Why scanning outgoing mail is a lot harder

An employee checks outgoing mail on a sorting conveyor

If your system is abused to send spam or viruses, that can cause a lot of problems. Other people’s PCs can become infected, and your domain or IP address can end up on a blacklist. With all the consequences that brings. So it’s no surprise that fairly soon after Onlinespamfilter started, we were asked whether outgoing email could be scanned as well.

Fighting spam is largely based on ‘trust’ or ‘reputation’. Before an email is actually offered to Onlinespamfilter, the sending server first sends a request: ‘We would like to deliver mail’. That request gives us the opportunity to check how reliable the server is, whether it plays by the rules, et cetera. Before we accept the email, we often already know very well whether we need to pay extra attention. If that server is reliable enough and the content is also judged to be reliable, the mail is let through. So you could say that a server’s reputation is an important part of assessing email. That reputation is a lot harder to check for customers’ outgoing mail; after all, we essentially trust the customer’s server by definition, unless there is evidence to the contrary. This effectively rules out a large part of the checks we can normally use when scanning email.

Structure and content

That’s why, when checking outgoing email, the focus is on different things. The most important aspects are the structure of the email and its content. Over the years, ‘the Internet’ has made agreements about how an email should be structured. These agreements are laid down in so-called Requests for Comments, or RFCs. Think, for example, of the syntax of email addresses, but also the use of text and HTML code in emails, and attachments in an email. Spammers are known for often not sticking to these. The structure of their emails often deviates a lot from these standards: no text part in the email, only HTML code, an unusual encoding of an attachment, or essential parts of the header missing, for example. For Onlinespamfilter, those are important ways to check outgoing mail for spam.

False alarm or not?

Unfortunately, there are also legitimate systems that don’t have their affairs in order. These can be printers and scanners, for example, but also accounting software or emails generated by web forms. When this kind of thing is detected, the mail is sometimes (at low volumes) not immediately classified as spam, but, because there is doubt about its reliability, sent via a specially reserved IP number. This helps us keep our regular IPs, which we use to send ‘clean’ messages, free from possible reputation damage. The chance that these emails will be marked as spam is also considerably higher than normal. If an email is rejected, you receive a warning about it via your mail server. In the user portal you can then still send the email if you wish. Nobody wants to receive spam. The problems that can result are very annoying. What is perhaps even more annoying is when your own system sends spam. Not only for yourself, but certainly also for all your contacts.

Would you like Onlinespamfilter to check all your outgoing email too? Then get in touch to see how we can help you with that.