Monitoring Your Mail Stream

One of the most important things for any mail sender to do is monitor their mail stream. There are a number of things that every mailer should pay attention to.  Some are things to monitor during delivery, some are things to monitor after delivery. All of these things tell senders important information about how their mail is being received by their recipients and the ISPs.

What to monitor during a send?

Many modern, commercial MTAs incorporate dashboard monitors into their systems so that senders can monitor sends as they are going out. Likewise, ESPs also have dashboards that give an overview of the progress of a send. These dashboards are a little different across installations but most will show you the active queue and the delayed queue. It can be useful to keep an eye on these values during a send.  are only two values that are useful to keep an eye on.

  • Active Queue size. The active queue is for messages received by the MTA and have not been sent yet. The number of mails in the active queue tells you how fast the send is progressing and how many messages have been sent.
  • Delayed Queue size. The number of mails in the delayed queue tells you how many messages have been sent but were temporarily refused by the receiving mail server (4xx response). Mails in the delayed queue may indicate some spam blocking or may be related to a problem on the receiving end.

What to monitor after a send

After a send there are a lot more things to monitor. Some of these things tell you about delivery and some of them tell you about how well a mail met user expectations and some tell you about what the ISPs are seeing from your send.

  • Feedback loop messages. The FBL messages tell you which of your recipients saw a message in their inbox and decided the mail was spam. For small volumes of mail this can be very noisy and not accurately represent how much users liked (or didn’t like) a particular mailing. Experience tells us, however, that for larger volumes of mail FBLs are an accurate measure of user reactions. High numbers of FBL messages, for mail with high inbox delivery, says that this was not something the users liked.
    • Pros
      • Accurate measure of user response for large numbers of recipients.
      • Tells you exactly what the user thought about the email.
      • One of the metrics ISPs use to make filtering decisions.
    • Cons
      • Very noisy for low volumes of mail.
      • Only some ISPs offer FBLs (see ISP Information for details)
      • Some ISPs do not share every FBL messages.
      • Need to know inbox delivery to be able to accurately determine complaint %.
  • Hotmail Smart Network Data Services. Microsoft provides senders with data about mail from specific IP addresses. Users need to sign in and authenticate the IP addresses, usually through clicking on a link to the registered owner of the IP range.
    • Pros
      • Gives number of spam trap hits per IP.
      • Gives number of complaints per IP.
    • Cons
      • Red/yellow/green color coding that sometimes doesn’t correlate with actual delivery.
      • No historic data.
  • Google Postmaster Service. The Google postmaster pages are new but provide data for an authenticated mail stream. It’s almost too new to give pros and cons.
    • Pros
      • Gives reputation data for IPs in high/medium/low/poor.
      • Gives reputation data for domains.
    • Cons
      • No API access.
      • No way to download data.
      • Some people sign up but never get any data and aren’t told why.
  • Open. Open rates are not an accurate measure of how many people actually read mail, as they require users to load images and not everyone loads images. However, they are a good trending measure and can be used to compare different mailings to the same group of people.
    • Pros
      • Easy data to get with modern sending infrastructure.
      • Every ESP will provide open data.
      • Trends can show how people react to mail over time.
    • Cons
      • Not always accurate.
      • Open rates can vary depending on target market.
  • Clicks. Tells you how many people were interested enough in your mail to click through.
    • Pros
      • This is what you want! You got people to click through and look at your website.
    • Cons
      • ISPs don’t actually measure clicks so don’t count towards ISP engagement data.
  • Unsubscribes. Unsubscribe rate tells you how many people so disliked your mail that they wanted no more mail from you. It’s a bit of an unconventional measure, but I like using it to see how recipients are responding to mail.
    • Pros
      • A form of negative feedback that doesn’t affect deliverability. ISPs don’t know when someone has unsubscribed using your link.
      • Useful as trending data.
    • Cons
      • Some MTAs don’t provide easy access to unsubscribe numbers.

There’s no such thing as a useless source of data. The more data and information you have, the better decisions you can make. Creating effective views of the above data sources can lead to much more direct and immediate feedback on how recipients are reacting to your mailings. This knowledge can drive more effective and productive emails in the future.

Related Posts

Barracuda filters clicking all links

Earlier this month I mentioned that a number of people were seeing issues with multiple links in emails being clicked by Barracuda filters. I invited readers to contact me and provide me with any information or evidence they had. Not only did a number of senders contact me, but one of the support reps at Barracuda also contacted me.
At issue is a part of the Barracuda email filter call the intent filter. There are 3 different modules to this filter.

Read More

Bad unsubscribe processes

We recently renewed our support contract with VMWare. It’s a weirdly complicated system, in that we can’t buy directly from VMWare, but have to buy through one of their resellers. In this case, we purchased the original hardware from Dell, so we renewed our contract through Dell.
Dell sends my email address over to VMWare as part of the transaction.
My only role in this is as CFO. I approve the purchase and pay the bill. I don’t do anything technical with the license.
The email failures start when VMWare decides that I need to receive mail about some user group meetings they’re holding all over the US. First off, I’m not the right person to be sending this mail to inside our company. I’m the billing contact, not the user contact. Then, they send me mail about meetings all over the US, when they know exactly where I’m located. Would it be so hard to do a semi-personalized version that highlighted the meetings in my local area then pointing out the other locations? Apparently, yes, it is so hard.
The biggest failures, though are in the unsubscribe process.
unsubscribe option
The unsubscribe page is no big deal. I get to unsub from all VMWare communications, and submit that request without having to figure out what my VMWare password is or anything.
After I hit submit, I’m taken to this page.
VMWareThank you
Wait? What?
“Thank you for registering?” I didn’t register! I don’t want you to contact me. Plus, this is a HP co-branded page when I’m not a customer of HP. VMWare knows this, they know they got my address from Dell.
The biggest problem is that I’m not sure that my address was actually unsubscribed. I suspect that someone copied a form from elsewhere on the site to use as an unsubscribe form. This person forgot to change the link after the “submit” button was clicked. But what else did they forget to change? Is the unsubscribe actually registered in the database?
I suppose only time will tell if VMWare actually processed my unsubscribe. If they didn’t they’re technically in violation of CAN SPAM.
The lesson, though, is someone should check unsubscribe forms. Someone in marketing should own the unsubscribe process, and that includes confirming that unsubscribe pages work well enough.

Read More

My holiday email prediction

I was on IRC with a group of ESP delivery specialists last week and one of them was looking for something to blog about. I suggested a list of holiday predictions. Not that I have a huge number of holiday predictions, but I did come up with one.
During the holiday season at least one retailer will decide that they have information so important that they will ignore my opt-0ut request and add me to their holiday blast list.
So what’s your holiday email prediction?
 

Read More