"transactional" report type (was Re: [taugh.com-johnl] Re: [feedback-report] New version of the ARF draft)

[Steve Atkins]

On May 24, 2007, at 4:37 PM, Damon Sauer wrote:


> On 5/24/07, J.D. Falk <jdfalk at yahoo-inc.com> wrote:
>
>> On 2007-05-24 15:45, Damon Sauer wrote:
>>
>> >> Seems to me that only the sender can know for sure whether the  
>> message
>> >> was supposed to be transactional.
>> >
>> > Yes, this is true, for the first round. In which we would  
>> research the
>> > issue via the ARF and let the ISP know that it was truely a
>> > transactional message and possibly provide further information  
>> to the
>> > ISP via a back-channel about which of our servers, IP range,  
>> domains,
>> > etc. are used in transactional mailings.
>>
>> Can't you do that anyway?  A report type of "spam" doesn't mean you
>> can't attempt to contact the report generator with questions/ 
>> concerns.
>>
>>
> Sure, I am just asking that a method be in place that would allow for
> the ~possibility~ that if  an ISP does do transactional tracking, that
> it can be labeled as such.
>
>
>> >  1) Diminish or absolve the scoring for mislabeled spam reports.
>>
>> Scoring and other spam filtering techniques are entirely up to the
>> system that received the message.  Adding more report types won't  
>> change
>> that.
>>
>
> This is where we differ in thought. I see transactional messages as an
> class of email entirely on its own. Such as personal, business,
> non-delivery report, domo, auto-respond, marketing, transactional and
> spam.
>
>>
>> ARF says absolutely nothing about what either the report generator or
>> the report recipient is supposed to do with the report.  That's  
>> been one
>> of the primary design goals all along, because we can't predict  
>> how the
>> best practices will evolve.
>>
>>
> I can predict that this would be very useful.
>

If you already have an arrangement with the receiving ISP then they
will know (via DKIM or source IP based arrangement, or similar) that  
you're
claiming that this mail is transactional.

If their user claims the mail is spam, then they're still going to  
send you
a "spam" type ARF report. But there's no need for any additional
notification in either direction, as they already know that you claim
the mail is transactional. And if you and the ISP both know that you
claim the mail is transactional then including that information in the
ARF report will not add any information at all.

 From your description it sounds like you want have an ISP take
every ARF report of mail sent by you that a user complained about
and piggyback on top of it the (pretty much unrelated) data as to
whether the ISP thinks you've told them the mail is transactional or  
not.

There is nothing to stop you and an ISP coming to a private
arrangement to do that, but it doesn't sound like something that
would be of much value to most senders or recipients, so I don't
see any value in standardizing the procedure.

(I think there would be value in standardising the way that a sender
claims something about an email to an ISP, and the ISP confirms
that's the data they have, but piggybacking it on spam complaints
doesn't seem to be the place).

Cheers,
   Steve