I started a podcast about a month ago, and for various reasons I decided to self-host it rather than use a podcast service. I am doing this mainly because I want the episodes to be available indefinitely, even if I stop making new ones, and I don’t want to pay for just hosting. I also don’t care about analytics, and I have the skills and desire to learn how to self-host.
I think this is the wrong choice for almost everyone who podcasts.
But, if you got this far, I will say that it’s probably right not to just put your mp3 files on your web-host. I haven’t really done the math, but these are large files, and if you get any kind of traffic, it will probably be expensive and possibly send you over your caps.
I’ve decided that the minimum I need to do is to use S3. I think it’s probably technically correct to also use a CDN, but I’ll cross that bridge if I get more traffic.
(If you have no idea what S3 or a CDN is, I really recommend you do not go down this route)
There are a lot of good guides out there for the specifics. I used these two:
In addition to setting up a bucket for your .mp3 files and artwork, I suggest you set up a separate bucket for logs and then send web access logs to that bucket. The AWS official docs are good to see how to do this.
By having logs stored you have enough to get some simple analytics. There are services that can read and graph the data in them.
I will post soon about how I scripted a simple way to get episode download counts.