Competing spam 'solutions'

From Knowledge@Wharton
Special to CNET News.com
January 25, 2004, 6:00 AM PST

New federal legislation in the United States to put a lid on Internet spam--the torrent of unwanted commercial e-mail promoting such things as Nigerian business investments, mortgages and body part enlargements--may help against the electronic onslaught.

But better solutions are probably to be found in economic and technological change, according to Wharton faculty.

The term "spam" comes from a Monty Python skit, in which a customer in a restaurant is deluged with demands to order canned Spam before screaming out, "I don't want any Spam." Under a tsunami of junk solicitations, now estimated to make up more than half of all e-mail, Internet users have made the same cry to Congress.

Known as the , the new federal legislation authorizes, but does not require, the Federal Trade Commission to set up a similar to the National Do Not Call Registry, which is now . It also carries criminal penalties of up to five years in prison for violators. "With this bill, Congress is saying that if you're a spammer, you could wind up in the slammer," said , a supporter of the legislation.

The federal provisions would supersede tougher legislation in several states, including , which took effect Jan. 1. The California legislation requires marketers to gain permission from consumers before sending e-mail messages. By contrast, the federal legislation puts the burden on consumers to opt out of junk e-mail.

According to Eric Clemons, a Wharton professor of operations and information management, the volume of junk e-mail has grown beyond an annoyance. "The Net could get crushed by volume," he says, adding that while the federal legislation has good intentions, a more effective way to stop spam is rooted in economics. "One entire approach to making this work has nothing to do with forbidding it--just to the sender."

He points out that bulk advertisers who use traditional mailings bear the burden of printing their materials and paying for postage. The sender has to decide whether that investment will pay off with new customers. But e-mail is essentially free for bulk mailers, because it does not cost much more to spam millions than it does to send a single e-mail. "If it costs nothing to the sender, he thinks, 'Why don't I send it to everybody on the off chance somebody might be interested?'" Clemons says. "There is no cost to the sender in wasting my time."

The best way to reduce spam is to attach a cost to sending it, he suggests. "If the sender has to think before he puts me on his mailing list, then it's going to limit the volume he sends and the volume I receive. Everybody is happier."

According to Clemons, there are several ways to shift the burden of spam to the sender. One approach would be to use technology that enables Internet service providers to require that PCs perform a short calculation before sending any e-mail.

The DMA supported the congressional legislation but objected to the "do not e-mail" provisions.

The delay would have little impact on normal e-mail users but would have a significant effect on spammers. This approach, which Microsoft and other technology companies are studying, addresses the issue of shifting the cost of bulk mail but falls short of benefiting the public, Clemons says. "Anything that just creates computational friction or some form of economic waste is silly."

A second--and better--solution, he says, would be to charge a small tax to send e-mail, then use the proceeds to upgrade the infrastructure of the Internet. The cost would not be enough to affect regular e-mailers, but in multiplying even a small tax by their tremendous volumes, it would make the tax prohibitive for spammers.

A third solution, studied by professors at the University of Michigan, would be for bulk mailers to post a bond that would be used to give e-mail recipients a micropayment to read messages. This could be waived, if the recipient felt that the message was worth reading.

Of the three approaches, Clemons prefers the second, because it is the easiest to administer, even though economists dislike the idea of any tax, because they argue that it can distort usage. "But you already have a distortion," Clemons says. "People are assuming that wasting my time is free." A system taxing bulk e-mail or requiring a bond could be administered by ISPs or other organizations, not necessarily government, Clemons says, adding that "the Net has done a marvelous job of organizing itself, so far."

Target the advertisers
Dan Hunter, professor of legal studies at Wharton, agrees that the cost of spam must shift from the receiver; in effect, the legislation attempts to do that with fines and jail time. However, he says, the federal law will be difficult to enforce, because spammers can easily move offshore and beyond Congress' jurisdiction. "At the moment, most of the big spammers are American citizens using foreign servers to mask their identity. The difficulty is going to be actually finding them and catching them."

Instead of focusing only on the spammers, Hunter says, government should go after the advertisers, who are likely to be located in the United States. "The people are often local. If you attack them as well as the spammers and the ISPs that host them, this multipronged approach is much more likely to work."

According to Hunter, the big obstacle to this plan is political opposition by powerful interests, including the . "Any attempt to actually stop people advertising is politically sensitive, because they have lots of money, and they don't want their market taken away."

The DMA, which represents 4,700 companies in the United States and 53 others abroad, supported the congressional legislation but objected to the "do not e-mail" provisions. Hunter says the marketers supported a weak version of the legislation that will now supplant the California law. "They supported the watered-down version; the one that's not going to work and does not affect their interests." Marketers had objected to the California legislation, because they said it would be difficult to identify California e-mail and would have left businesses open to frivolous consumer lawsuits.

According to John Mozena, co-founder of the , not only direct mailers--but also other large corporations and business associations--joined with the DMA to back the federal legislation over the California law or tougher versions debated earlier in Congress.

One solution would be for filters to add a message to the sender saying the message was deleted as possible spam.

"It's not just the DMA. Every company of any size uses direct marketing," Mozena says, adding that the legislation does not go far enough to protect e-mail users from spam. "It fails the most basic test of antispam legislation. It doesn't tell anybody not to spam. It regulates spam and requires truthfulness in spam, but it doesn't stop any marketer anywhere from sending us messages."

Still, Mozena says CAUCE believes that spam can be legislated out of in-boxes. "We agree with the basic concept that there needs to be financial disincentives for spammers. But we continue to believe the only way to do that is through laws and the courts. We have yet to see any suggestion for a technologically backed way of assessing the cost of the marketing--and getting it back to the original sender--that can be deployed in the real world."

Higher usage, higher fees
According to Thomas Lee, a Wharton professor of operations and information management, ISPs are already exploring a new pricing structure that's known as "quality of service"-based pricing, which could be more effective in controlling spam than legislation. Broadband providers, he says, charge a person who sends a few e-mails a week the same amount as a teenager who downloads music in huge chunks of data. "If you are an ISP, you would like to grade your level of service so people using more bandwidth pay more."

He, too, says one of the main difficulties with the new U.S. legislation is that it does not address the problem of foreign spammers. But if ISPs charge additional fees for heavy users, that would be a disincentive to spammers around the world, because their local ISP would be charged a higher interconnect fee by U.S. ISPs to get the messages in front of American e-mail users. In turn, the foreign ISPs would then charge their spamming customers more.

Lee acknowledges that the fees would have to be carefully structured to avoid hurting individuals who legitimately want to send packets of information. He also notes that whatever pricing thresholds are set, spammers may be able to get around them, just as money launderers get around U.S. banking requirements to report cash deposits of $10,000 by making deposits of $9,999.

Spammers could get around filters in the same way, says Lee. If they figure out that antispam filters are bouncing out mail going to more than 5,000 recipients, the spammers will send their messages to 4,999. "That's a constant game."

Balaji Padmanabhan, professor of operations and information at Wharton, suggests that the idea of paid e-mail might be a long-term solution.



		2003 in review Spam: The scourge can't be contained Efforts to stop the deluge were ineffective.

"At some point, it would be nice to see someone experimenting with this concept so we know the problems and benefits."

In the short run, however, Padmanabhan suggests that technology can be used to stop much of the spam overload. The software for spam filters is easier to write than virus detection software and can easily be updated, as spammers find new ways around the system. Antispam filters would be most effective if they are applied by ISPs that can keep a closer watch on Internet traffic trends than individual PC users can. The danger with any filter, he acknowledges, is the possibility of creating false positives that delete legitimate e-mails. "You're worried about missing an e-mail from a high school friend who has not been in touch for 20 years."

One solution would be for filters to add a message to the sender saying the message was deleted as possible spam. "If the sender is a friend from high school, they can at least know to work a little harder at finding another way to get in touch with you...The bottom line," says Padmanabhan, "is that good spam filters will go a long way toward dealing with the spam problem."

Wendy Moe, who earned a Ph.D. at Wharton and is now a marketing professor at the McCombs School of Business at the University of Texas, says that of the three ways she receives unsolicited advertising--mail, telephone and e-mail--she finds e-mail the least intrusive. Aggressive legislation, taxation or technological filtering of spam may choke off development of a potentially valuable consumer service, she suggests. "Intelligent marketing makes spam a service, as opposed to an annoyance."

When Moe uses e-mail systems through an employer or a private server, she receives far less spam than when she uses a big commercial network such as . She suggests that consumers complain to ISPs who allow their lists of e-mail addresses to get into the hands of spammers. "That's not something the legislature needs to deal with."

She is not ready to fully endorse the federal legislation, nor is she ready to say there will never be a time when spammers do not need legal sanctions. "There are extreme cases. There are situations where the market can only do so much. But I don't think we've had enough time to let the market try and sort it out," she says.

To read more articles like this one, visit Knowledge@Wharton.

Interested in more research studies like these? If so, sign up for the Knowledge@Wharton Newsletter, a free service of the Wharton School of the University of Pennsylvania.

Competing spam 'solutions'

Get the CNET Daily News newsletter