I was having a discussion today with a few industry colleagues about engagement and open rates. It was a good discussion and inspired a couple blog posts. Engagement totally matters, Engagement affects deliverability, and ISPs should be the last of your concerns.
I think they’ve covered the engagement issue pretty well, but what I wanted to talk about was metrics, specifically opens. Open is a fairly simple word, and it’s used in email all the time. Recipients open email. Mailbox providers measure that open. Senders measure that open.
It’s critical to remember, though, that open rates as measured by free mailbox provider and open rates tracked by a sender are not really the same thing. They’re measured in very different ways, and there is not a 1:1 mapping between the two measurements.
- Free mailbox providers actually track that the message was opened. They can see the status change from “unread” to “read.”
- ESPs track when a recipient loads an image.
ESPs can’t track the status of a message inside the recipient’s mailbox. And free mailbox providers don’t need to track image loads.
So even though both groups claim they are tracking opens, how they’re tracking gives different data to the people measuring the information. Gmail sees me open mail all the time. Most of my clients never see me open an email in my gmail account.
Free mailbox providers and senders are using the exact same word (Open) to describe different things (rendering an image vs. actually opening the mail). I think these things are different enough to say that an open as measured by a free mailbox provider and an open measured by the sender are not the same at all.
The crux of it is that even though mailbox providers use the metric of “open” to look at engagement and even though senders use the metric of “open” to look at engagement, they’re actually looking at two totally different things.
RSS - Posts
This is interesting and I had never really thought of the Read/Unread status as a way to track opens. For Gmail, I don’t use their web interface and I have my mail client set to not automatically load images. One thing I find about this is that some messages I receive regularly and expect to receive and always open (without displaying images), I have found that Gmail will occasionally send these to the junk folder. I figured it was because I wasn’t using their interface and they couldn’t see my regular engagement with the email. But with read/unread over IMAP, they should still be able to tell, correct?
How can “opens” be tracked if the recipient mail system cache the images?
The images have unique URLs for each recipient, and you typically only care that a user has looked at the mail, not how many times they have.
The first time the user looks at the mail the image won’t be cached, so it’ll be fetched and counted as an “open”. If they open the mail again then the image might be cached, or might not, but you’d ignore the duplicate open anyway, so it doesn’t really matter. (If you do care about multiple image loads then you can add cache-control and expires headers to the image to reduce the chance of it being cached).
If an ISP preloads images then you might see an “open” even if the user didn’t actually open the mail – but that’s not common behaviour.
Steve,
I’m not referring the user’s web browser caching the image. I’m referring to the ESP of the recipient caching all images within the email and rewriting the image sources in the HTML of the body of the message. I suppose this may follow what you’re suggesting with “pre-loading” images. From my understanding, this is pretty common among the big ESPs.
Under this scenario, you’re getting a falsified sense of an ‘open’ as you’ll see the ESP downloading the image and not the actual recipient. I suppose you can compare the IP of the view to check whether it was within the subnet owned by the ESP.