Will Amazon SES Change Everything?
Amazon Simple Email Service (universally known as SES) is the new email server platform recently announced by Amazon. It's an obvious (and really smart) direction for Amazon to go. At a very high level, here's how it works:
· Amazon runs some email servers in the cloud
· Sender can submit messages via an API
· Amazon sends messages
· Sender can pull some basic analytics via API (SMTP accepted rate, etc.)
There has been a lot of excitement about Amazon SES among our client base -- and in certain dark corners where email technologists like to hang out.
So how important is this? After a bunch of conversations with my colleagues at Return Path, our investors and some super-smart email technologists, here's my take:
1. I expect that a lot of really interesting applications will be built on top of Amazon. By itself, SES could serve as a very basic transactional mail system. If your current needs could be met by setting up a sendmail server and sending out transactional mail, this is a lot less work. However, SES currently lacks some of the functionality that would be good to have in a transactional mail system: link tracking, native DKIM support and a few other features. I don't think that it is Amazon's intention to build out all the "bells and whistles." Their intention is to have people build on top of their platform and they can add the advanced features. That's really cool. For an example of what is possible, see what MailChimp is doing with Amazon SES. (http://blog.mailchimp.com/mailchimp-launches-transactional-email-service-on-top-of-amazon-ses/]
2. I expect that Amazon will do a good job at preventing the worst spam from being emitted from their system. The potential problem with a cloud-based email service is that it can easily be overrun by spammers. We've looked at the mail emitted from SES IPs, and we haven't seen the kind of egregious spam we might have expected. Amazon has put some thought into this but they will have to continue to adjust their techniques as spammers adapt.
3. I don't expect that Amazon SES (or other cloud-based MTA services) will replace most Email Service Providers. If you take a look at a lot of the effort and code that goes into creating campaign-based email, most of it isn't in the MTA. The templating, campaign management, professional services and customer service provided by ESPs is a lot of work. I don't expect those kinds of products and services to be replaced by Amazon. In fact, I think SES could be a net positive for ESPs, especially smaller players who might look at moving some of their backend to Amazon SES. Systems like this could also bring new clients into email -- as their needs grow ,some will eventually move over to more robust platforms.
4. I don't think that Amazon SES (or other cloud-based MTA services) is a magic bullet for deliverability. Most of the reputation factors that are measured by mailbox providers to determine inbox placement are not a function of the mail server that the mail goes through. Complaint rates, unknown user rates, how much time a user spends on a message, and spam trap hits are not factors controlled by the mail server. All of those elements are controlled by the company sending the mail through the mail server. If you have bad practices that drive bad metrics, no amount of mail server magic is going to save you.
What are your thoughts?