A new approach to handling blog spam

A friend of mine expressed great frustration at not being able to post a comment on my blog the other day. Because of blog spammers, I have grown progressively less comment-friendly, requiring CAPTCHA tests and moderating every post.

I just signed up for Mollom, a new Web service from Dries Buytaert, founder of the Drupal project. This lets me turn anonymous posting back on, and remove the mandatory CAPTCHA challenge. Mollum performs content analysis and, if it "thinks" a post might be spam, it intervenes with a CAPTCHA challenge. Mollom also tries to detect obscene and violent language, and over time it'll get smarter about that.

Like a lot of Web 2.0 services, Mollom "learns" from its users. If I mark an item as spam, that data gets passed back to the big Mollom brain to help everybody else. Mollom benefits from participation, and low-volume service (which is plenty adequate for personal blogs) is free.

Currently, Mollom claims to catch 99.69% of spam attempts, and it reports that 77% of all comment postings are spam.

The toughest challenge is a subtle one. A lot of spam today is human-written by people who have read something about SEO and go around posting plain-vanilla, meaningless comments like "good job" and "I want to know more about this" -- and then slipping in the URL of the site they're trying to pump in the Google pagerank. It will be interesting to see how Mollom handles this problem. I suspect that a couple will slip through, but one they're marked, those URLs will get the Mollom death penalty. I will enjoy that.

Comments

I am a Wordpress user but am researching Drupal to see if it is something I want to take the time to learn and implement for some web projects this year. This is a good point for Drupal Thanks for the post Bill