A grand experiment in "pay-to-play" e-mail will begin in the next few days, and the results will influence how we communicate for years to come.
Goodmail Systems, which plans to charge some portion of 1 cent per e-mail to ensure that senders' messages are delivered into AOL inboxes, will go live in a matter of weeks. Will this idea generate lucrative streams of revenue -- some of which will be shared with AOL -- or will it end up as a mouse that failed to click?
Much has been written about whether it is "right" to charge for e-mails that, since their invention, have been free for anyone to send. That's a question the marketplace will decide, like it or not. But there's a more interesting story here, aside from the pennies that may or may not add up to millions.
Criticizing Goodmail as a Simplistic Solution
Goodmail Systems works by selling "tokens" to large senders of e-mail, such as newsletter publishers and corporate marketers. When the system is fully operating, these companies' e-mails will be digitally signed by Goodmail's servers and passed along to AOL.
AOL has agreed to place these messages into its subscribers' inboxes, eliminating any chance that the e-mails will be shunted into a "junk" folder. Even better, such messages will be delivered with their images fully visible and their links clickable. By default, AOL disables images and links in e-mails unless its subscribers specifically turn them on for a particular message or sender.
Goodmail's criteria for giving "certified mail" status to certain senders -- in addition to their willingness to pay the fee -- is purely based on the number of complaints received from AOL customers. If more than 2.2 recipients per 1,000 messages (0.22 percent) click AOL's "Report As Spam" button, you'll reportedly be warned and ultimately kicked out of the program.
That single-mindedness is at the heart of some e-mail experts' biggest criticisms of Goodmail.
"They're starting to realize that they need to look at more points of data," says George Bilbrey, a general manager of Return Path, which helps companies measure and improve their e-mail deliverability to major ISPs. "We look at a lot of different data sources," he says, including how many bad addresses a company is sending mail to (a warning sign to ISPs of spammers), how quickly unsubscribe requests are honored, and whether e-mail addresses have been "harvested" or scraped willy-nilly from Web pages.
In an April 20 panel on deliverability at Marketing Sherpa's E-Mail Summit in Chicago, representatives of other firms specializing in e-mail services traded jabs with Goodmail's CEO Richard Gingras.
In an interview after the summit, J.F. Sullivan, a vice president of the deliverability service Habeas, said, "My concern with Goodmail is it reinforces the idea that complaints are the only mechanism to determine what is legitimate mail. ... There's no incentive, no accountability, in this design that forces the company [Goodmail] to respond" if a sender suffers a sudden spike of spam-button accusations.
Goodmail Defends Its Value
Gingras confirmed in a telephone interview that the 0.22 percent complaint rate is being used as a threshhold to separate the steak from the spam. "But that is simply an element of the qualifying rate for senders," he says. "The actual complaint rate they must have after they're in the system will be more nuanced than that." Transactional messages -- which confirm e-commerce purchases, for example -- are expected to generate a lower complaint rate than marketing messages or newsletters, he explained.
Two of the statements Gingras made during the Chicago panel raised the eyebrows of the other assembled speakers as well as members of the audience. Three out of four corporate applicants for the Goodmail system are being rejected, he said. And no company that has been in business less than one full year will be allowed to apply at all.
"The majority of applicants [that are being rejected]," Gingras tells me, "were tied to not having low enough complaint rates."
Even a very low complaint rate can't help startups, which Goodmail won't evaluate. "I can imagine a year from now, we'll have a different policy to handle newer companies," Gingras says. "There are certainly lots of techniques someone could use to vet a company. At this point, we're not supporting those techniques."
How Important is it to Blanket AOL?
Goodmail's value for companies, especially considering that many legitimate senders are being turned away, is questioned by Michelle Eichner, a vice president of e-mail auditing firm Pivotal Veracity and one of Gingras's co-panelists.
"It's just not very well thought out, based on the numbers [Gingras] mentioned that are being rejected," Eichner says. "All of our clients are getting into the AOL inbox 92 percent of the time," she adds. "I don't know at this point that there's a tremendous amount of motivation for any of our clients to pay for Goodmail."
A major factor to consider in the pay-to-play equation is how important it is for a given company to reach an additional fraction of the AOL members who opted into a list. A few years ago, AOL was the 800-pound gorilla of U.S. e-mail accounts. But AOL has been shrinking for years, losing more than 2.8 million U.S. customers in 2005 alone.
The giant has slipped so far that it's not even the 2nd largest domain in most companies' e-mail databases any more. On business-to-consumer lists, AOL addresses represent fewer than 13 percent of the names, according to a recent study by Email Labs. Yahoo, Hotmail, and MSN add up to more than 32 percent. In business-to-business lists, AOL is under 3 percent. Its big competitors are far ahead at 12 percent.
Authentication Replaces the Need for Buy-In
Even for those companies that are desperate to reach every last subscriber, there's a more effective alternative to Goodmail that costs little or nothing -- e-mail authentication.
Authentication systems enable senders to prove who they are and allow ISPs to monitor whether those senders generate legitimate mail or spam. Two major techniques to authenticate senders have caught on -- Sender ID Framework (SIDF), which is strongly supported by Microsoft, and DomainKeys Identified Mail (DKIM), developed by Yahoo, Cisco, and others.
I previously wrote about these proposals on Sept. 28, 2004. Since that time, their growth rate has surprised even me.
As of last month, 75 percent of Fortune 100 companies are using Sender ID to back up their marketing e-mails, and 45 percent are authenticating them with DomainKeys, according to a ClickZ column by Kirill Popov and Loren McDonald.
Sender ID (particularly its simple subset, SPF) provides a list of authorized servers that may originate e-mails bearing a company's name. DomainKeys takes this a step farther by encoding e-mail messages so they can't be altered in transit and can be traced back to one specific domain.
Sources tell me AOL plans to implement both mechanisms by the end of this year. The high adoption rate of these techniques means that some ISPs have begun to use these markers to sort legitimate mail from the spam.
According to the Email Sender & Provider Coalition, an organization of major ISPs and e-mailers, every one of the top 18 ISPs in the U.S. is currently tagging its outgoing mail with Sender ID and/or DomainKeys. In a statement by the group, many of the same ISPs are said to already be using the absence of the two methods as part of their "scoring" of incoming mail.
Using one or both of these authentication approaches on your company's outbound mail is "absolutely" a good idea, says Dave Lewis, a vice president of StrongMail Systems and another Chicago summit panelist. In fact, authenticating your mail may give you all the help you need. "If you've got a good reputation, your mail is going to be delivered to the AOL.com domain," Lewis says.
Conclusion
In this space next week, I'll dig into today's swift uptake rate for e-mail authentication -- and how it can make a difference for your company.
Images
A grand experiment in "pay-to-play" e-mail will begin in the next few days, and the results will influence how we communicate for years to come.
