There are a number of reliable ways to detect a read. The most common is to sent an HTML-format message with an embedded graphic link to the sender's server.

Seems to me that that can, in principle, produce both false negatives (images turned off in client) and false positives (if a client were to parse HTML-format message and retrieve embedded graphics autonomously before the user actually displayed the message for reading).

Even ignoring the fact that "my client has displayed the message, including any embedded images" is not equivalent to "I have read the message" (which is especially a factor in clients with a preview pane, where simply being the most recent message when the client is opened may result in the former condition, but quite often not the latter if the user is opening the client to send a message or look for a particular message.)

Note: this will not work if the person you are emailing is using gmail because gmail caches (and changes the address of) all images in the email so it appears that the image was opened immediately.

That's not reliable enough. A lot of email clients (or users) have images off by default, at least when reading emails from new senders.

> That's not reliable enough.

It's reliable enough for a spammer. Remember spam is a numbers game -- things only have to work as hoped a small percentage of the time for the system to work.

Try it with my Mutt:)

This is a future feature and we are examining which method is best and whether people actually want it. Would you adjust the content of a bump based on whether or not the person has viewed the email? May cross the Creepy Line. Maybe not. Open to feedback.

Maybe pixels are placed in the emails?

Most likely yes. Same as how Sendgrid and Yesware do it.