Member
My .xyz for a Minecraft server notification bot: fine. My .xyz for a "please review our hosting" campaign: 60% spam folder.
Different use case, same domain, same Postfix config. The content and recipient relationship matter more than people admit.
Member
This. I manage thirty sites. The ones with double opt-in, established lists, regular cadence — they could send from .party and land in inbox. The ones who bought a list and slapped it on fresh .com? Spam city.
It's the behavior, not the badge.
Then explain why my .com on same behavior outperformed .xyz. I controlled for that.
That was my original question, yeah. Still open.
I white-label hosting. My upstream doesn't know my clients' domains. I see the aggregate.
New gTLDs have higher complaint rates in my tickets. Not bounce rates — complaint rates. "Why is this in spam?" The answer is usually "because someone reported it." The TLD didn't cause it, but the TLD correlates with senders who get reported.
I don't choose my clients' TLDs. I just observe.
So the thread consensus is: TLD reputation is a proxy for sender reputation, not a technical block.
Thread consensus is a dangerous phrase. We have:
No consensus. Good data, though.
Following for the data.
What we'd need: same sender, same IP, same content, same recipient list, different TLD, proper volume. That's an A/B test nobody here has run.
Closest is garykwh, but he changed /24s. So it's TLD plus IP reputation, confounded.
I know. I know. My budget, my confound.
If someone wants to fund a clean test, I'll run it. Otherwise I'm going to keep using .com for anything that needs to arrive.
