Dave Taylor (of Intuitive.com and MicroURL.com) just emailed me that this morning’s HTML Surfnetkids newsletter tripped his spam filters with the following message:
X-Spam-Status: Yes, hits=6.9 required=5.0
X-Spam-Report: 6.90 hits, 5 required;
* 0.6 — BODY: One hundred percent guaranteed
* 0.3 — BODY: No such thing as a free lunch (1)
* 0.3 — BODY: Satisfaction Guaranteed
* 0.3 — BODY: Asks you to click below
* 0.2 — BODY: Sign up Free Today
* 0.1 — BODY: Free Trial
* 0.1 — BODY: Contains ‘Dear Somebody’
——–>>> * 3.0 — BODY: Image tag with an ID code to identify you
* 1.1 — BODY: Spam phrases score is 03 to 05 (medium)
[score: 3]
* 0.3 — BODY: HTML font color is red
* 0.2 — BODY: A WHOLE LINE OF YELLING DETECTED
* 0.2 — BODY: Includes a URL link to send an email
* -0.2 — Mailing list headers are suspicious
* 0.4 — HTML-only mail, with no text version
The biggest problem (the 3 pointer annotated with —–>>>) was an advertiser’s image and linking URL that contains “?id=1042&sub=Q3C036”. (Note to self… remember to check advertisers’ URLs for spamminess.)
Interestingly enough, when I used Lyris Content Checker (which uses Spam Assassin) the same message only rates a 3.3 — which means it should have gotten through MOST filters.
Ah, the joys of email publishing! In the future, I’ll be talking more and more about publishing and receiving content via RSS.
BTW — Dave also told me his new service MicroURL could have avoided the spam filter problem by creating a non-spammy URL.
Mea culpa. I have certain spam things that I give a high score to:
score WEB_BUGS 3.0
score NO_REAL_NAME 2.5
score BASE64_ENC_TEXT 4.0
score MISSING_HEADERS 2.0
score MAY_BE_FORGED 2.5
score GAPPY_SUBJECT 2.5
score MSG_ID_ADDED_BY_MTA_3 2.5
score EXPECT_TO_EARN 2.0
I just want to have spam COMPLETELY removed from my mailbox.
I just checked the two Spam Assassin reports side-by-side and yours assigns 3 points to “Image tag with an ID code to identify you” while the Lyris Content Checker assigns .1 points.
Which just goes to show us that TRYING to be smart (and checking email content against one spam filter or another) ONLY gives us email publishers a FALSE sense of security.
Even more curious, Barbara, I too am running SpamAssassin, and that’s where all those scores came from. 🙂