Mean What You Say

While speaking with clients there is a lot of terminology (see bold items) that seems to be thrown out to describe things other than the email industry meaning. It is important to make sure when discussing deliverability strategies and problems all parties are speaking the same language. I’m going to try to help remove any confusion surrounding some of those terms here.

One area that seems problematic are bounces. The generic term bounce refers to any message that didn’t reach its intended party due to any error from the receiving server. Some of these errors are permanent errors (hard bounces) such as “unknown user” while others are transient errors (soft bounces) like “mailbox full”. Companies often set these addresses to a held status when they reach certain bounce thresholds. When you then look at your deliverability rate, you are looking at the number of subscribers not in held status who did not bounce. Also commonly used (and likely a better and more telling metric) is to calculate deliverability rate based on delivered to inbox rather than merely did not bounce, but neither should include addresses that are flagged as held prior to the send and not actually sent an email. It should be noted that it is possible to have a large held file and very good deliverability.

Another greatly misunderstood item is the confirmation mail. A confirmation mail has only one definition in the world of email. It is an email sent in response to a potential subscriber having opted in to receive your emails which requires the user to take an action (usually follow a link) to remain a subscriber. There is no other definition of this term. Probes (emails sent to test deliverability to an email address typically sent after a certain number of bounces) can be sent to addresses that are suspected to be deliverable, but you do not want to reconfirm and often are either in danger of being held or already have been held. Both of these are types of transactional mail, email sent in response to a business request (invoices, renewal notices, delivery notices, etc.)

There are a few different kinds of blacklists (aka blocklists), lists of domain names or IP addresses that meet a set of criteria specific to that list and are often used in filtering email. A DNSBL is a blacklist that lists IP addresses (the numeric address that represents where your email is coming from in a format similar to 199.1.145.68). An RHSBL is a blacklist that lists domain names (the alpha-numeric address that represents where your mail is coming from in a format similar to bmail1.digital.halldata.com). The third type of blacklist is a URIBL/SURBL. These list domain names of links contained within the emails you send.

The industry does have other words that it is trying to change, remove, or create. Today an article was posted on Email Insider blog talking about the dangers of continuing to use such terms as blast, broadcast, and bulk in the current industry environment. It was arguing that mailers continuing to use these terms (and the practices associated with them) were hurting not only themselves, but the entire email industry. Recently there was an article in the Email Experience Council blog recommending that the hyphen be removed from the term e-mail. New terms are springing up as well. For instance bacn has been getting a lot of press lately. Bacn is basically being used to describe mails that are not quite spam and not quite personal email. By consistently using terms we can have an impact on the perception of email both within the industry and from the outside.

What your subscribers are telling you

When sending out your emails, it is important to realize that you aren’t the only person judging the success of your campaign. The ISPs and recipients are also making their own judgment calls. Some of the feedback that your recipients give you is positive in the form of clicks and opens. There are other types of feedback though, and these come in the form of unsubscribes, complaints, and bounces. Most of the major ISPs have feedback loops available for senders to monitor these other types of feedback.

Feedback loops are essentially a mechanism that ISPs use to let senders know when a subscriber has complained. They come in a variety of formats. Some contain no more information than the email address of the complainer, others contain the entire mail with complete header information, and still others scrub any identifying information from the headers of emails to ensure that you don’t engage in list washing. Hallmark Data Systems works to ensure that all available feedback loops are in place on your IPs. We have established feedback loops with AOL, Yahoo, Hotmail/MSN, usa.net, Excite, and United Online.

Every list is going to have a few complaints. Some people complain merely as a means of unsubscribing, but most are actually using it as a means to let you know that something has happened that they don’t like. They could be telling you that the unsubscribe mechanism is too difficult to use or broken. According to a recent Return Path study 33% of users try to use the unsubscribe link first and 12% only complain as a last resort of unsuccessful unsubscribe attempts. They could also be telling you that they are receiving more email than they expected. It should be clear when subscribers opt in that they know how often they will be hearing from you.

It's a good idea to look at patterns in your complaints, since this can provide valuable information. Complaints can spawn from poor acquisition practices, irrelevant content, difficult, strict, or broken unsub mechanisms, or sending emails too frequently. It is important to track down the sources of complaints. Complaints leave a bad taste in the mouth of the complainer (and in the ear of anyone they tell) that tarnishes your brand, as well as diminish the size and effectiveness of your list. Also, when your complaint rates are elevated, your emails are more likely to be blocked. By looking for patterns in your complainers, you gain insight into the sources of those complaints and ultimately can stop the complaints before they start. Any data points that see high percentages of complainers, whether it is specific content, address acquisition source, frequency, etc., can tell you about potential problems you may need to address. If it is a problem with frequency or acquisition source, it may be as simple as a need to update the opt in language to set subscriber expectations. Hallmark Data Systems can make reports available to our clients that address many of these issues as well as others. For further information about these or other reports please contact your account manager.

Get off to a good start

Many spam messages come from IPs that have never sent email before. Machines become infected with viruses and start sending emails. It should come as no surprise that when senders change their IP addresses, ISPs and inbox providers look at these new IP addresses with suspicion. Some major ISPs throttle the amount of mail that they accept from fresh IP addresses and others filter them more harshly until they can establish a good reputation. When changing an ESP or simply adding more IP addresses to your mail server there are a few things one should keep in mind.

1) Let your readers know

In the world of B2B email, many of your recipients are in corporate domains. By adding some simple text to your email, you can ask your recipients to have any corporate white listings updated to reflect your new sending IP. If they haven’t asked their networking team to white list your domain in the past, this is a good opportunity to give them a reminder that they might want to white list you so that they won’t miss any of your content. It is also possible that these new IP addresses will suddenly be sending to domains that were blocking you before. This reminder allows your readers to help make certain that they continue receiving your emails and serves as a reminder that they may have signed up in the past and hadn’t been receiving your emails from some time. This message should be added to emails both before and after the move to maximize their impact. Because the recipient may not have been receiving your emails it is important to put in each email the date that they signed up for your list as well as which list it was.

2) Start small

Sending large volumes from a fresh IP space is a surefire way to hit blocks. You should slowly ramp up your email program so that throttles won’t keep your mail out. It should probably take you a few weeks to a couple of months to send the same volumes that you did before, especially if you are moving from an IP space that was severely blocked. Especially in these early stages you should aggressively scrub any hard bounces and spam complainers so that the rates quickly reduce. Also pay close attention to spam trap hits. If the lists haven’t been scrubbed vigorously you may need to reconfirm at least portions of your list. By starting small, you can target which lists have the problems and minimize the negative impact that these new sends may create.

3) Put your best foot forward

Before sending email you should always be cognizant of the fact that what you are sending can have a positive or negative impact on both your IP reputation and your brand reputation. This is especially important in the first few months of sending. Relevant content sent at the frequency your subscribers want will help keep spam complaints down. Your readers are your first line of defense when it comes to deliverability. If they want your content and aren’t complaining about it, then the servers that host them will be more likely to work with you to make sure that your subscribers get the message. If you do happen to set off sirens by sending from a new IP it is also important that the emails being seen by the network administrators and anti-spam teams see emails arriving that don’t look anything like spam.

All of these things combined will have positive effects on your reputation in the long term. By keeping complaint rates and unknown user rates low you increase the chance of white listing your domains at the major ISPs. You stand a chance of keeping corporate white lists that are already in place, and potentially gaining some new white listings at corporations that may have blocked you before. Most importantly though, they will help keep your mails from getting blocked right out of the gate.

White on White: A review of Silverpop's latest white paper on white listing

Over the last few months I have been talking with some of our clients about white listing. Silverpop, an email service provider, released a whitepaper today entitled Unlocking the Secret World of White Listing. This document outlines what senders need to know about white listing and verifies many of the points that I have been making to our clients.

When ISPs are considering white listing you they are looking at a variety of items. Bounce rates, complaint rates, volume, and authentication/reputation are all considered when applying for a white list, and ISPs look at your previous sending patterns to determine what they expect those to be.

A few highlights from the whitepaper (which I encourage everyone to read):

“Keep in mind that the topic of complaints is very sensitive for ISPs. After all, those unhappy people are their customers.”

“Some ISPs keep records of addresses that have bounced, and if you continue to send to addresses they’ve told you are no good, you not only risk any chance of being white listed but also endanger the future deliverability of your messages.”

At the end of the day, this is why it is imperative to adhere to best practices when sending. For a comprehensive listing of what ISPs expect senders to adhere to, please read MAAWG Sender Best Communications Practices - Version 1.1.

Well Done Is Better Than Well Said

The primary driver of deliverability in today’s environment isn’t content, blacklistings, or even as simple as spam complaints. The most important aspect of deliverability is reputation. Senders can have good reputations or bad reputations and the impact of either can be rather severe in terms of deliverability. This reputation isn’t based on your company name, business model, customer base, or anything outside what you are sending to email subscribers. The old adage of actions speak louder than words couldn’t be more true when it comes to this.

What is reputation –

Reputation in terms of deliverability is made up of many metrics, and the way that it is determined varies widely, however the general composition seems to have a few consistent players. ISPs and some corporate domains maintain a repository of all of the mails that you send and a lot of data that they see about those mailings and tie those back to the specific IP that is sending that mail:

Spam complaints – Spam complaint rates typically are the biggest determining factor when it comes to your reputation. The definition of spam is very fluid, but the most important one is what it means to the people that control whether or not your mails reach their destination. To the people that control the flow of your mail, spam is being defined as any mail that a recipient doesn’t want. It doesn’t matter to them if it was asked for or not (although opt-in information can help get blocks removed in these scenarios). High complaint rates lead to a bad reputation while low complaint rates will lead to a high reputation.

Unknown user rates – These are percentages of bad addresses that you send to servers. If you don’t work at actively scrubbing bounced addresses from your list or reactivate held addresses when changing email service providers, you are likely to have high unknown user rates. This basically tips off the ISP or corporate domain that your list may consist of people that didn’t ask for your mails and very much resembles spammer behavior.

Spam traps – There are 2 types of spam trap addresses. There are addresses that are planted on websites and have never subscribed to anything nor are they active email addresses. The other type of spam trap is an address that was formerly a valid address that has since been abandoned. When ISPs see email being sent to spam trap addresses it is considered spam de facto. The type of trap, number of hits, and type of mail being sent to the trap all influence the impact that these have on reputation.

Sending infrastructure – This consists of all of the technical aspects of email. rDNS must match forward DNS. Are you using authentication and is it properly set up? Always abide by the preferred connection settings of the receiving mail server. These are a few of the items looked at under this category. At Hallmark Data Systems, we ensure that all of these things are in place for every client.

Sending permanence – This is essentially how your emails are sent. How often do you send and to how many people? It is also about how consistent your emailing patterns are. They are basically comparing your sending strategies to those of spammers.

How reputation impacts deliverability –

This varies widely from domain to domain, but the consistent factor is if your reputation is lower than they accept, your mails will most likely be filtered or blocked. Each domain weights the categories differently, and some domains consider items not listed here while others don’t monitor all of the ones listed. According to a study recent Return Path study, only 17% of emails are filtered due to content. At Return Path's Email Marketers Leadership Forum in Chicago, George Bilbrey from Return Path said for B2B it is closer to 25%, but still most mails are filtered based on reputation, not content.

For a basic idea of what your reputation might look like your best bet is to track the feedback you get when you send a mailing. Know what percentage of your list hard bounces and complains. Bear in mind that ISPs calculate these metrics largely on the basis of inbox delivery and ignore both bounces and items in the spam/junk folder, so any metrics you see are generally lower than what they are calculating.