Migrating from Mailchimp to Mailgun (+Metric Sneak Peak)

The Status Quo

In my effort to be as cheap as possible, I constantly evaluate all of the third party services I pay for. Two weeks ago SmartyStreets (address validation) got some optimizations. So far I've seen promising results! This week we're on to Mailchimp.

If you didn't know, Mailchimp is used for email marketing. They're one of the more popular services used for spamming the crap out of your inbox. I use them to send a newsletter 2-3 times a year. They were an easy choice when I started beestat as their free tier worked fine for my expected volume.

Unfortuntely, Mailchimp is now costing me $30/month. This is with their cheapest monthly plan and between 1,500 and 2,500 contacts. It was good while it lasted, but it's time to end it.

Why do you need a service for this?

Before we get to Mailgun, let's look at the purpose of these companies. At their core, these are services that let you manage lists of email addresses and send emails to them. They have tons of really cool marketing and tracking features. For example, here's statistics from my most recent email:

This information is somewhat useful, but I'm not doing aggressive email marketing and I don't use this information for anything other than casual interest.

Probably the most useful part of any mailing service is their ability to send lots of emails in a way that is not considered spam and is generally tracked to be successful or not. I could technically set up my own email server, but I have no interest in the level of configuration and management that requires.

Enter Mailgun

So Mailchimp is feature overkill, and running my own server is maintenance overkill. The middle-ground is in services like Mailgun. With this type of service you get the ability to easily send emails without all the crazy marketing and tracking features. And at my volume it will only cost $2 per email I send.

That's right...Mailchimp wants to charge me $30/mo just for the privilege of being signed up with them. Mailgun is free until I send an email, at which point they'll barely charge me anything. 

I currently use Mailgun to send all emails for the beestat community. That's about 200 emails/month which is free. Adding my newsletter to this made perfect sense.

Technical Setup

Switching things behind the scenes took an evening of effort and changing a few processes. With Mailchimp, any time a new user subscribed I would blast an email off asking the user if they want to subscribe. This was handled automatically by Mailchimp so it was an easy choice. Mailgun, being a simpler service, doesn't have this capability.

Instead I just ask you in beestat. You may have seen this prompt recently if you are not a subscriber:

So I had to spend a few minutes building this, but to save $30/month it's an easy win. The added benefit is that I'm not sending Mailchimp the email address of all my users as soon as they sign up. I'm also pretty sure Mailgun has less aggressive tracking and analytics.

Other than some changes to the workflow, making API calls to Mailgun was overall trivial to implement. You can check out the code if you're interested.

In the end it took me about four hours to update beestat and migrate everything over. Benefits include increased privacy and savings of $30/mo which can now be used for more worthwhile things. Since switching to Mailgun I have gotten 300 new subscribers. This will settle down but it's interesting to see the massive volume spike.

Metric Sneak Peak

Disclaimer: These still need a lot of work.

There's a lot to unpack here. I currently have four different metrics, each of which can have multiple instances for heating, cooling, and each individual stage:

- Setpoint
- Runtime per degree day
- Balance Point
- Differential (This one might go away, but you can see there's a lot of people who use the default 0.5°F)

I have most of the backend generation figured out and adding metrics is painless now. The GUI is in a decent place but it's very cluttered and difficult to find what you want. I need a couple more weeks of work to clean things up and validate the data. At that point I'll be able to release all the new features to Patrons. Early access will be a little different, but I'll provide all the details when we get there.

Until then, happy to take feedback or answer questions about any of the above. :)