An intriguing forthcoming paper on spam (no full-text available yet) from one Guido Schryen: an attempt to get some hard data into the debate. Here are the key findings:

  1. Web placements attract more than two-thirds (70%) of all honeypot spam emails, followed by newsgroup placements (28.6%) and newsletter subscriptions (1.4%)
  2. The proportions of spam relating to the email addresses’ top-level domain can be statistically assumed to be uniformly distributed
  3. More than 43% of addresses on the web have been abused, whereas about 27% was the case for addresses on newsgroups and only about 4% was the case for addresses used for a newsletter subscription
  4. Regarding the development of email addresses’ attractiveness for spammers over time, the service “web sites” features a negative linear relationship, whereas the service “Usenet” shows a negative exponential relationship.
  5. Only 1.54% of the spam emails showed an interrelation between the topic of the spam email and that of the location where the recipient’s address was placed, so that spammers are assumed to send their emails in a “context insensitive” manner.

Surprised by how low the last one is, although I shouldn’t be…