Broadcast Email Is Unreliable

[aebrown]

Why would spam filtering cause this kind of random failure? I don't know much about e-mail, but I do not believe that the website e-mail broadcast mechanism attempts to validate e-mail accounts before broadcasting. It might be an interesting question to explore though. That is the only scenario where I would agree that the website admin's e-mail domain might be an issue.

In many cases, spam filtering by the ISP involves several more variables than just the text of the message. Obviously certain text can be considered to be spam ("need help moving 23.6 Million Dollars from Nigeria"). But there are many other factors:

• Similarity to other messages that have been flagged as spam
• Sending a message to a large number of recipients
• Sending from an account different from the sending domain
• Black lists, white lists, and gray lists for sending domain

Typically, the various potential violations are given ratings and weights and so a message ends up with a final score that determines if it is considered spam. Since there are so many factors involved in the rating, and they can even depend on other traffic through your ISP, it can seem somewhat random. It can even be the case that the exact same message gets filtered one day and sent through just fine on another.

[doughy]

This is not a spam filter issue. In my case, ALL of the emails either go through, or don't. If it were a spam filter issue, certain individuals would receive the email, and others wouldn't. I'm on gmail, same as Techgy, and our results are different.

There is a bug in the email broadcast code. Has anyone actually looked into it from the church IT department? We need one of the developers to take it head on and investigate.

[techgy]

This is not a spam filter issue. In my case, ALL of the emails either go through, or don't. If it were a spam filter issue, certain individuals would receive the email, and others wouldn't. I'm on gmail, same as Techgy, and our results are different.

There is a bug in the email broadcast code. Has anyone actually looked into it from the church IT department? We need one of the developers to take it head on and investigate.

As I had mentioned in my earlier post, it's possible that some of the problems may be attributed to a blocking issue. Obviously, in your situation, that's not the case, but until you had responded with a domain, that question was still hanging.

[dmaynes]

Typically, the various potential violations are given ratings and weights and so a message ends up with a final score that determines if it is considered spam. Since there are so many factors involved in the rating, and they can even depend on other traffic through your ISP, it can seem somewhat random. It can even be the case that the exact same message gets filtered one day and sent through just fine on another.

Why would spam filtering prevent e-mails from being sent? I thought that spam filtering only related to incoming e-mails? Also, there is no way that e-mails sent from the ward website actually result in connecting to my Exchange server in order to send the e-mails. So the sending of the e-mails does not involve the actual e-mail domain server where I have my account.

I have confirmed with other members that only one e-mail is being received by them, even though I have multiple attempts.

[aebrown]

Why would spam filtering prevent e-mails from being sent? I thought that spam filtering only related to incoming e-mails? Also, there is no way that e-mails sent from the ward website actually result in connecting to my Exchange server in order to send the e-mails. So the sending of the e-mails does not involve the actual e-mail domain server where I have my account.

I have confirmed with other members that only one e-mail is being received by them, even though I have multiple attempts.

You're right that your domain is not involved in the sending of the mail, but there is some mail server involved at some level in the sending. I don't know just what mail servers the Church may use.

However, it is quite common for spam filtering to be done at the sending level as well as the receiving level. ISPs and other mail servers don't want to be involved in the sending of spam, so they apply algorithms to block outgoing messages that seem to be spam.

In such cases, a bounce message would be sent. However, since you are not the actual sender, you would not receive the bounce -- some Church server would. So if this is a factor, it should be a source of informaiton to help the Church track down what is happening.

[dmaynes]

In such cases, a bounce message would be sent. However, since you are not the actual sender, you would not receive the bounce -- some Church server would. So if this is a factor, it should be a source of informaiton to help the Church track down what is happening.

I would be willing to help track this down if we could set up a dummy ward website. I don't think it would be good to send out lots of test broadcast messages to my ward members.

Based on your explanation, it seems reasonable that a spam filter on an e-mail server for the LUWS is acting in an undesirable way.

I have also wondered if the e-mail length was an issue.

I can try testing with a bogus e-mail address and domain to see if that results in a failure.

[techgy]

I would be willing to help track this down if we could set up a dummy ward website. I don't think it would be good to send out lots of test broadcast messages to my ward members.

Based on your explanation, it seems reasonable that a spam filter on an e-mail server for the LUWS is acting in an undesirable way.

I have also wondered if the e-mail length was an issue.

I can try testing with a bogus e-mail address and domain to see if that results in a failure.

Your suggestion regarding the length of an Email is a valid consideration.
I've found from personal experience that the shorter the message the greater possibility I have of it being considered as junk mail. The wording is also important.

[lajackson]

Perhaps part of the problem is that the message is sent from a Church server, but the message says it is from my ISP (gmail, in my case). Since the message is not really coming from where it says it is coming from, many mail clients get very nervous. I know mine does. I find many valid e-mails in the spam folder.

[AileneRHerrick]

Your suggestion regarding the length of an Email is a valid consideration.
I've found from personal experience that the shorter the message the greater possibility I have of it being considered as junk mail. The wording is also important.

I've found that if I make my messages too long, they don't get sent through at all! But I guess that's an entirely different matter. The only time I had problems with a report of spam is with a new News and Information item (and it contained a link).

[dmaynes]

You're right that your domain is not involved in the sending of the mail, but there is some mail server involved at some level in the sending. I don't know just what mail servers the Church may use.

However, it is quite common for spam filtering to be done at the sending level as well as the receiving level. ISPs and other mail servers don't want to be involved in the sending of spam, so they apply algorithms to block outgoing messages that seem to be spam.

In such cases, a bounce message would be sent. However, since you are not the actual sender, you would not receive the bounce -- some Church server would. So if this is a factor, it should be a source of informaiton to help the Church track down what is happening.

I send out a weekly e-mail on Wednesday morning. This morning's effort required 6 attempts to send the e-mail to all the ward before it was sent and delivered (1 success in 6 attempts). Before each general ward broadcast attempt, I sent an individual e-mail to myself. All of those e-mails were received (6 successes in 6 attempts). I have not seen any e-mail delivery/send failures for individual e-mails. They always come through, even though the e-mail content is exactly the same.

There is something very different about the general broadcasts and the individual e-mails. With 12 attempts and 5 failures, the probability that all 5 failures would come from the general broadcast and none would occur with the individual e-mails, if they are the same procedure, is less than one in one hundred.

I have bounce-back (undeliverable messages) receipts coming back from two Church e-mail servers: mail1.gmhwh.org and mail6.gmhwh.org.

There is no indication of spam filtering (I'm not sure that I would see any).

If the issue is spam filtering, why would the filtering be applied for general broadcasts, but not individual e-mails?

Also, the server that is handling the general broadcast is different than the server handling the individual e-mails, because the time stamp on the sent messages for general broadcasts is one hour earlier than the current time. And the time stamp on the sent messages for the individual e-mails is the current time. So the time on the e-mail server handling the general broadcasts is wrong.

I'm not getting the e-mail verification messages. I just think the server is dropping the request to send the broadcast, or some software packing/parsing bug is causing the request to not be created. That is the simplest explanation.

Thanks,
Dennis