> I’ve been hosting my own email for many years and I have almost zero deliverability obstacles.
This is almost certainly false. If this is not false, please do show how you measured your deliverability (across all those years) and how those measurements show "almost no" issues.
Not the person you are replying to, but I've been self hosting an email server for ~3 years (which is not that long tbh, so take this data point for what it is), and have similerly also had almost zero deliverability issues.
I can tell because I also run a mailman mailing list (~500 people) on that same server and domain, I measure initiall discoverability by how many bounces I get (almost always none) and also that the high percentage of read receipts I get, and the fact that engagement with the list is pretty even across domains (IE, people with gmail, outlook, yahoo, and aol emails all frequently read and engage with the list).
Interesting! Can you share the numbers for read receipts and "engagement" for different domains? I'm very surprised if the % for gmail and outlook truly is not lower than the % for yahoo and aol, as you indicated.
I'd have to do some digging since its been a while since I've sent out a newsletter (the only thing I use the tracking pixel on to check read recepts) but IIRC a little under half of the list pinged the server with the pixel . Which seems about right to me, esp given that some clients block tracking pixels.
> I'm very surprised if the % for gmail and outlook truly is not lower than the % for yahoo and aol, as you indicated.
the distribution of domains is far from even, but I can tell by checking my inbox from today that one of our most frequent posters with an aol domain sent a news article around and it started a thread with multiple gmail users chatting with the original poster about it.
Thats not hard numbers, but everyone on the list has agreed that the list works better on the self hosted server then it did on google groups.
> This might be indicative of deliverability issues affecting some domains only.
I mean maybe? I remember when we switched from google groups that I checked and the % of opens and that it seemed fairly evenly distributed given that the percentage of gmail users far outweighs the percentage of yahoo/aol users, and to be honest I've not bothered to check much more beyond that point. But the core group of crotchety old folks who frequently send round interesting local stories are a decent mix. The last couple of days I see 4 different gmail accounts, one work domain (that I assume is outlook but might be some other provider), the aforementioned aol account, and the burner riseup account that I use when I am feeling crotchety.
> The fact that some gmail users can see emails doesn't mean 100% of gmail users are delivered your email.
Sure, its possible but I've not seen any evidence of it after we made the transition. I've seen email sent from gmail accounts to gmail accounts not get delivered, expecting 100% is never going to be realistic even with the largest mail providers.
Unless and until I start getting in person complaints (which, trust me, I would if there was a problem) or see people mention emails going into spam, I'm going to keep on keeping on.
It's not that difficult. I host my email server on an AWS machine (so the IP that gests out is clean), and all I had to do is to setup DKIM signature and create the SPF, DMARC and DKIM DNS records for the domains.
Setting all up is complex, but there are project that give all the tools in a Docker container. Since I wanted to learn it I did all manually, and did set up postifx, dovecot, opendkim, postfixadmin and even squirrelmail as a "retro" web interface, with a postgresql database to keep the mailboxes and the accounts. It was an interesting experience, plus now I have full control of the mail system, e.g. I can decide to route the mail in whatever directory just by changing the SQL query that decides in which mailbox a message should be delivered.
> It's not that difficult. I host my email server on an AWS machine (so the IP that gests out is clean), and all I had to do is to setup DKIM signature and create the SPF, DMARC and DKIM DNS records for the domains.
The most important thing is to setup DKIM, SPF and DMARC correctly. I mean sign the emails with the correct key, but also setup the correct records, e.g. SPF record must be TXT, if you create a (deprecated) SPF type record GMail won't recognize it, even if you also have the TXT one! Of course they don't mention it in the documentation, and I've spent 1 month of trial and error to figure it out. And the DKIM record needs to be formatted exactly correctly (antother thing that took me time to figure it out)
With DMARC you can specify that the recipient sends you a periodc report of the delivery status, this is useful to debug eventual problems. For example GMail and Outlook send this out, and gives you information about the fact that your DMARC is working correctly and SPF and DKIM checks are passed.
Anyway when you manage to get it working, it did work for around 4 years without any issues. I've even reinstalled the mailserver to another machine (keeping the same IP) and I did not have any issue in the migration.
Ok, so you didn't actually measure your deliverability, so a portion of the emails you send are not being delivered, just like everyone else who self hosts email.
This is almost certainly false. If this is not false, please do show how you measured your deliverability (across all those years) and how those measurements show "almost no" issues.