Mollom Update

Mollom logoIn earlier blog postings I've talked about using Mollom's services to mitigate spam that originates through a website's Contact page. In this post I'll show effective it can be.

I'm not going to identify the site/client where this happened; it just isn't necessary. This client uses Google Apps to host their email and has done so for a long time. Google does a very good job of filtering/stopping spam. But, like all spam mitigation techniques it isn't perfect. Something changed and a small, but annoying, number of email messages began getting through Google's filtering. It was time to do something.

Our first approach was to install the Mollom contributed module; create a Mollom (Free) account; and configure the module. The results were dramatic. With Mollom providing the first line of defense and Google providing a second, spam submissions stopped! That doesn't mean that all spam will be stopped because there are ways that crafty spammers could get through, but those avenues are difficult enough that the spammers are likely to "climb easier trees." Besides, these techniques do not provide the spammers with any feedback to indicate that their spam attempts have been rejected.

How effective is Mollom? Fortunately, Mollom provides us with a graphic report of its activity. Here's a picture of that report from this site:

Sample Mollom Statistics Report

Just in case those statistics aren't easy to read on your screen here they are in text: "Mollom blocked 15692 spam attempts in the past 29 days." We think that's pretty significant; well over 500/day!.

Those with a keen eye will note that it doesn't appear than any "Ham (not spam)" was accepted; that's incorrect! In fact, ham messages were successfully processed. Their number, however, is small enough that they get lost in the imprecision of the graphing mechanism.

But let's not stop with just Mollom, good as it is. Most of Mollom's effectiveness comes from its text analysis function. This requires that our website communicate with a Mollom server to pass the original message to Mollom and receive Mollom's response. While it's free from a monetary standpoint, it isn't free when we consider computer and telecommunication resources. Can we eliminate some of that overhead? Of course we can (you didn't expect the answer would be "No," did you?).

Honeypot is another Drupal contributed module that stops some spam submissions. It's not necessary to describe the technique here. Suffice it to say that Honeypot stops some spam if particular conditions are met. It does this without the need to communicate with any other service OR interacting with the site visitor (human or robot/spambot). Thus, Honeypot will stop some spam messages even before the website invokes the services of Mollom.

What we've now implemented is a three-tier structure for mitigating spam: Honeypot, Mollom, and Google's anti spam processing facilities. That's a pretty effective lineup.

Troutreach Technologies will implement these facilities, at no additional charge, on any client website either as part of our initial build or under our Quarterly Maintenance Plan. Mollom and Google Apps require that the client create the appropriate accounts with these vendors.