SophosLabs has out a report they ran which shows the words used in spam which are mangled in order to disguise them and get them by the spam filters. It looks like even though Viagra was at the top of the list in previous studies done over the years, Cialis has now taken top "honors".
Before giving you the list of the top 25 words which are disguised in spam emails, the report adds "NOTE: Information contained in this report may be considered offensive by some customers." The reason for that of course is that sex sells - pretty much all of the words on that list are either directly related to sex, or are drugs related to sex.
I'm curious how the pharmaceutical companies see this. Sure, their public facing comments have to be that they are against spam and they likely very much are against it since they don't want to taint their image (or rather have it tainted by others in the case of spam). But I imagine it also serves as a large indicator of popular opinion as well - whatever spam words are hot are words which will make money. That means Cialis is currently more likely to make money than Viagra.
Posted by Eric at March 24, 2005 09:17 AM
| TrackBack
That is a fairly common thing to see and one of the many reasons that basic text matches are a risky endeavor in terms of blocking spam - especially if that is the final check on a message.
Software like SpamAssassin does a bit of that matching (arguably a lot more advanced, but it is the general concept), but instead it weights the result of a hit and then if enough weight is surpassed, then it is marked as spam - but any one hit (usually) isn't enough to call it spam.
We deal a lot with UK companies and "Essex" shows up fairly often. You can see where I am going with that one. A year or two ago here I can remember people talking about how that mail was getting blocked (not by our office, but a law firm we deal with).
Posted by: Eric at March 24, 2005 02:09 PM
Yeah, if you ever want to ruin your day, do what I did as a systems administrator: while testing out some new spam software I thought I'd filter for certain words like viagra and cialis and archive those messages for deletion.
An hour or so later, I noticed that the filter had collected thousands of messages more than I had anticipated. A cursory look at one revealed that filtering for "cialis" resulted in every message with "specialist" in it being tagged as a false positive. Users began to complain immediately that they "couldn't find" certain e-mail messages they knew they had received...
Oops! I definitely on't make that mistake again.
Posted by: Reed Wiedower at March 24, 2005 12:36 PM