Question for my IT Peeps re: Web E-Blasting

That sounds like a good approach in general. If, instead of an e-blast, you send one e-mail every two seconds, Road Runner shouldn't identify you as spam, and you'd still get your "e-blast" out within a bit over an hour. If you have a list of contacts on your own computer, you might even be able to set it up that way yourself, without help from the Colorado guys.

Just to make sure I understand you: Am I correct in assuming that an "e-blast" consists of emails? Or is there something else involved in addition?

I don't know what is going on in your setup so here's my quick take on all that could be going on.

The heuristic (as opposed to blacklist) approaches to spam filtering usually use the following signals:

1) Email content. Things like having HTML (when was the last time you wrote a personal email in HTML?), links, certain words (free, viagra, humongous dong), and all can increase the spam score (remember, the automated approaches often don't just take one signal and say if it does this it is spam, they say if it does this add x amount of points and at x level we consider it spam).

2) User complaints. Many users will just use their inbox "mark as spam" feature instead of bothering to unsubscribe, even if you run an opt-in list and respect it. Many email providers will incorporate this data into their spam filters.

3) DNS configuration. Depending on how the domain is configured (Sender Policy Framework is the term to tell the geeks to make sure is not the issue) it may specify that only certain IP addresses can send email from that domain. Sometimes companies will do this and list their mail server but neglect to list their webserver, which may send out mails like in this case.

4) Email source. You may find yourself blacklisted not because you are but because you are sending email from a host that has seen spam from others or even to a destination URL that also hosts someone else who spams.

5) Last, but not least, the way the email is sent does matter. And yes hitting a host with volume at once is one thing that may certainly make a big difference. A key component to spam is volume, that's basically what defines it as spam as we wouldn't even have a name for this if it were just the rare targeted email.

You ask whether programmers can make the mail go out more intelligently and the answer is always a yes, given the right time and resources but it may actually be a daunting task on your platform and they modifications to make it mail more intelligently may not be in the scope of their job.

If you find you can't really work it out you may consider exporting the emails and using a separate mail solution that is more amenable to mail delivery needs.

Yes, Thomas, the e-blast is an e-mail, it just originates from the website, using the member registration list, and is processed through the site rather than through my e-mail program and IP address. I write it, save it and send it all through the admin area of the website. Easy peasy until the board president doesn't get his (because he uses Road Runner) and wants me to fix it.

Let me just start by denying that I have EVER sent an email with "humongous dong" anywhere in it, whether personal or professional. Though, now I'm a bit tempted.

The reason for non delivery that I get from RR with each returned e-mail is always that it was identified as SPAM due to the volume. Following the RR links and code explanations in the returned e-mails, I can confirm that as the reason.

I had considered that the company hosting our site also hosts hundreds (a thousand?? who knows) of sites that also can send e-blasts and since they all go through the same Colorado company's IP as I do, even without one of them sending Spam it would be enough volume across the continent that I would think all of us would be getting blocked by RR.

This has only been happening since sometime last summer. We've had the same web host and website for about six years. I'll ask about the DNS configuration / Sender Policy Framework, ( You're gonna make me sound so smart, Robert! ), but it's sounding to me like that would be the issue only if they changed it last summer, OR... if the Board president changed the return e-mail address to his own (which he did last summer) and sent the e-blast from his house (which he did). Would that mess up the identifiers? If RR system says "This looks like Spam... let me go backwards and see where this came from and if it is originating from the place that is on record" would it only go back to the Colorado host or would it see that it's coming from someone that is not my work IP or my return address?

Ahh, yes, if they are saying it's the volume then I'd be inclined to trust them.

If the software can't easily be modified to send in batches you might want to use another solution (ranging from newsletter software you can purchase for the server to third party email services) for sending out the emails.

Oops, I didn't get that last paragraph in my previous post prior to your response.

Do you know if him changing the return e-mail address and sending an e-blast from his house, and then changing it back to my e-mail as the return address (and doing this numerous times over a few months) would have any affect? I didn't know if that would appear as a hijack of sorts.

If you end up looking for a third party to manage your mailing lists and compose your mailing campaigns/newsletters, Constant Contact is an excellent one and very inexpensive and easy to use.

During the presidential election, I used them to keep in communication with about 5,000 members of several Obama groups. We sent out monthly newsletters, weekly updates and daily reminders

They require that you comply with the industry standards regarding spam and do a good job of educating you on what those are and how you can accomplish the same thing without being defined as "spam."

I researched about 5 or 6 companies and they were the best we could get for everything we wanted it to do and not have to pay much for the service. Many organizations and companies use them for their mass communications.

http://www.constantcontact.com/index.jsp

If you end up using them and need some help setting it up, I'd be willing to lend a hand.

If you are sending the E-mail as coming from your E-mail address, and the server sending the E-mail is not listed as one of the mail servers for your E-mail domain, it can lead to you being flagged as spam. (This is similar to, but slightly different from, the DNS/Sender Policy Framework stuff Robert was talking about.)

Sender Policy Framework means that you specifically tell the world what servers are allowed to send E-mail as coming from your domain.

Some E-mail systems are so paranoid that they even try to touch back to the sending server.

Thanks, Butterfly. I'll pass along that if this can't be worked out with RR, that is an option. They aren't much on spending money.

I'm not sure I understand. When RR returns the RR customer emails with the codes for why, part of the report indicates that the IP address they (RR) identified is the one in Colorado. They have never said anything about my IP address, just the one that hosts the site. The e-mail is done on the website and is sent from the website through the Colorado server IP address. So, wouldn't they be the only mail server for our e-mail domain?

Whether I compose it on the website and send it from the office IP address or if someone else does it from their home IP address, it comes back from RR as flagging the Colorado IP.

RR did say that if flagged, one of the things the automated system will scan is the contents of the e-mail to see if the return e-mail address is valid. The return e-mail address until last summer was my work e-mail address. When the board member has sent eblasts he has, several times, changed the return e-mail address to his own personal e-mail address which happens to be an RR account.

Is my work email address registered somewhere in Colorado or in the domain name registration as the official return address? Or, should that not matter?

If I want to send E-mail to "[email protected]", I must have a way of looking up where to deliver the message. You look it up using the "MX" record in DNS. The MX records point to the bogus.com mail servers, to which I then connect.

Once I connect, I have to tell the recipient mail system the destination E-mail address and the sender's E-mail address.

At this point, the recipient mail system knows two things. It knows the IP address from which I connected, and it knows the E-mail address which I claim is sending the E-mail.

The recipient mail system can at this point look up the MX record(s) for the sender's E-mail domain and see whether it matches the IP address from which the connection initiated.

It's kind of like caller ID. If someone calls your house claiming to be Bi-Polar Bear, but the caller ID shows a phone number which you don't recognize, you are right to be suspicious.

Ah, I get it. Thank you, DrewDad.