April 15, 2004

gmail and spam filtering

A quick note about Gmail and spam filtering. Their filters out of the gate are pretty tight, but not perfect. They’re not catching everything, and there have been quite a few false positives – not on personal mail, but on things like the New York Times daily news headlines. I assume that the “report as spam” feature (which moves messages to your spam list) feeds back into Gmail’s spam filtering tools; and I also assume that the converse is true – if you mark a mis-classified email as “not spam,” that action will contribute to their rulesets. It’s only a week into the beta, but a week of consistently telling the system that the New York Times is not spamming me hasn’t had an impact on their filters. But hey, I’m just one guy.

That said, there’s a “yet to be documented” feature of Gmail to help work around these issues. If you add the sender of the message to your contacts list, Gmail will consider those messages “not spam.” Which means they already have the basics in place for a more robust whitelist filtering system. And, of course, it’s a no-brainer to imagine layering in your Orkut’s social network to further extend your whitelist. (A no-brainer to imagine. Not a no-brainer to implement in a seamless and elegant way.)