For more details on when we bounce spam, see what each spam setting does.
In order to bounce a message, it must be rejected before we receive the whole message. At that point, the only things we know about a message are:
- What server is trying to send it to us
- Who it is coming from
- Who it is going to
What we don't have is:
- The message content
- The Subject:
- The From: header, which includes real name and, sometimes, a "sanitized" email address
This is because you have to tell the connecting computer that you want to bounce the message BEFORE they send it.
The great majority of Pobox spam conditions are based on the reputation of the connecting computer. So, we can make a decision about which condition would catch that message, how accurate that condition is, and decide to bounce the message without knowing what's in it.
(The one condition that runs on the whole message, Cloudmark, cannot be set to bounce your mail; it can only hold messages.)
Why can't we get the whole message before deciding to bounce it? Because that would either require us to accept the message, then bounce it, which is a big anti-spam no-no (causes bounces to go to forged addresses), or hold a connection open while we scanned the message content, which would slow down mail delivery for all customers. And, if a message was addressed to more than one person (like if it were CCed), it wouldn't work at all, because once you accept the message, you either have to bounce the message for everyone, or accept it for everyone.
This means:
- Subjects are not available for bounced messages. Instead, we replace it with the condition that caught your message.
- The From: address you see may be a "behind-the-scenes" address, used for tracking or other purposes. It doesn't contain a real name.
- You cannot release bounced messages; you can only add their address as a Trusted Sender, so future messages will be delivered.