18 December 2012

The most subtle spam I've ever seen - updated


All of you who maintain blogs know that the battle against spam comments never ends.  I spend time every morning weeding the comments to remove obvious spam, tricky spam, and some of the borderline my-business-is-linked-in-my-name (the latter I judge on a case-by-case basis, depending on the value and relevance of the comment).

Yesterday I encountered a new form of trickery.  The only reason I spotted it was that this type of wording is typical for spam - generic, broadly applicable, and complimentary - designed to be retained by the recipient blogger.

I had to look closer to discover that the two periods between "posting" and "thanks" are in red, and were in fact links - to two different sites.

I'm not sure it's right to call this "spam," since the person is not trying to get traffic.  I presume it's a form of black hat "search engine optimization," by which someone hopes to improve the status of their blog or website by being "linked" to mine (and I didn't know that search engines rank sites by links in the comments - I thought they just used links in the body of the posts).

This would have slipped by me if the "comment" had been in any way relevant to the text of my post.

Addendum:  For a thorough explanation of this phenomenon by a knowledgeable person, see the discussion by Benjamin in the Comments.

Updated:  Here it is again (that red period was a link to spam):


I monitor every single comment written on this blog.  They all come to my email inbox for review, which at times becomes a tedious chore, but which I think helps improve the quality of the blog.  Any comment from an "anonymous" gets extra scrutiny, and is subject to a more vigorous deletion policy.

Posted for the many readers who maintain their own blogs.  Don't let yours get "used" in this way.

18 comments:

  1. This was a nice posting.
    Thank you for sharing it.

    No, I'm completely serious.

    ReplyDelete
  2. I figure if someone is curious about me, they'll either click on my avatar above or do a Google search on my name or other things relative to myself. The only way that is going to happen is if I am making a substantive and quality commentary relevant to the given topic in any thread, and readers find that compelling enough to want to find out more about the individual who wrote it.

    I would much prefer the visitor counts and contribution on my own web site and blog to be based on a firm foundation of reality, not on the shifty sands of subterfuge. I gain nothing by such trickery. I benefit from people who would come to my site trough their own curiosity and volition, as that guarantees regular readership. I also believe in the good neighbor policy, whereby my own lot will improve if I aid my neighbors to improve their own. The best way to do that is to be a contributor instead of a spammer.

    ReplyDelete
    Replies
    1. Oh, I have no problem at all with regular identifiable readers who have links to their work. In fact, I often visit just out of curiosity, and not infrequently find blogworthy material.

      Delete
    2. Well, I guess I could sneak in a reference to williamdrichards.com and my blog. With your permission of course! ;-)

      Delete
    3. Ummm... they both look the same...

      Delete
  3. I'm thinking it'd be a clever ruse to write a virus that intercepted real blog comments and added link tags to the spaces between words. I tell you, I could be a millionaire if I had fewer principles.

    ReplyDelete
  4. As someone who's stumbled into a job in the internet marketing industry, I'd love to throw in a few facts that I think most people aren't aware of, or at least don't fully grasp the ramifications of.

    Those periods are indeed a form of what we refer to as "link spam" - in this case, they are the sub-category of "comment spam." The assumption that they're not meant to drive traffic is absolutely correct. Their sole purpose is to create links pointing at sites that are trying to soak up money from unsuspecting people when they appear on the top of the search engine results pages - or "SERPS." Yes. This is a black-hat technique. I personally tend to avoid that phrase, because I think it implies that such internet marketers have the level of sophistication of much more intelligent people who are the classic unprincipled hackers/crackers. I prefer to just refer to link spammers with profanity.

    While this may or may not be the official position of Google, it is widely understood that links pointing to one's website (or "backlinks") have the greatest effect on a page's rankings on the SERPS. There are a number of things that one can do to sink their page, but in general if you want to rise to the top (and that first result gets roughly 30% of the traffic for the average search) then you need links.

    When Minnesotastan takes a break and offers links to blogs written by readers, this is EXACTLY what Google is looking for - a respected webmaster offering links that are intended to imply that the targets are trusted. This is the way that Google sorts its index. Unfortunately, the spammer in this case sought to directly co-opt this function by adding links him or herself. This is one of the most classic schemes out there. Blog comments and message boards which allow free access are inundated by these on a daily basis.

    The irony is that Google is hip to this. Recent updates to Google's algorithm have targeted sites with "suspicious backlink profiles." If one's website has nothing but thousands of links from websites that either A) are nothing but giant linking pages themselves or B) have anchor text that consists almost entirely of a single keyword related to the targeted page ("Viagra" comes to mind) then a site can pretty much vanish overnight. The whole operation is a good deal more sophisticated than this, but this is something of a nutshell overview.

    It's also for these reasons that there is a link attribute developed by Google and others called "nofollow." This attribute handled differently by different search engines, but in the case of Google, it is a signal that a link is not considered to be an endorsement of the targeted page. Many blogs (most Wordpress blogs for instance) have this attribute applied universally by default. In fact, it was developed specifically for blog comments sections. In general, it stops links from passing any value. Some surmise that such links actually still benefit websites, especially in great numbers, but this is a matter of debate. I personally recommend implementing global nofollow in comments in most cases. It's one thing to allow a commenter to add a functional link that can be clicked - it's quite another to allow them to create the impression of an endorsement.

    I would like to add that I would not be writing these things if my employer were in any way involved in them - nor would I be showing up to work every day. I am proud that the basic focus of what I do is simply to help clients develop better websites that have high visibility to search engines and excellent usability for users. That's how things should work, and it's why Google doesn't seek out and delete SEO companies from its database - the best SEOs are essentially trying to make the web a better place - even when you're just shopping for shoes.

    ReplyDelete
    Replies
    1. Wow. Thank you, Benjamin, for a definitive explanation.

      I don't know why your comment got trapped in the spam filter; perhaps length is a screening criterion for the Blogger algorithm?

      Delete
    2. Haha. It's because I'm in internet marketing - and therefore a spammer!

      Delete
    3. This largely explains (and highlights) why I've been conditioned to skip the first three to five results in any given Google search.

      In fact, unless it's a "word of mouth" referral, I (probably mistakenly) believe that my internet purchasing choices (i.e., where to shop, look for deals, etc.) are largely set in stone.

      Delete
  5. Over at talking points memo, there's a post about an take-down notice the site owner received from a company that had placed this kind of comment spam on the website in the past. Apparently, Google's new algorithm docks points for these types of links, so the company's previous SEO efforts are now hurting its ranking. Solution? Send TPM a take-down notice demanding the removal of online articles that contain the company's comment spam.

    http://talkingpointsmemo.com/archives/2012/07/chutzpam.php?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+Talking-Points-Memo+%28Talking+Points+Memo%3A+by+Joshua+Micah+Marshall%29&utm_content=Google+Feedfetcher

    ReplyDelete
  6. This has nothing to do with spamming, but rather my reaction to the fact that you review every comment. A few weeks ago I got "sucked in" to your site, and was clicking some articles that I hadn't seen. I got absorbed in one, and left a comment/answer to someone who had previously commented. Right as I finalized the comment, I realized that the conversation had ended two years ago. I felt simultaneously like a time-traveler and a bit of a dummy. I thought, "Man, I bet Minnesotastan is going to read this and think I am seriously late to the party."

    I blame it on the good content. ;)

    ReplyDelete
    Replies
    1. Actually the older stuff gets quite a few comments because for whatever reason the posts often wind up high on Google searches. And now that I've added this "you might also like" internal reference widget at the bottom of posts, I think more people are visiting the old material.

      Delete
    2. I concur; I honestly can't remember exactly how I discovered this blog, but it is definitely a favorite!

      Delete
    3. I know I'm pretty late to the party but I too don't remember how I found this blog but visit frequently :)

      Delete
  7. Just guessing but it seems like a proper comment system would convert incoming text to plain text, thereby removing hidden links before posting. A hard coded link (the URL spelled out in the clear, i.e. http://tywkiwdbi.blogspot.com/2012/07/most-subtle-spam-ive-ever-seen.html) should pass through.

    ReplyDelete