Some of those folks are clients or sales prospects but some of them are actually industry colleagues that have customers sending spam. In either case, I’ve been thinking a lot about best practices and what we all mean when we talk about best practices. In conversing with various people it’s clear that the term doesn’t mean what the speakers think it means.

For me, best practice means sending mail in a way that create happy and engaged recipients. There are a lot of details wrapped up in there, but all implementation choices stem from the answer to the question “what will make our customers happy.” But a lot of marketers, email and otherwise, don’t focus on what makes their recipients or targets happy.

In fact, for many people I talk to when they say “best practice” what they really mean is “send as much mail as recipients will tolerate.” This isn’t that surprising, the advertising and marketing industries survive by pushing things as far as the target will tolerate (emphasis added).

Just as it did over-the-air, those purveyors of television advertising will push the limits of toleration in their quest for profit. There’s nothing wrong with that, but let’s understand that in so doing, we’re pushing people who can now push back.

The unwanted messages theme is the same with mass media in print. A paper with nothing but ads is called a sale paper or advertising supplement. There’s only so far you can stretch the display advertising model before people begin to complain, too. Why? Because there’s no demand for unwanted messages.

The ability of people to push back is magnified in the email space. Recipients can push back against unwanted messages directly as individuals by using the this-is-spam button in their mail clients. They can block certain senders, they can filter mail out of their inbox. But even more than that, if many recipients push back against a particular sender, the ISPs notice. Their individual pushbacks are noticed the ISP acts to block or filter mail for all their users.

Best practices aren’t just about authentication, or personalization or any of the specific actions people are thinking about when they mention best practices. The term best practices is really shorthand for “don’t send spam.” Unfortunately, there are a lot of companies that send spam and still proclaim they’re following all best practices.

Really. Spam isn’t a best practice.

[Transactional mail is] an automated message, sent on a per-user basis, usually as the result of a direct action by the user or strongly associated with the user.

In Mike’s decision were things like sales receipts, opt-in notifications / welcome messages, social networking messages and the like.

Kelly disagreed with Mike and said she looked to the FTC and her current working definition was was mail that the user could not opt-out of.

Margaret pointed out that users should be able to unsubscribe from transactional mail.

Finally Steven posted this definition:

I think transactional email is anything which serves a functional purpose which is useful to the receiver. Shipping information is the classic example, invitations to buy are not.

I think these various comments demonstrate something I’ve been thinking for a while. Transactional email is one of the most interesting use cases of email. The traditional examples of transactional emails are those triggered by an action of the recipient; things like shipping notices and purchase receipts. Then there are transactional emails triggered by the actions of a friend of the recipient but that expect interaction or action by the recipient. Requests to connect through the various social networking sites or forward to a friend links on webpages fit into this category. Email as a command, subscription and unsubscription requests, are also a common category that fits clearly into most people’s definition of transactional email.

Then there is email as a notification. These notifications are often programatically generated and may or may not require action on the part of the recipient. Often, they are generated by monitoring scripts, and only fire when certain thresholds are crossed.

I think one of the important factors of a transactional email is that they are not generated by a person. If a mail is generated by a person then it is a one-to-one email. Of course, like most definitions, there are some exceptions, like some sales receipts are drafted and sent by hand not by an automated system.

What do you think? Are these types of messages transactional messages? Or is there some other term that would encompass this use of email?

I thought I’d ask readers here: How do you define transactional email? I’m interested in this both from the perspective of a sender and from the perspective of a receiver.

To better explain the vitriol, a little history of the two terms might help.

My personal recollection and experience is that the term “confirmed opt-in” was coined by posters in the newsgroup news.admin.net-abuse.email around 1997 or 1998. There was some discussion about marketers / spammers (a lot of the posters did not distinguish between the two) trying to use the term “double opt-in” instead of “confirmed opt-in.” Many posters believed (and many still do) that this was a deliberate attempt by marketers to make the process seem overly burdensome and unworkable.

During the 2003 FTC spam hearings, Rebecca Lieb shared formal definitions for 5 different subscription types including “Confirmed opt-in” and “double opt-in”. These definitions are still up on ClickZ.

Confirmed Opt-In
[...]confirmed opt-in lists confirm, by email, your subscription as soon as your name has been added to the list. They allow you to unsubscribe immediately by replying or clicking on a link within the email. [...]
Double Opt-In
A double opt-in list means not only must the user take an action to add himself to a list, but he then receives a confirmation of his subscription. He must reply to be added to the list. [...]

What we have here are two terms describing the same process and two different processes being described by the same term. Not only that, but the term describing two processes is also one of the terms terms used to describe the single process. I am confused just trying to describe the situation.

Adding drama to the confusion, there are some people who believe very strongly that marketers specifically published different definitions of confirmed opt-in to confuse discussions with anti-spammers. Whether or not this was a deliberate decision by marketers, the reality is that it has set the stage for years and years of confusion, obfuscation and controversy.

This is something I deal with on a regular basis. In order to make things clear with clients, ISPs and blocklists I cannot just use a term and be sure we all agree on what that term means. Instead, I have to define terms at the beginning of the conversation and make sure that everyone involved is using the same terminology.

So what happened on the URIBL list? Al answers that question at SpamResource.

Some random guy, nobody I know, he posted a request for help. He said, more or less: “Hey, blacklist XYZ has listed my double opt-in server. What should I do?”

Approximately 13 seconds after posting, he was verbally attacked in response. He was accused of being a spammer, and ridiculed, for daring to use such a term as “double opt-in.”

[...] a discussion forum made up of supposed thought leaders, people who actively work to stop spam, accused the guy of being a spammer. They didn’t accuse him of being a spammer because he sends spam — but instead, they called him a spammer because he used a term that they do not like.

There are anti-spam groups and people running blocklists who still treat the term “double opt-in” as a sort of reverse-shibboleth. People who use the term double opt-in, are not a member of the anti-spam community, and are unwelcome and shunned. Even those of us who have been around for many, many years are treated as outsiders and spammers if we accidentally use the wrong terminology in front of anti-spammers.

There is no solution I can see. The confusing terminology has been around long enough that there is no way to clarify things. Inventing new terminology is only going to increase confusion. Expecting either marketers or anti-spammers to abandon “their” terminology is patently ridiculous. The best those of us trying to deal with both groups can do is to be bilingual.

The eec made a really bad and ugly mistake but you can take my word for it that they have learned from it and that it will not happen again. I am not going to blog about this because I really do believe in the value of the EEC and what it brings to the industry. It’s okay to call out a mistake, but do you really need to destroy an organization that is so worthwile?

Just to be clear I had not heard of the EEC before this and when the story broke I blew it off as no big deal, some organization did something stupid and spammed. It was only after I did a little research that I realized this was THE organization that was supposed to be leading the pack in email marketing. They are

[...] a global professional organization that strives to enhance the image of email marketing and communications, while celebrating and actively advocating its critical importance in business, and its ROI value.

And, yet, they send mail that was perceived by many of their recipients as spam. While I have not seen copies of the mail, two posters commented that the mail did not comply with CAN SPAM. One of those said there was no opt-out link. Putting aside any of the permission or relevancy questions, if this is true then it takes it from a bad idea to illegal activity. How does this organization maintain any credibility as a leader in the email marketing space?

As for the negative comments, I fully expect that if Word to the Wise pulled something like this, there would be a lot of negativity and people holding us accountable for our actions. I do not see with the EEC should expect anything different from their base.

There was a funny comment from EEC Member pointing out that the EEC had brought us standardization of the spelling of email.

On my Email non-viable for acquisition post, Josh disagreed. He says

I think saying that “email is not viable for customer acquisition” might be too broad of a statement. I wouldn’t have any problem with “Purchasing lists is not viable for customer acquisition.”

I think his point is well taken. There are places where you buy a mailing, or buy an advertisement and that does drive acquisition as well as sales. I am still wary of using email for acquisition as most of the companies who come to me with that business model mean purchasing lists or co-reg when they say acquisition.

There have been a number of comments about Postini. Jay Levitt had a couple of comments that sum up the frustration that many of us have had with Postini.

I too tried to get a human at Postini. I took three different back-channel routes to get there. They all landed at the same person – apparently the one guy who sends out “we’re not responsible no matter what” form letters to anyone who writes to Postini. He told me, and I can’t make this up:

Postini was scoring my e-mails as “spammy” because Postini had previously scored my e-mails as spammy.

Dennis also commented about Postini:

I was told that if you take a document originally typed on an application such as MS Word and then copy and paste this into the marketing e-mail it gives it funky html code that for some reason gives your e-mail a lower score in Postini.

Cutting and pasting from MS Word has a myriad of problems, not just Postini delivery. One thing I emphasize with my clients is that their email structure must be clean and standards compliant. So many spammers out there are using badly formatted HTML mails, that the ISPs are looking at the technical structure of your email and using that as part of their filtering decisions. This confirmation from Postini only reinforces that.

Have a good weekend, everyone!

“We’re blocked at ISP-A, so we’re just going to stop mailing all our recipients at ISP-A.” Pure spammer speak. The speaker sees no value in any individual recipient, so instead of actually figuring out what about their mail is causing problems, they are going to drop 30% of their list. We talk a lot on this blog about relevancy and user experience. If a sender does not care about their email enough to invest a small amount of time into fixing a problem, then why should recipients care about the mail they are sending?

A better solution then just throwing away 30% of a list is to determine the underlying reasons for  delivery issues, and actually make adjustments to  address collection processes and  user experience. Build a sustainable, long term email marketing program that builds a loyal customer base.

“We have a new system to unsubscribe people immediately, but are concerned about implementing it due to database shrink.” First off, the law says that senders must stop mailing people that ask. Secondly, if people do not want email, they are not going to be an overall asset. They are likely to never purchase from the email, and they are very likely to hit the ‘this is spam’ button and lower the overall delivery rate of a list.

Let people unsubscribe. Users who do not want email from a sender are cruft. They lower the ROI for a list, they lower aggregate performance. Senders should not want unwilling or unhappy recipients on their list.

“We found out a lot of our addresses are at non-existent domains, so we want to correct the typos.” “Correcting” email addresses is an exercise in trying to read recipients minds. I seems intuitive that someone who typed yahooooo.com meant yahoo.com, or that hotmial.com meant hotmail.com, but there is no way to know for sure. There is also the possibility that the user is deliberately mistyping addresses to avoid getting mail from the sender. It could be that the user who mistyped their domain also mistyped their username. In any case, “fixing” the domain could result in a sender sending spam.

Data hygiene is critical, and any sender should be monitoring and checking the information input into their subscription forms. There are even services which offer real time monitoring of the data that is being entered into webforms. Once the data is in the database, though, senders should not arbitrarily change it.

It’s a bit different perspective, but very complimentary to my post yesterday about what is and is not spam. He and I have both heard from ISP people about how many requests for whitelisting or unblocking are prefaced with, “We comply with CAN SPAM” and how meaningless that statement really is. Al has a longer discussion of why.

The problem with this claim is that CAN SPAM lists the minimal standards an email must meet in order to avoid prosecution. CAN SPAM does not define what is spam, it only defines the things senders must do in order to not be violating the act. There is no legal definition of spam or of what is not spam.

To add to the confusion there are a number of confusing and contradictory definitions of spam. Definitions people have used over the years include:

• unsolicited bulk email
• unsolicited commercial email
• mail I don’t want
• mail I don’t think my customers want
• mail that is identical/similar to mail that hit my spamtrap
• mail that was sent to a non-existent address at my domain
• mail that contains HTML
• unsolicited email
• mail that advertises Viagra or porn sites or similar
• mail that other people send

I rarely use the word spam. There are so many different definitions of spam, I have no way to know if my clients understand what I am saying, so I avoid the term completely. I do think it is important for senders to understand the definitions of spam as used by entities responsible for filtering large amounts of incoming email.

Spamhaus and some other blocking lists use “unsolicited bulk email” as their definition. Generally, they have addresses that have never been used to sign up for email, and if a mailer sends mail to them, the mailer is sending unsolicited bulk email and is eligible for listing on the blocklist. The lists believe that if a mailer is sending one piece of email to a user who did not request it, then they are likely mailing many other users who did not request any mail. This definition centers around permission, and only sending email when you have the permission of the recipient.

Many of the large ISPs use “mail our users complain about” as their definition. With this definition, they do not have to argue permission status with a sender. The data shows that their customers complain about mail from that sender or with that URL. The ISPs are going to block, or deliver to the bulk folder, email that their users do not want.

Filters and some blocking lists use “mail that has characteristics of mail we know is unsolicited bulk mail” as their definition. These characteristics can be things like an invalid HELO string, or lack of reverse DNS on the connecting IP address, or badly formatted HTML. Mail that looks like spam, in the technical sense, is often treated like spam.

Resolving a block or listing requires first understanding the definition that entity is using. For blocklists senders usually must make changes to eliminate any possibility an address will get on the list without permission of the owner of that address. For ISPs, senders must decrease the complaints from users, usually accomplished by improving the signup process, getting a FBL from the ISP and and sending more relevant email. For filters, fixing the technical issues, cleaning up HTML and sending mail that does not look like spam will resolve many of the issues.

Complying with the law is not sufficient to meet the standards of recipients. If e360 is sending mail users are complaining about, then the recipient ISPs are going to treat the mail as spam and filter or block it. If e360 is sending mail to people who have not requested it, then posters in NANAE are going to claim e360 is spamming. Is e360 sending mail that complies with CAN SPAM? I expect that they are. Does this mean they are not spamming as defined by some people? Of course not.

E-mail delivers a return on investment so high, it’s practically embarrassing.

It doesn’t require getting fuzzy with the metrics.
But as long as we continue to call the percentage of graphics displayed in a given campaign its “open rate,” we’re being dishonest with everyone who doesn’t know what the metric truly means. And as the national “open rate” continues to drop, the lie gets even bigger.

I have to wholeheartedly agree with Ken here. “Opens” have always been something easy to measure, but hardly anyone actually understands what it means. Open does not mean that someone opened the email, open just means that an image in the email was loaded from the sender’s server. There are a lot of reasons an image might not get loaded even when the email is opened and read by the recipient. Some people, like me, choose not to load images by default. Some ISPs block images by default. Some companies block images. A very small fraction of people use mail clients that do not render images at all. All of these factors will affect open rates.

Measuring performance, real performance, of email marketing is important. Open rates are really not a measure of performance.

There are two ways that recipients give permission to receive newsletters or marketing email, explicit permission and implicit permission. Recipients give explicit permission to receive marketing email when they sign up for such email. Implicit permission covers situations where a user provides an email address, either during the course of a purchase, a download or other interaction with a company. There may be some language in the company’s privacy policy explaining that recipients may receive marketing email, but the recipient may not be aware they will receive email.

The easier situation is explicit permission. There are two basic ways a company can gather explicit permission to send marketing email: single opt-in and double (confirmed) opt-in.

Single opt-in: Recipient provides an email address to the sender for the express purpose of receiving marketing email.

Double (or confirmed) opt-in: Recipient provides an email address to the sender for the express purpose of receiving marketing email. The sender then sends an initial email to the recipient that requires a positive action on the part of the recipient (click a link, log into a web page or reply to the email) before the address is added to the sender’s list.

There can be problems with both types of opt-in, but barring fake or typoed email addresses being given to the sender, there is an social contract that the sender will send email to the recipient. I’ll talk about single and double opt-in in later posts.

Implicit permission covers a lot of situations where email is commonly sent in response to a recipient giving the sender and email address. In these cases, though, the recipient may not be aware they are consenting to receive email. This behavior may annoy recipients as well as causing delivery problems for the sender. Common cases of implicit permission include website registration, product purchase and free downloads.

More responsible companies often change implicit opt-in to explicit opt-in. They do this by making it clear to users that they are agreeing to receive email at the point where the user gives the company an email address. Not only is the information about how email addresses will be used in the company’s privacy policy, but there is a clear and conspicuous notice at the point where the user must provide their email address. The recipient knows what the sender will do with the email address and is given the opportunity to express their preferences. If users do agree to receive email, the company will send a message to that recipient with relevant information about how their email address will be used, how often they will receive email and how they can opt-out.

Explicit opt-in is the best practice for building a list, however, there are still companies that successfully use implicit opt-in to build marketing lists. Companies successfully using implicit opt in usually are collecting emails as part of a sales transaction. There is very little incentive for their customers to give them an email address not belonging to the customer.

Outside of purchasers, however, implicit opt-in leaves a company open to getting email addresses that do not actually belong to the person providing the company with the email address. This most often occurs when the sender is providing some service, be it software downloads, music or access to content, in return for a “payment” of a valid email address. In order to protect against users inputting other, valid addresses into the form, the sender must verify that the address actually belongs to their user before sending any sort of marketing email. The easiest way for senders to do this is to send a link to the recipient email. This link can be the download link, or the password to get to restricted content. Because the recipient must be able to receive and act on email, the only addresses the sender has belong to actual users of the site.

In some rare cases, implicit opt-in can be used to build a list that performs well. However, senders must be aware of the risks of annoying their customer base and the recipient ISPs. Mitigating these risks can be done, but it often takes more effort than just using explicit opt-in in the first place.