automatically) when a new article is posted and Googlebot is usually there within minutes. The spam post would be gone in 1 to 24 hours and this URL will return 410 Gone but they would not return to read it again for months because there are no longer links to those deleted posts. I was even thinking about creating a page with just links to the recently deleted posts. Never did it: it would look just plain silly: a whole page of links that return 410.

And, yes, sometimes I do miss a spam post completely. It can be a busy day and other new posts push it too far back or the subject of the spam is too close to the subject of the site (happens to my sewing site all the time).

So, I think my approach to spam control is reasonable (automated filters, then manual review) but not infallible.

But if anyone knows a mechanism of announcing to Google that the page is removed, I would appreciate if you share. I do use WMT -> Site Configuration -> Crawler Access -> Remove URL in the most egregious cases but it's a manual process and in many cases I just don't have time. Do they have an API for removals, BTW?

[1script]

@tedster: Yes, the spam gets there via usual post submission mechanism. I filter what I think is 75% of it but some does get through. It used to be relatively simple: don't allow posts with an unreasonable amount of URLs and stop words in them. But spammers are always adjusting (their job, right?) and the most problematic recent type is a post with just one URL and a hole bunch of keywords that are relatively close to the theme of the site. It actually works against me in two ways: I'm more likely to miss it and then, it may start ranking because of those stuffed keywords and I could suffer a penalty for keyword stuffing. I'm adjusting my filters, too, but it's always an uphill battle...

@martinibuster: Thank you for the negative answer! I'm heart broken though: I thought I've finally found a reasonable explanation... In my mind the worst thing you can do was to host links to bad sites (even if not HTML-formatted in this case). Since my penalty is rather harsh (90% G* traffic is gone) I thought it made sense.

Well, I'm off to searching for something even more egregious than bad links. What that might be?!

[martinibuster]

Well 1script, every case is different. I don't know how Google found those spam pages, but as far as I know they were not part of the site architecture. In the case outlined above, there was an old chat script on their server, spammers found it, and started making posts on the old chat script.

Links to spam might be harming your rankings but even if they're not, it's a good idea to remove them anyway because that's a bad user experience. Do searches like this to help find the spam:

site:example.com viagra

If your forum is being spammed by bots then introduce a better sign up challenge.

[blend27]

I know of a site/forum that uses this type of spam collection on its pages to accomplish the exact opposite: Prevent SCRAPING AND SPAMBOTS from accessing the other sites/forums the owner of the sites manages by collecting/studying common techniques and IPs of the spammers and bots that are running in the wild.

It’s called a "Dirty HONEYPOT - Catch and Release".

pssssss...

[1script]

@blend27:

Yeap, good idea. I'm doing it for a couple of years myself. Except I don't have a sacrificial site, I'm just sharing the banned users, emails and post subjects between all my sites. I would not overly rely on that though because there are great many angles spammers are working. Still need a pair (or several) of human eyes and even that fails to catch all 100%.

[micklearn]

In this day and age, if I owned a forum or any site where user generated content is involved: I would try to find a way that wouldn't allow anyone to post a URL, even if it isn't an active link, until the destination was reviewed and determined to be relevant/helpful -- through means of automation or human eyes ($$$)...and/or...

I recently read that over 90% of all emails are spam. Is it wrong to assume that, eventually, most of the links out there will be/are close to that percentage? Here's hoping that attempts to spam a site via forum signatures, forum profiles and comments have reached a peak.

The only solution that I can come up with right now, is to eliminate that very option: the ability to post/place a URL on a site that one doesn't own.

Why give a spammer the option/opportunity to drop a link? So that one can identify them, and let them know that you have done that to them? Essentially enabling them to go and find the next angle?

Eliminate the link-drop options, feels like the right approach in my opinion.

[martinibuster]

I would try to find a way that wouldn't allow anyone to post a URL

The intent of your idea is good and has merit but the application would prove impractical. Discussion and membership satisfaction would dampen if members could not link out to a news source, to an authoritative source of information, to photos, and to citations from other websites.

There are workarounds that minimize the spam problem such as stopwords, active moderation, a system for banning problem IPs, a bot-stopping challenge and removal of profile links. With all that in place the spam problem gets whittled down pretty fast.

[FranticFish]

All user generated content should use rel=nofollow tag. Unfortunately this won't stop the spam.

Correct. Nofollow is lazy, and it's not a fix.

1) Doesn't stop bots or humans spamming you, or wasting space with 'I agree' or 'Great post' comments
2) Text in nofollow links is still indexed
3) Nofollow bleeds PR (at least Google says it does)

Even if you don't allow links (which I think is daft), the only solution is to moderate your content - and some of that has to be done by hand.

If a site isn't moderated then why should I bother visiting it?

Automated systems encourage spammers, NOT the other way round.

[1script]

All user generated content should use rel=nofollow tag and that includes member profile user "homepage" links.

Guys, the whole point of this thread was to discuss an issue of having URLs in the text of a UGC page but not as links, nofollow or do follow. Just plain text. Also, I have no user profile pages. So, there should technically be no reason to spam my forum because you cannot gather links from it but I assume that spammers can only survive it they automate their operations. I doubt anyone spamming there has actually been to any of my sites to verify if it merits spamming.

In any case, since the forums are doing very poorly in Google in the last two weeks, I am trying to find out if this "lite" version of parasitic hosting might have contributed to the drastic reduction in Google rankings. I think the consensus from this thread is that it was probably not the reason for the penalty.

I do not advocate leaving your forums unattended if you disable URL into links conversion. I was merely suggesting that no matter how you try there will always be some URLs posted that you'd rather not to link to or even mention and so it would be important to know the implications.

[tedster]

If the ranking drops are due to unlinked "citations" (text-only URLs), I mentioned above that this would be the first such report I've run into.

<conjecture>
There has been conjecture in the SEO community (including the black hat community) that such citations have a Google benefit. IF that is so (there's been no confirmation that I know about, so that's a big IF) it would make some sense to see a spam prevention arm to Google's approach.
</conjecture>

[micklearn]

@1script: What I meant was that monitoring links/sites in a forum should be the top priority for anyone who owns one. If that's not an option and a small percentage of linked or unlinked references might be causing harm, kill that option for a while and see if it makes a difference...there is another forum out there where links are all that really matter to them in attempts to driving traffic to their "sites"...

@martinibuster: Just curious, why did you only quote the first part of that sentence in your reply? Also, by "automation", I meant what you described in the second paragraph of your reply. I'm sorry that I did not expand on that topic a bit more.

@FranticFish: Daft?! I'll be sure to let a friend know that -- who recently experienced a rather large "attack" of forum/profile spammers (think thousands). I hope that you're not saying that martinibuster's second paragraph above is mostly false? ("There are workarounds...") Except for his mention of active moderation, isn't the rest of it describing automated solutions, and helpful in solving the problem?

[martinibuster]

I would try to find a way that wouldn't allow anyone to post a URL, even if it isn't an active link, until the destination was reviewed and determined to be relevant/helpful -- through means of automation or human eyes ($$$)...and/or...

The intent of your idea is good and has merit but the application would prove impractical. Discussion and membership satisfaction would dampen if members could not link out to a news source, to an authoritative source of information, to photos, and to citations from other websites.

There are workarounds that minimize the spam problem such as stopwords, active moderation, a system for banning problem IPs, a bot-stopping challenge and removal of profile links. With all that in place the spam problem gets whittled down pretty fast.

[edited by: martinibuster at 5:22 am (utc) on Mar 30, 2010]

[edited by: bill at 2:52 am (utc) on Apr 2, 2010]
[edit reason] tidy up [/edit]

[tangor]

But, yes, I sympathize with your frustration and I do remember names of my sites. But since I also want to remember the names of my children, I cannot claim that I review every single post being made. And, as you correctly point out, automation often fails.

I was about to respond with something clever and sarcastic until that third sentence. We are on the same page. I suppose it is a matter of scale. My forums are in the thousand(s) posts per month. If it got to thousand(s) per day it would be beyond my personal abilities. I flag for moderation keywords... and add to that list almost on a daily basis for the general (open) areas. My password/password (secure areas) are not accessible to search engines and I have less concern there, but wield the "club" more directly since those who are in that area should know better.

Spam is a constant problem, but it can be managed. Keep the bots out, make it difficult for automation to insert (captacha, double opt in, etc.).

Links in forums are what makes it feel alive to the users and visitors. Figure out a way to control that with sanitizing methods (keywords, known bad urls), banning known bad neighborhood IPs and ... for me this has been the best way ... and looking at rapidity and frequency of posting.

I've never experienced the "parasitic hosting lite" suggested. I just don't like spam in my forums and deal with it from that point of view.

[FranticFish]

@micklearn

Well your friend needs to make his sign up process a bit more rigorous then IMO.

I'm not saying that MB's comments are false.

What I'm saying is that if your business model is to provide a framework for other people to build your site for you then you better have a proper system in place to control abuse.

[softwaretool]

thats right. i agree with Frantic Fish