Only speaking from my own experience here, but changing the d= value (and
nothing else) usually causes catastrophic inboxing issues at gmail. None of
the other major providers seem to care. But a new d=, even just adding or
removing a subdomain, causes immediate bulking at gmail. If the mail is
good, it seems to recover within a few days even without taking any action.

Double-signing, and/or rolling out the new domain slowly is definitely the
way to go.

Luke

On Tue, Aug 28, 2018 at 3:11 PM Steve Atkins <st...@blighty.com> wrote:

>
> > On Aug 28, 2018, at 2:46 PM, Jonathan Leist <jleis...@gmail.com> wrote:
> >
> > Hello,
> >
> > We're currently exploring the possibility of migrating from signing as
> the individual hostnames of our sending IPs to signing as the org domain
> aligned with those hostnames (e.g. signing as example.com instead of
> mail1.example.com). Our main concern is in regards to deliverability, as
> we'd presumably lose years of sending history that we've accumulated with
> those signing domains.
>
> Relevant sending history when it comes to delivery decisions is typically
> measured in weeks so I wouldn't worry about anyone tracking your reputation
> from June, let alone 2017.
>
> > To potentially mitigate impact from the change, I'm considering having
> the d= be the org domain, while i= could remain the actual hostname we've
> historically signed with. So with the example above, they'd be d=
> example.com and i=@mail1.example.com. Would anyone know off hand whether
> we could expect that to help preserve the reputation we've built as a
> sender, given that the i= also carries reputation (from what I've read)?
>
> That wouldn't hurt anything. But I doubt it'd have much effect, as
> recipients are going to use either the d= or the domain part of the i=, not
> both. I'd expect them to just use the d=, mostly.
>
> The "DKIM Way" would be to sign twice, with the old domain and the new
> one, for a while.
>
> But if you're not seeing delivery issues today and you're not changing IP
> addresses, just the d= signing domain, I wouldn't expect much impact from
> just changing the d=. Trying it with a single MTA would let you see any
> impact, and dribble the new d= value into your mail stream.
>
> Cheers,
>   Steve
>
>
> _______________________________________________
> mailop mailing list
> mailop@mailop.org
> https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
>
_______________________________________________
mailop mailing list
mailop@mailop.org
https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop

Reply via email to