Recent Posts

Suing spammers

I’m off to MAAWG next week and seem to have had barely enough time to breathe lately, much less blog. I have a half written post, but it’s taking a little more research to put together. That can wait until I get the chance to do the research.
Instead I thought I’d talk about the North Coast Journal article “The Rise and Fall of a Spam Crusader.” It’s quite an interesting article and looks into the personal and business sacrifices that people make in order to chase down spammers.
In my experience a lot of the serial litigators have very poor practices around data collection and analysis. They don’t collect evidence, they just collect email and then make assertions and assumptions. This not every effective when having to convince a judge that you are right.
The article actually does nothing to change this impression. The cases ASIS won are the cases where the defendants didn’t respond. That also means that ASIS couldn’t collect.
I do disagree with Mr. Singleton, the lawyer, where he says CAN SPAM is dead. In many cases I’ve seen there aren’t clear CAN SPAM violations. So if he’s trying to sue these spammers under CAN SPAM his cause of action is wrong. Secondly, the article goes on to talk about the broader implications.

Blasting the message!

Sending frequency is an important part of any email campaign. Too little mail and recipients forget about the mail and don’t open it when it does arrive. Too much mail and folks start complaining, like John Cole over at Balloon Juice.

Know your target audience…

… and the device they’re probably going to read your email on.
@lauter from MailChimp points and laughs at an advertising email from Blackberry-the-company that’s completely unreadable when read on Blackberry-the-device.

That’s really bad marketing on a bunch of different levels.

Beware the TINS Army

When consulting with clients, I spend a lot of time trying to help them better understand the concept of sender reputation. Spam reports, feedback loops, and other data that comes from a collection of positive and negative reputational feedback about a company sending email.
Certainly, the “This is not spam” action – moving an email from the spam folder to the inbox, or clicking the “not spam” button in a web mail’s interface, is a strong positive reputational action. Some webmail providers use this data to decide which bulked senders deserve being let out of the penalty box – which should have their mail once again delivered to the inbox.
A client recently theorized that a great solution to their delivery problems would be to do this “en masse.” Sign up for hundreds or thousands of webmail accounts, send my mail to them, and click on the “not spam” button for each of my own emails. That’ll greatly improve my sending reputation, right?
NO! ISPs have already thought of this. They watch for this. They’re really good at picking up on things like this. I know for a fact that Yahoo and Hotmail and AOL notice stuff like this, and I strongly suspect other webmail providers notice it as well.
What happens when Yahoo or Hotmail pick up on this type of unwanted activity? Well, if it’s at Yahoo, they’re likely to block all mail from you, 100%, forever. I’ve seen it happen more than once. Yahoo might even identify all of your netblocks, ones beyond the ones sending today’s mail or originating today’s activity. And good luck trying to convince them that you’re not a spammer – you have a better chance of winning the lottery two weeks in a row.
As for Hotmail – what would Hotmail do? Ask Boris Mizhen. Microsoft is currently suing him, alleging that he and/or his agents or associates engaged in this very practice.

Spam isn't a best practice

I’m hearing a lot of claims about best practices recently and I’m wondering what people really mean by the term. All too often people tell me that they comply with “all best practices” followed by a list of things they do that are clearly not best practices.
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).

Does your signup pass muster?

On Eric Goldman’s blog, Venkat discusses a recent fifth circuit decision about an online signup process and what the court will look at when considering a claim that a user didn’t read an online disclaimer.

Zombie Apocalypse

I hope my series on zombie addresses has convinced you that there are zombie addresses on your list and that you should be concerned about the effect they have on delivery and metrics. Today I’d like to talk about what you can do to get rid of zombie addresses without affecting too many actual subscribers.

One thing that many companies struggle with while dealing with zombie addresses is letting go of addresses. They are so tied up in the idea that a bigger list is better that they can’t let them go. Even if a particular address has not had any activity in 18 or 24 months, they insist that they can’t give it up, it might come back and the customer might make a giant purchase. No. It’s a zombie. It’s not coming back, except to eat your brains.
The first step to dealing with zombies is to acknowledge their existence. They are there, they are on your lists and they are dirtying up your lists. Pretending they’re not there does not make them go away. They are zombies. In no case is there a human inside. There is no potential sale lurking, waiting to jump out and act on that perfectly crafted offer.
The second thing to remember is that the humans that used to have the zombie addresses found you once and they are still interested in what you’re offering then they will find you again. They may even already be back on your list with their new email address.
While you can’t identify zombie addresses specifically, you can identify addresses that act like zombie addresses. These are addresses that have no activity over a long period of time, more than 12 months. For these addresses that haven’t had activity in 12 – 18 – 24 months, you want to confirm with the recipient that they are there and want to continue to receive mail from you.
The best way to notify them is to send an email asking if they want to remain on your list. If they fail to act, you will remove them from future mailings. Short, sweet and will let you drop off zombie addresses without much effort on your part.
I know, I know, you aren’t ready to let go so fast. After all, some people have come back after 24 months and made a purchase from the perfect offer. They’re not dead yet! OK. But you can’t get a response from them through email. They just don’t care enough about what you’re sending. That’s when you contact them through another channel.
For instance, if the email address is tied to a web account, say a social networking site or bank account or a web forum, you can also contact the user through your website. Next time they log in, send them a message that says their email address has been removed due to inactivity, but if they want to reactivate they can do so at the subscriber preference center or profile page. When they do, send them an email to confirm that this is the address where they want to receive mail. At this point you can give them a link or a magic cookie to past into the website to verify the address.
Or if you’re a bigger retailer you can send alerts to your customer service staff, so when the account holder contacts you by phone with a question or an order you can get an updated email address. If you have a loyalty program, have an alert come up at the point of sale and the clerk can ask for an updated email address.
I even know one company that would send postcards to their zombie accounts in an effort to re-engage them and get an active email address from them.
If the person never comes back, if they don’t ever interact with your business again, if none of the channels work to contact them and update the address then it really is best to just let the relationship go. It may not be you, or anything you’ve done. People move on, their interests change and that’s part of life. They may have moved outside of your service area, or they may have joined your list for a specific product that they don’t need or you don’t sell. They may have died and turned into a real zombie. In any case, they are not a viable prospect for your mail.
Email addresses and business relationships are not forever. Letting zombie addresses go is important for the health of any email marketing program.

Zombie email: Part 3

Last week, in Zombie email: part 1 and part 2 I talked a little about the history of email addresses and how changes in the ISP industry in the early to mid 2000’s brought about the rise of zombie email addresses. Today we’ll look at the effect zombie addresses have on email stats and why ISPs are starting to monitor zombie addresses.
A zombie address, despite the fervent belief of some email marketers, doesn’t come back to life. The person who initially registered that address has decided to stop using that email address. The defining factor of a zombie address is that there isn’t now and won’t be anyone in the future reading email sent to that address. There is no human there to read or react to any email sent to that address.
A zombie address does not represent an actual recipient, they’re just remnants of a recipient that once was present.
Having a list containing any significant number of zombie addresses can throw off metrics enough to mislead a sender about the effectiveness of their email marketing program. Sometimes, the zombie addresses make the metrics look worse, sometimes they make metrics look better. In either case, the metrics don’t accurately represent the performance of a marketing program.
Zombie email addresses do bulk out a mailing list, making lists look bigger. They’re not real addresses, so they don’t reflect quality, but they do impress marketers that think bigger is always better. But, in reality, you may as well add thousands of addresses at non-existent domains for the real value these addresses bring to your list.
Zombie email addresses on a list depresses any metric that use “number of emails sent” or “number of emails accepted” as a denominator. If 10% of a list is zombie addresses, then an open rate reported as 15% will actually be an open rate of 16.7%. The more zombie addresses on a list, the more the statistics will be depressed.
In addition to having lower open rates, lists with more zombie addresses also have a lower complaint rate. In fact, in the recent past spammers have padded their lists with zombie addresses as a way to artificially lower their complaint rates.
Spammers using addresses created just to bulk up the denominator and lower complaint rates have led ISPs to start monitoring the types of addresses on a particular list. I first heard about ISPs looking at recipient profiles at a meeting in 2006, so it is not, in any way, a new technique for ISPs. What is new is the number of zombie addresses on legitimate, well maintained lists, and the fact that they are present in high enough volume to affect reputation and delivery.
ISPs use zombie addresses to monitor the reputation of a sender because it is a more accurate way to measure what the recipients think about an email and that sender. Senders ignore zombie addresses because they make some stats look bigger (total list size) and better (lower complaint rates). Many senders also believe that addresses come back to life, despite all evidence to the contrary, and will not purge an address for any reason other than it bounces. They’d rather live with inaccurate and misleading metrics than removing non-performing addresses.
Tomorrow, in the final post of this series, we’ll examine how senders can identify potential zombie addresses and what steps they can take protect themselves from the negative reputation hit from zombie addresses. (Zombie Apocalypse)

Zombie email: Part 2

In zombie email: part 1 I talked about how email addresses were tightly tied to internet access in the very early years of the internet. We didn’t have to worry about zombie email addresses because when an account was shut down, or ignored for a long time then mail would start bouncing and a sender could stop sending to that account.
There were two major changes to email accounts in the early 2000’s that led to the rise of zombie emails.
People started decoupling their internet access from their email addresses. Free addresses were easy to get and could be checked from everywhere. No longer did they have to dial in to get email, they could access it from outside the office and outside the home. Mobile devices, including the first generation of smart phones and laptops, helped drive people to use email addresses that they could access from any network. The easy access to free mail accounts and the permanence led people to adopt those addresses as their primary address.
When people changed addresses, for whatever reason, they didn’t have to stop paying. There was no way to tell the free ISPs to stop accepting mail for that address. Free mail providers would let addresses linger for months or years after the user had stopped logging in. Sometimes those addresses would fill up and start bouncing email, but they were not often turned off by the ISPs.
The lack of purging of abandoned addresses was the start of dead addresses accumulating on mailing lists. But there weren’t that many addresses in this state, and eventually they would fill up with mail. When they were full the ISP would stop accepting new mail for that account, and the address would bounce off a mailing list.
Everything changed with the entrance of Gmail onto the scene. When Gmail launched in 2004 they were providing a whole GB of storage for email accounts a totally unheard of storage capacity. Within a year they were providing multiple gigabytes of storage. Other freemail systems followed Gmail’s lead and now all free accounts have nearly unlimited storage. Plus, any mail in the spam folder was purged after a few weeks and bulk mail doesn’t count against the users’ storage quota. Now, an abandoned email account will almost never fill up thus senders can’t use over quota bounces to identify abandoned accounts.
Now we’re stuck in a situation where SMTP replies can’t be used to identify that there is no one home inside a particular email account. Senders can’t distinguish between a quiet subscriber and an abandoned address. ISPs, however, can and are using zombie addresses as a measure of a senders reputation.
On Monday we’ll talk about why and how zombie addresses can affect delivery. (Zombie emails: part 3)
Tuesday, we’ll talk about strategies to protect your list from being taken over by zombies. (Zombie Apocalypse)

Zombie email: Part 1

Zombie email addresses: those email addresses that never really die, eat your brains and destroy your email delivery. To understand zombie addresses and why they’re just now becoming a problem, we really need to understand some of the history of email addresses.
In the early days of the net, people got an email address usually associated directly with their access to the Internet. Many of them ended with .edu or .gov. I even had one that ended in .BITNET for a while. The first ISPs followed this convention. Users signed up for an account at a local dialup and were assigned an email address, and that was their email address. It wasn’t until the late 1990’s where there was widespread access to multiple email addresses.
What this means is that when people left a job, or canceled their Internet access their email address went away. Addresses that were abandoned would, after a short period of time, start bouncing back with user unknown, giving everyone the opportunity to stop mailing that account.
Even with the advent of multiple addresses for a single account and the easy availability of free addresses from places like Hotmail addresses that had been abandoned would still bounce off a list. Why? Because accounts had limited storage. My first dialup account had, I think, 10MB of space. It may have been as much as 20MB, but it wasn’t very much. Accounts receiving a lot of mail that weren’t checked frequently would fill up and start bouncing mail. Senders would be able to remove abandoned accounts because they were full.
Tomorrow we’ll talk about two things happened in the early 2000’s that changed email and led to the rise of zombie email.
Zombie Email: Part 2
Zombie Email: Part 3
Zombie Apocalypse