Mailchimp for example, provides this information. It's one reason why Mailchimp and SendGrid et al are so useful. It pays for itself when we send out a newsletter to thousands of emails.
Still doesn't answer the question. How does mailchimp know? Do MUAs commonly send some kind of response to the sender when the user flags a message as spam? Or is this something they're getting via webmail services?
That's a good question I couldn't find the answer for. What I can tell you is what happens as a MailChimp user: if I send out an email and someone clicks the 'this is SPAM' button in their mail client, MailChimp notifies me that this happened so I guess that the email client somehow notifies the sending mail server using the List-Unsubscribe header.
I bet this only work for aol.com and a few others.. I highly doubt that Thunderbird & friends notify the sender. Of course It could be some sort of hack, like the image thing.
I've occasionally seen GMail ask “do you want us to unsubscribe you as well” when flagging stuff as spam (I refuse, because I don't want to confirm my address to spammers). Perhaps some provider has come up with a way to differentiate “mark as spam” from plain “unsubscribe”, but I do find it surprising.
Most likely there's some indirect, aggregate feedback as well.
