Anti-SPAM Measures

If you’re a sysop, the following is likely not exactly news to you.

I blogged quite a while ago about how my SPAM volume varied. Recently, I’ve had cause to move my email server, and consequently implemented my SPAM filtering a bit differently. I’m still using the same filter, mind you, I just reject more emails before handing them to the filter.

The result? Where previously my SPAM ratio was about 3:1, that is three SPAM messages for each good message, it’s now closer to 2:1. For the first time since I can remember, my SPAM filter actually processes more good messages. Note that the history isn’t a full 14 days yet, so the totals might change.

SPAM filter performance

So what’s different? Well, two things:

  1. I’ve implemented SPF with a SOFTFAIL policy. I should’ve known that would help, but I didn’t realize just how much it did.
  2. I’m also rejecting any email with an empty sender part in the SMTP envelope.

That second thing turned out to be even more effective than using SPF, but that may be to do with the SOFTFAIL policy being comparatively lenient. Which it pretty much has to be, because not every domain provides SPF information yet, nor will that likely happen.

Odd though, how many SMTP servers apparently do not verify what follows the “MAIL FROM” command…