Laura,
Again, thanks for your insights! Inline:
>>>> - Anyone else seeing this behaviour from gmail recently?
>>>> - Could the newly created, custom MAIL-FROM-domain cause a behaviour like
>>>> this? The MAIL-FROM-Domain has not yet been used before, but the sending
>>>> email address was the same
>>>
>>> You changed a domain that is important in authentication. That changed the
>>> identity of the message. That meant you were treated as a semi-new sender
>>> by the machine learning filters.
>>
>> I was aware that different IPs, domains or email addresses require prior
>> warmup – but I was not aware that the return-path/MAIL-FROM-domain also
>> requires a warmup. Interesting.
>
> It’s a new domain, a new identity. It’s actually one of the authenticated
> domains so of COURSE it’s important.
>
> There’s also, currently, a significant problem with spammers creating
> subdomains and hijacking reputations. Right now I’d say that filters are more
> than overly suspicious of new SPF domains. That’s just the current threat
> model they’re working with.
I was under the wrong assumption that the domain part of the from address field
(i.e. @syniumsoftware.com) was the most important part for building a
reputation – and not the return-path/MAIL FROM. Learning new things everyday.
> I will note that one ESP had a huge problem with delivery across their
> customer base when they rolled out a new domain in the One-Click
> List-Unsubscribe header to comply with the new Yahoogle recommendations. ANY
> domain in the email can have an impact on delivery.
Interesting. I would not have anticipated that :-)
>>>> Any insights or hints on how to investigate this would be highly
>>>> appreciated.
>>>
>>> You may need to back off and go through a short warmup phase to introduce
>>> that this SPF + this DKIM + this Header From from SES shared pool is a
>>> valid source of mail.
>>
>> Will do this over the coming weeks. I’ll spread out my next campaign from
>> one hour to 5-15 days.
>
> Are you really only sending monthly? That’s going to make it tough to really
> establish a reputation.
We’re sending out 6-9 newsletters a year (new app-updates etc). Before the
change in configuration, we were satisfied with all metrics (sales, delivery
and read rate etc).
>>> ALSO - did you actually have any delivery problems? One thing I’ve noticed
>>> (at Gmail in particular) is that changes in an authenticated domain often
>>> result in a lower rate of pre-fetching of images *without* any
>>> corresponding change in delivery.
>>
>> Yes. All other metrics (hits on homepage, sales, customer responses,
>> unsubscribes) were WAY lower than anticipated compared to previous
>> campaigns. It’s not a measurement thing that’s off.
>
> Fair enough. It’s worth mentioning, though. Also, have you tracked the
> metrics per ISP / recipient mailserver? You may discover the problem is
> related to one specific MX / Filter and that will help inform what to do
> about it.
No, we could not do that reliably. Most of our customers are Apple users – and
are using Apple Mail.app as the mail client of choice. Mail.app strips nearly
all information from tracking pixels for privacy reasons so it is very hard to
find out which service was impacted the most.
The only hint we currently have is the HUGE spike of failed deliveries (41%) in
the Google Postmaster Tools concerning gmail.com <http://gmail.com/>.
Again, THANK YOU very much for your comments!
Best,
Mendel
_______________________________________________
mailop mailing list
mailop@mailop.org
https://list.mailop.org/listinfo/mailop
