Recently The Rocket Science Group, the company behind both Mandrill and MailChimp, decided to change things up. They decided to roll Mandrill, their transactional E-Mail service, into MailChimp as a paid addon available to paid MailChimp accounts only. A lot of people freaked out or got really upset, most of them focusing on the fact that many people who were using Mandrill for free or close to it, were going to have to start paying at least $40 per month. That’s $20 per month for a MailChimp account, which they might not even have a use for, and $20 per month for the lowest tier of the Mandrill transactional email addon.
I was upset too. Not because of the additional cost, but because of the way Mandrill users were treated. An email went out to all Mandrill users on February 24th, mandating that all existing users needed to have a paid MailChimp account set up and connected to Mandrill by April 27th. That gave nine weeks.
Nine weeks isn’t a lot of time, and Rocket Science knew this. Is it enough time to set up a MailChimp account, link it to your Mandrill account, and pay them the extra money every month? Yes. Was it enough time to research an alternative, set up an account elsewhere, rewrite all your transactional email code to use a new API, train users to use a new interface at this new solution, test all the new code, and deploy? Not really. At least not in many cases.
However, that’s exactly what I did. Not because I was upset at the extra monthly cost, but because I didn’t like being treated like that.
I started with some simple research to find alternatives. As it turned out, plenty of other people were doing the same research and posting their findings, which greatly simplified the process. What I found is that since my usage was pretty straight forward, almost any of the available alternatives would work for me. I ended up choosing SendGrid.
The actual development wasn’t particularly interesting. The APIs are different, as you would expect, so all the code needed to be changed but was ultimately similar enough to be pretty simple.
One of our hangups came with tags. In Mandrill we used tags to label various kinds of E-Mails, what server the E-Mail was triggered from, etc. We use those to help us track where deliverability issues occur, as well as to help us track down bugs when they happen. The problem was that SendGrid didn’t have these. Luckily, SendGrid DOES have what they call “Unique Arguments“. Basically, they let us do the same thing, adding in our own unique key/value pairs, with the only downside being that their web app doesn’t give you much in the way of working with those (like viewing all bounces with a specific value for one of the arguments).
And that seems to be the only real downside so far. The web app for SendGrid doesn’t seem to be quite as powerful or fully functional as the one Mandrill has. Having said that, delivery, responsiveness, speed, etc all seem completely on par. I’ll happily give up the zoomy UI though, if they’ll treat their customers with a little more respect.