Bennett Haselton has written in with his latest report. He starts "Goodmail has announced partnerships with four new ISPs who will charge for "reliable" delivery of your e-mail messages if you want to bypass their spam filters. The news will probably generate another round of editorials like the ones written a year ago about AOL's plan to use Goodmail, including this one from Esther Dyson (for it) and this one from the EFF (against it)." Follow the magical clicky clicker below to read the rest of this story.
If I could ask one serious question of anyone who was defending pay-per-email, or sitting on the fence about it, this would be it: Suppose you sent an extremely urgent e-mail to your doctor or your lawyer, who for the sake of argument you're not able to reach by phone. The recipient's ISP owner happens to see the message before the user retrieves it, and realizes how urgently you need to get it through. So he moves it to the recipient's "spam" folder, and then calls you up and says: pay me $1,000 to move it to the recipient's inbox, or they'll never see it.
Does the ISP have the right to do that? If not, why not?
Perhaps you'd say that Goodmail's 1/4-penny-per-message is reasonable, but $1,000 for one message is too much. But then who decides what is "too much"? The marketplace? Then isn't the ISP admin just another player in the market, and $1,000 is what they want to charge? If you don't like it, you can go somewh... oh, wait, you can't, because there's no other way to get through to the recipient. If you ever get through to your doctor or lawyer, they might switch ISPs after they hear what happened, but should that be your only recourse?
The problem with the ISP charging $1,000 to deliver your message is not that $1,000 is "too much", but that they're charging for a service that has already been paid for. If your doctor or lawyer pays for an e-mail address, they're doing so with the understanding that their ISP will make a reasonable effort to deliver the non-spam e-mails that people try to send them. If their ISP then turns around and asks you for $1,000 to deliver the e-mail, then they're trying to double-bill for the same service, and if they block the message because you don't pay the $1,000, then the ISP is cheating the recipient out of a service that they've already purchased. And it's not just the recipient being cheated; if the recipient has an arrangement with you, as your doctor or lawyer would, then the ISP is interfering in their business relationship with you.
Now, if an ISP using Goodmail offers to let you bypass their filters by paying 1/4 penny per message, how is that different from the doctor example? Well, on the face of it, it's different in at least two ways: first, because the ISP is charging "only" 1/4 penny per message instead of $1,000, and second, they're not saying that your mail will be blocked if you don't pay, only that it might be. But are these qualitative differences, or just differences in degree?
Take the cost-per-message. I have a (verified opt-in) mailing list of about 50,000 people that I send mail to twice a week. In the aggregate, it is just important for me to get mail out to those subscribers, as it is for some people to get a single mail through to their doctor or lawyer. Also, in the aggregate, it would cost me about $1,000 per month if the ISPs collectively asked for 1/4 penny per message and threatened to block them otherwise. So is there any real difference between requesting $1,000 to unblock 50,000 e-mails, and requesting $1,000 to unblock a single e-mail, if you're just doing it because you know the sender urgently needs to get them through? (It's not a reflection of the ISP's costs -- downloading and storing 50,000 messages at 3 K each, costs almost nothing, certainly not anything close to $1,000. And again, I would argue it's a moot point anyway, because those services have already been paid for.)
And how much difference is there, really, between saying that a message (or a group of messages) might be blocked, and saying that a message definitely will be blocked? If it's bad for your doctor's ISP to call you up and say, "Give me $1,000 or there's a 100% chance that your message doesn't get through," what if they say, "Give me $1,000 or there's a 50% chance that your message doesn't get through," isn't that at least 50% as bad? You could say that in my doctor example, the blocking was deliberate, but in the case of the spam filter, it's accidental. But if an ISP chooses not to fix problems with its spam filter, then in a way it's still deliberately creating a certain percentage of cases where the spam filter will block legitimate mail, even if those cases occur at random.
There is one more difference between Goodmail and the scenarios I've described, which is that Goodmail not only lets you bypass an ISP's spam filters, it also certifies that you are trusted and not a phisher. If an ISP like AOL controls the user-interface that a user uses to check their mail, it can display the blue-ribbon "CertifiedEmail" icon next to a Goodmail-certified message. In this case, an ISP can plausibly claim that they're letting all legitimate e-mail get through, but they're still offering a benefit to Goodmail senders. The problem with this is that since phishing only works on users who are gullible to begin with, a phish could just as easily display the CertifiedEmail icon in the body of the message to try and gain a user's trust. It's all very well to say that a user should know that the CertifiedEmail icon only "counts" when it's displayed in the inbox, not in the message itself. But a user who knows that, would probably also know that their bank's Web page is not 22.214.171.124. And besides, most users of Comcast, Cox, RoadRunner and Verizon will be using their own mail clients like Eudora which won't display the "CertifiedEmail" icon anyway.
So it seems pretty clear that the main benefit of using Goodmail will be deliverability. And that's the basic Catch-22: If an ISP gives the same deliverability to non-Goodmail-certified messages, then who's going to use it? On the other hand, if an ISP gives better deliverability to Goodmail-certified messages than to other messages (much more likely), then they are to some extent misrepresenting the services they sell to their users, since users expect an ISP to make the best effort to deliver all legitimate e-mails, not just the ones from paying senders.
Goodmail likens their service to FedEx or UPS for "enhanced delivery" of paper mail as a way of getting the recipient's attention. But the difference is that if you're trying to reach your lawyer, then the office complex where he works (or the city that maintains the streets to his house) is providing the service that he expects and has paid for, namely, allowing different companies to deliver stuff to him there -- and because you have different choices, that means FedEx, UPS and the USPS have to compete with each other, and that keeps the delivery prices down. On the other hand, if an ISP blocks you from mailing their customer unless you pay their fee, then the ISP is going against what the customer expects them to do, and it is precisely that betrayal of trust that gives the ISP a monopoly on your ability to reach the customer -- which leads to them charging monopoly-style prices, like $1,000 to receive and store a few tens of thousands of messages.
There is a lot of debate about whether "the market" would fix problems of legitimate e-mail being lost. Esther Dyson's editorial was a classic libertarian defense of the free market as the arbiter of systems like Goodmail: "If it's a good model, it will succeed and improve over time. If it's a bad model, it will fail. Why not let the customers decide?" Actually I don't think the free market does fix most e-mail deliverability problems -- I've been involved in a few business that sent bulk e-mail (to subscribers who requested it and confirmed their subscriptions), and have had conversations with dozens of others, and we've all had problems sending to Hotmail, AOL, and Yahoo, and I've never, ever heard anyone say that their deliverability problems were solved by "the market". (Usually the problems just come and go, and nobody knows why.) But in a way this is all beside the point. Even if the market would stop more egregious abuses, what gives ISPs the right to charge senders for e-mail services that their customers have already paid for?
I actually met Richard Gingras, the CEO of Goodmail, and Charles Stiles, the postmaster of AOL, at a conference in Seattle last year where they were on a panel defending against the Goodmail controversy. They seemed like nice guys who were genuinely blindsided by the criticism that Goodmail had been receiving. It's easy to see the point of view of Goodmail's defenders -- if Bob wants to pay Alice to "certify" Bob, why would it be anybody else's business? It isn't, until it leads ISPs to steer people towards a system where if you want to be treated like a non-spammer, you have to pay -- even if, strictly speaking, the recipient is already paying to receive your mail.
As for the much-vaunted free whitelisting privileges that non-Goodmail senders will continue to enjoy, in the pre-Goodmail era I once found that AOL was blocking some of my mail to their users, so I called their postmaster department and learned the following facts:
- The first person I talked to, said that he checked the logs and our mail was being blocked because we didn't have reverse DNS set up. I thought this was odd because we did have it configured, but I thanked him and hung up.
- Then, I called back and got someone different. I asked them the same question and they said that according to his logs, our mail was being blocked because someone else at our ISP was sending spam. I asked him why they were blocking our IP address, if it was different from the IP of the alleged spammer; he paused and said, "Is there anything else I can help you with?", and this repeated several times as I thought my phone or his headset wasn't working, before I realized he was just being a dork.
- Then, I called back and got yet another person, and this person said that he could see our mail was being blocked because it contained banned content. I was pretty sure that was wrong, because you get a different-looking bounce if you're sending mail that contains a banned string, but I took a note of that anyway.
- Then, I called back and got a fourth person, who said that our mail was being blocked because some of their users had flagged mail from our IP address as spam. He paused for a brief conversation in the background, then came back and added, "This has already been explained to you, sir." I said that since I had gotten four different explanations in four different phone calls, I figured I could just keep calling and tallying the votes that I got for each explanation, until one of them emerged as the winner.
Much later I found out from someone else about the AOL whitelisting program, which I'm currently trying to see if it prevents us from getting blocked. But if none of the people answering the phone at the postmaster department knew or told me about it (and I confirmed that it did exist at the time), how many other organizations or businesses don't know?
ISPs adopting Goodmail say that while Goodmail senders can bypass their spam filters, non-Goodmail senders will continue to enjoy the same deliverability rates that they have in the past. That's what I'm afraid of.