« July 2007 | Main | October 2007 »

September 08, 2007

Op-ed in Wall Street Journal advocates hybrid solution to spam

Three researchers published an op-ed in today's Wall Street Journal (subscription only) suggesting that two practical methods to greatly reduce spam are now technically workable, but will not be implemented without cooperation on standards by the major email providers. They urge the providers to agree on a hybrid system:

To break this logjam, we advocate a hybrid system that would allow email users to choose their preferred email system. Those who want anonymity and no incremental cost for email can continue to send emails under the current system, without authentication and without sender bonds. Those who want the lowestcosts and don't care about anonymity (most legitimate businesses would likely fall into this category) can send email that is user authenticated, but not bonded. People who want anonymity but are willing to pay to demonstrate the value they place on the recipient's attention can post a bond. Payment could be made anonymously via a clearinghouse, using the electronic equivalent of a tiny traveler's check bundled with each message. Those with especially high-value messages can make them both authenticated and bonded.

The authors are Jonathan Koomey (Lawrence Berkeley National Labs), Marshall van Alstyne (Boston U) and Erik Brynjolfsson (MIT Sloan).

The ideas are not new; they are trying to create public pressure. The authentication system in play is DKIM, a standard approved by the IETF earlier this year. The sender bond method was detailed in a paper by Thede Loder, Rick Wash and van Alstyne. Loder has started a company offering the service (Boxbe); Wash is currently one of my Ph.D. students (though he did this research while working with van Alstyne while Marshall was my colleague at Michigan).

Posted by jmm at 10:37 AM | Comments (0)

September 02, 2007

Incentive wars: Buy blog comments

Raising the cost of polluting is one way to reduce pollution. Whether the pollution is a by-product (e.g., effluent from a factory), or whether it is the polluter's product (e.g., spam advertising), if the polluter is forced to bear more of the social cost of producing the pollution, he or she will have a good reason to produce less.

Most techniques for discouraging or blocking spam can be interpreted as raising the cost. My Ph.D. student Lian Jian told me about a lovely example that makes this point clearly: http://buyblogcomments.com. Most bloggers moderate comments so they can screen out spam by hand; they also often put up technical barriers to block spambots. These efforts are a form of raising costs for spammers, and the market has responded by putting a price on getting around these efforts. BuyBlogComments is a service that pays humans to enter "quality comments" that are related to blog postings, so they won't be deleted, yet that also include the URL the spammer wants to disseminate. For example, you can buy 100 blog comments for $24.99. They probably don't have 100% success, but suppose they do: we now have a reasonable estimate that comment moderation is at most going to eliminate spam comments that are worth less than $0.25 to the spammer.


Posted by jmm at 05:09 PM | Comments (1)