First letter of email address determines spam load

Albuquerque Express (IANS) Tuesday 2nd September, 2008

How much spam you get depends on the first letter in your e-mail address, a Cambridge study reveals.

Analysis of more than 500 million junk messages has found that addresses that began with more common letters were likely to receive 40 per cent of their mail from spammers. Those starting with less common letters, by contrast, would receive less than a fifth of their mail as spam.

According to the study, if the first part of an email address (that part before the '@' symbol) starts with a J, A, U, I, R, P, M, or S, then it is likely to get proportionately twice as much junk email sent to it than an email address beginning with Q, Z, W, Y or F.

Dr. Richard Clayton, a computer scientist at Cambridge University who carried out the study, said he believed the difference could be explained by the first set of letters being more likely to appear at the start of names than the second set.

The study did not draw a conclusion as to the exact cause of this phenomenon. However, Dr Clayton said there was some evidence to suggest that it could be due in part to the way some spammers launch 'Rumpelstiltskin' attacks. This is where the spammers use dictionary words and proper names in ascending alphabetical order in front of large numbers of domain names to generate their junk lists.

Using this spam generation technique, junk emailers would, over time, email more addresses that started with letters in the first half of the alphabet than the second half.

The research looked at 550 million email messages sent to customers of the Demon Internet service, one of the UK's largest ISPs, between February 1 and March 27 this year, according to The Telegraph.

Dr. Clayton, who presented his findings at a conference on email and anti-spam in Mountain View, California last month, concluded: 'Measuring incoming email has shown that the first letter of email addresses makes a difference to the proportion of incoming spam.' He said the results could help in the development of anti-spam measures.