In a recent conversation with an anti-spam advocate over the best way to stop unwanted messages, he suggested to me that all the challenges in his industry and mine would go away if emailers would only double opt-in all their recipients.

It has been a while since I've heard this argument, but it's not surprising that it came up again. It's a very logical point of view, and it would seem almost foolproof. The problem is that it doesn't always work.

In my book, "The Quiet Revolution In Email Marketing," I cited a ClickZ article that suggested double opt-in failure rates upward of 50 percent. In other words, for every 10 people who sign up for a double opt-in based permission email newsletter (or promotion), only five will actually complete the process. This kind of drop-off can undermine even the most successful email programs and begs the question: Why do so many people fail to complete the process? Here are some possibilities:

• The actual confirmation email isn't sent out immediately. By the time someone receives a confirmation, the person has forgotten all about signing up or has decided he or she no longer wants to be on the newsletter list.

• The recipient doesn't recognize the "From" field. In the past especially, a lot of double opt-ins would come from auto-responder systems that didn't allow the brand of the actual marketer to be displayed. So recipients would get a strange message and be unsure of what to do with it.

• Recipients have been trained never to click on anything in a message. I've spoken to people who are so concerned about phishing and Internet security that they won't click on anything in an email--not even a link to verify their permission.

All these reasons may be valid, but I believe the largest driver of failed double opt-ins is spam filtering. A recent article by Ken Magill of Direct online magazine brought this home, and it's worth reading if you've ever considered using a double opt-in. To summarize, a very legitimate mailer found its opt-in confirmation messages being blocked by an ISP spam filter. This is more common than most ISPs want to admit, but this particular case was worse than most. It turns out that some malicious person was providing a spam trap as their opt-in address. And, of course, the confirmation was then sent to that spam trap address, completely convincing the ISP's spam filter that the marketer was spamming.

The bottom line is that double opt-in is a powerful approach, but it's not a cure-all. On the positive side, I recall one of our clients using double opt-in to reconfirm a list they'd inherited, and the overall response rates were very high. On the negative side, malicious users, aggressive spam filters and poor execution all can combine to make double opt-in a problem. And, in my view, the only way to make customers even angrier than sending them unsolicited email, is NOT to send them the critical email updates they went out of their way to request. There's no easy answer for these challenges, but I'd love to hear from folks out there who have had strong experiences either way with double opt-in.