On 2026/01/09 22:43, John Fawcett via mailop wrote: > > I am not sure it makes sense to talk about "the worse the domain > reputation" for domains that actually have never sent a single spam, > have never had any complaints raised by recipients and meet all the > requirements (DKIM, SPF, DMARC etc), but nevertheless have email > blocked or sent to spam folder because they are not big senders for > which statistics are calculated. Such domains have 0% spam whereas > Google is accepting email from senders that have up to 0.3% spam > providing they send enough of it.
"domains that actually have never sent a single spam ... Such domains have 0% spam" well yeah, but no receiving server can know that they've never sent a single spam. It's either "we haven't seen any mail from there yet", "we haven't seen spam from there yet", or "we have seen spam" (with the latter two giving different confidence levels about predictions depending on how much mail has been seen). So it's not possible for a receiver to assert that "such domains have 0% spam". While a bunch of private servers that don't send spam fit into "haven't seen mail from there yet" or "haven't seen spam from there yet (but not enough overall mail to predict)" I'm certain that is the case for a bunch of spam senders too. New mail coming from a higher volume sender with 0.3% spam could easily be much less likely to be spam than mail from an unknown / barely known sender. *Especially* for mail to a wide sample of recipients (the majority of whom will be far less likely to get mail from unknown sources than, say, people like us who have been on the net for longer and communicating with people from mailing lists etc). Feels like these systems could really do with some kind of "user is likely to receive legit mail from random places on the net" flag to reduce false positives without adding too many false negatives. _______________________________________________ mailop mailing list [email protected] https://list.mailop.org/listinfo/mailop
