Spammers are getting more sophisticated in generating fake sites– here’s one

A guest blog post by Dennis Yu, CEO of BlitzLocal, a local advertising company.

I found 25 new comments today when I logged into one of my blogs. Normally, akismet catches the spam, but this spam was particularly believable, even though it didn’t make specific reference to the content of the blog post. See here:

I’ll bet most people who see comments like this would approve them– they look so natural and even compliment us. Who doesn’t want to approve compliments?

So then I took at look at bestnewspolitics.com, which sounds like a spam domain. It also, at first glance, looks legit– when you skim the article titles and content. But then if you actually read the articles, you know it’s fake. This spammer is far more sophisticated than the auto-generated content you normally see:

And then you notice the ads for adult friendfinder, weight loss, and spammy links all over.

As excellent of a job they’ve done, they could still improve by:

• Finding images that actually have something to do with the articles.
• Comment spamming by leaving just ONE comment, not 20.
• Having a slightly more professional wordpress theme.
• Generating fake comments on their own blog posts to look natural– I don’t see comments on any of the articles on the site.
• Reducing the number of blogroll links– there are a few dozen, all with spammy anchor text

What surprises me is that nobody has filed a Google Spam Report on these folks yet. And their content generation is sophisticated enough that it will probably pass Google hand review. As you know, Google employs thousands that follow a checklist to see if a particular site is spammy.

I’ve seen folks who’ve created systems that generate fake profiles on twitter, tweeting realistic messages. And then when you check their profile, you see they seem to be on Facebook, LinkedIn, and even seem to be employed part-time. They ask for your forgiveness for the poor English because they are a foreign exchange student or something. So you naturally don’t judge the awkward grammar of the auto-content generator.

Pretty clever, eh?