Lessons from Comment Spam

A while ago I posted about comment spam, and discussed the structure by which spammers post seemingly innocuous comments on blogs automatically in order to return later with a more aggressive post that takes advantage of them being approved the first time.

At the time I was getting a lot of comments thanking me for my “awesome post, it saved much time”, and I wrote that there were around 720,000 occurrences of this in the Google index.  That’s gone up to around 770,000 at the time of writing, but I’m seeing a lot fewer posts using the same syntax, which is interesting.

The latest tranche of spam I’m seeing is generally similar in theme to the above, congratulatory about the quality of writing and insights I’m providing (thanks spambots :)), but it’s riddled with random misspellings.  I guess this is a technique to try and avoid the posts being picked up by Akismet, but they still are.  the other thing that I’m seeing at the moment is a lot of spam posts that link out to Facebook, Bing, Yahoo and Google:

Random Spam Comments
Random Spam Comments

It’s pretty clear from this that the goal of the initial spam comment, as before, is to appear complimentary and innocent enough to get approved so that the auto commenter can return later in order to post the actual spam payload, which will presumably be filled with links promoting ways to enhance the dimensions of your willy.

What’s a better way of comment spamming?

This kind of rubbish is pretty obvious, it’s lazy and ineffective.  You might get away with a few successful posts here and there, but like a lot of lazy SEO, it’s ultimately self destructive.  If you want to be effective at comment spam, you should think things through a bit more.

Firstly, look for relevant blogs.  If you’re planning to spam your poker site, look for sites about poker rather than sites about Hello Kitty.  You can still use the technique I mentioned in my last post about scouring Google for sites that include “powered by wordpress”, in the footer.

If you’re doing an automated scrape of Google for relevant posts you can violate, it’s possible to be pretty specific and still get a good target list:

A sample spam search
A search for places to spam

So now it’s time to comment.  If you want to make your comment more useful and likely to get past the spam blocker, you want it to seem like a legitimate entry in the list of comments.  So try and make it more relevant to the post you’re commenting on.  Something like “that saved me loads of time” is obvious, so why not try and ask a relevant question or request more information.

There are loads of posts explaining how to use regular expressions to find specific strings of text within a web page, so why not read one.  Then add a loop into your comment spamming script that runs through the content of the post looking for some content that can then be recycled into the comment that you will post.

A lot of posts about poker strategies will talk about when to play and when to fold based on your hand.  You could look for the following phrase  in the content of the post:

time to fold

You might grab this text and the section before and after it up to the nearest full stop, which might give you some content like:

it’s generally the right time to fold when you know you have a weaker hand than everyone else based on the flop.

Your script could then rephrase this content into a question based on some rules that add additional words and remove selected stop words from the content you’ve scraped:

You say it’s the time to fold when you have a weaker hand

You would also want to add in some kind of specific case to the comment:

You say it’s the time to fold when you have a weaker hand, but what if I’m confident that I can bluff?

You then have a comment that you can post which is relevant to the article you’re commenting on, is realistic, is potentially more useful, and is therefore more likely to get approved.

Constructive Comment Spam

Ultimately, if you’re the kind of person who does comment spam and thinks it’s viable, effective SEO, then you’re probably not terribly interested in adding value to other people’s blogs with your behaviour, but if you refine your technique to make your spam constructive and useful, rather than just annoying and worthless, then you’ll probably find that the rewards are commensurate with your additional effort.

XKCD comment spam
Constructive Comment Spam - courtesy of XKCD