Musings on spam


I get a fair bit of spam. Every day I delete about 400 messages that my spam filter catches; this blog has amassed over 7,000 spam comments in six months or so; and now, Twitter is getting spammy too. I’ve noticed a rash of twitter-spam-bot followers recently, and am quite confused as to what they are trying to achieve.

All of the accounts (I’ve blocked nine over the last two days, including Britney Reed, Mallory Burt, Meagan Wolf, Suzanne Ratliff, Lacey Potts, Katy Bean, Beryl Buchanan, Bettye Stein, and, just now, Eliza Holden). Each of them had posted exactly one tweet (a rather boring message along the lines of ‘I’m bored, e-mail me here’), and had a moderate number of followers. The images were all the same style, also obviously automatically generated. This spam seems strange because they aren’t trying to sell me anything through links, and who is going to type in their e-mail addresses?

Then there is the other kind of Twitter spam I get, in which somebody into marketing with 10,000 followers and lots of banal tweets, wants to tell me all about cruises in Florida or how I can make money in through some scheme/scam. At least I understand (on some level) what those people are doing. If this trend continues, I am going to have to get a Topify account and a Twitter client that makes it easy to report spam. See this TechCrunch article for links to various clients and an interesting discussion of Twitter spam blocking strategies.

I also don’t understand why I get so much e-mail and blog comment spam in Russian. Surely they could target it better than that! What’s the point of posting Russian-language comments on a blog that doesn’t have anything to do with Russian or Russia?

Although as it happens, I find some of it quite educational in a perverse way, and have learned a great deal of current Russian slang on a variety of topics from the e-mail and blog comments. I wonder if there’s any money in that kind of a service.

Update: Having received a couple of additional Twitter spam-bot followers this morning, I took a look at whom a particular bot (Lupe Dudley) was following: In the 36 (of 1540) people shown by default, there were 15 Javiers, 14 Mauricios and 7 Jimmys.  I suppose this means that either they are all manufactured (but then why with such repetitive names?) or that the bots scan twitter subscribers based on some name pattern. In any case, this is yet another bit of mysterious weirdness around this twitter spam.


  1. I’ve noticed that certain subjects on my blog attract more spam than others. Do you have a sense for the topics and/or the post characteristics that are particularly magnetic for spam?

  2. I just looked at the last 20 spam comments I got, and they refer to posts 1563,718,1635,1597,725,529,200,610,637,350,1488,1608,637,1608,350,274,562,1631,993,1608. I am not prepared to build a regression model, and I am not going to copy the 7150 spam comments out by hand to get their attributes, but it seems pretty much all over the board. The numbers themselves are not meaningful (as far as I know) except that they indicate the order in which the posts were created. For calibration purposes, 562 was April 1st, and 993 was June 3.

  3. […] Referrer spam, and that doesn’t include the various forms of internet fraud. I might also add twitter spam to this list. Why isn’t that in Wikipedia yet?Anyway, the point is that while antivirus […]

Comments are closed.