Webmail.us customers have come to expect a high-level of accuracy with our spam filters. Since 2002 we have done very well at keeping our customer’s inboxes clean, by constantly evolving our filters to stay one step ahead of spammers. In recent months however, we are often asked why so much more spam is getting through the filters. We have slipped a little bit and customers have noticed.
There’s a lot of work that goes on behind the scenes in order to fight spam… Machine learning algorithms automatically generate new rules based on customer spam reports (the “Blacklist Sender” button in webmail), new spam sources are constantly being identified, and our engineers write custom rules as necessary. Spammers are very smart though, and this alone can’t beat them. A vital component of effective spam filtering is the ability to implement new filtering strategies faster than spammers find ways to beat your old strategies. And this is where we have been slipping.
Our focus lately has been on growth, and ensuring that our systems can scale as our customer base multiplies. During 2002 through 2005 we were adding new spam fighting strategies monthly. As of today we have not added a new strategy in almost 6 months.
Today we took the first step to reverse this. We hired a software engineer who’s time will be 100% dedicated to building our new spam fighting strategies. We have many cool projects planned on this front, all of which will focus on one thing… beating spammers. Go get ‘em Mike!
PS: On one random day last week (Wednesday) we blocked 30,308,827 spam emails, and accepted 4,783,041 non-spam emails. That is 86% spam.