Engineering MailChannels Rolls Out Improved Spam Detection Engine By Sridhar Kakkillaya | 2 minute read To maintain an excellent email reputation, MailChannels works hard to block spam and phishing attacks from the more than 26 million unique senders whose email we process each month. Unlike other transactional email services, MailChannels is designed to deliver email from completely unprotected web servers. In fact, the primary service we offer to our customers is to magically deliver legitimate email for their end users while blocking all the bad stuff. As you can imagine, blocking spam and phishing attacks from unprotected web servers is challenging. Spammers frequently change their tactics and content to evade our algorithms, forcing us to continually invest in countermeasures. It’s an arms race, but with millions of adversaries facing off against one “good guy”. As part of our ongoing investment in new algorithms and techniques to combat spam, MailChannels recently rolled out a significantly improved content analysis system. While our previous content analysis technology also leveraged machine learning, the new engine is faster at processing feedback from receiving systems like Gmail™ and Office 365™ to identify suspicious or harmful content. And when you’re fighting outgoing spam, speed is everything. While the new engine is extremely good at catching spam, it does sometimes make mistakes. Fortunately, as always, if MailChannels blocks a legitimate email message by mistake, the sender can correct this by clicking on the “Not Spam” button. Learn more about reporting false positives.