This is a bit of a random post, with a bit of a random data point, but it surprised me and I thought those who send via SES might want to know. Because I am not entirely sure that it's well known.
On the Amazon SES FAQ page, they answer the question, "Can my email deliverability affected by bounces or complaints that are caused by other Amazon SES users?"
The answer is basically, "mumble mumble, no not really, but." (They don't really address the potential for shared IP reputation issues; and I'm not sure I want to pick at that scab here. Maybe in a followup post.)
The interesting bit is this, though:
An exception to this rule occurs when a recipient's email address generates a hard bounce. When a recipient's email address generates a hard bounce, Amazon SES adds that address to a global suppression list. If you try to send an email to an address that is on the global suppression list, the call to Amazon SES succeeds, but Amazon SES treats the email as a hard bounce instead of attempting to send it.
Emails that you send to addresses on the global suppression list count toward your sending quota and your bounce rate. An email address can remain on the suppression list for up to 14 days.
They're saying that if somebody else's attempt to mail user xyz@example.com bounces with a hard bounce, they will stop trying to send to it for a time. Up to 14 days. It's a good way to reduce SMTP bounce attempts from a shared service, but it concerns me, because it's a global thing -- based on multi-client data -- not specific to an individual client (only). Could that hard bounce have been due to an odd technical issue or a spam block that reports as a hard bounce? One thing I recall from my days working for a large email sending platform is that clients really were not so keen on the idea of suppressing mail to user X because other client Y wasn't able to get mail through to them successfully. I'm surprised to hear that SES works this way.
This is a bit of a random post, with a bit of a random data point, but it surprised me and I thought those who send via SES might want to know. Because I am not entirely sure that it's well known.
On the Amazon SES FAQ page, they answer the question, "Can my email deliverability affected by bounces or complaints that are caused by other Amazon SES users?"
The answer is basically, "mumble mumble, no not really, but." (They don't really address the potential for shared IP reputation issues; and I'm not sure I want to pick at that scab here. Maybe in a followup post.)
The interesting bit is this, though:
They're saying that if somebody else's attempt to mail user xyz@example.com bounces with a hard bounce, they will stop trying to send to it for a time. Up to 14 days. It's a good way to reduce SMTP bounce attempts from a shared service, but it concerns me, because it's a global thing -- based on multi-client data -- not specific to an individual client (only). Could that hard bounce have been due to an odd technical issue or a spam block that reports as a hard bounce? One thing I recall from my days working for a large email sending platform is that clients really were not so keen on the idea of suppressing mail to user X because other client Y wasn't able to get mail through to them successfully. I'm surprised to hear that SES works this way.
Comments
Post a Comment
Comments policy: Al is always right. Kidding, mostly. Be polite, please and thank you.