Faking X-mailer email field never rewards

October 13th, 2007
By Simon

Some email clients can fake the sent email’s X-mailer text. This is a very dangerous option to choose if you want your emails to stay away from spam folder.

When the email is composed, the X-mailer text is not the only common field filled by the origial sending software: there is also the X-MimeOLE field.

Microsoft Outlook Express attaches this MimeOLE: V6.00.2900.3138. This number depends obviously by the version of Outlook Express we are using, but not many bulk emailers can fake the MimeOle field too. Composing an email message pretending to be read as a common Outlook Express’ email and mistaking the final composition can result in a catastrophic spam score (if we were playing monopoly this means go straight to jail!).

If we are sending our email with a powerful bulk emailer that enables to change the X-Mailer text, and we pretend to be sending from Outlook Express in order to optimize our inbox rate, we are doing a great mistake: being caught by the Forged_Mua_Outlook SpamAssassin common spam rule after our email sending operation causes our email to get 3.4 spam points (very high spam rate, and very bad)…

Plus, if we are not careful in email composition, we can get additional 2.4 points for the Forged_Outlook_tags broken SpamAssassin common Rule, since Outlook has a very particular way of building html for the emails, it’s pretty easy to see if an email has been composed by another email client’s html composer.

Conclusions:
Faking X-mailer field is not a useful way to raise email inbox rate, even though it seems appealing. Keeping your email client or bulk e-mail sender identity clean (or at least leaving empty the x-mailer field) is the best option for your clear conscience and for your inbox rate.

Tests: you can do your own tests with our SpamAssassin engine spam checker software: try to modificate the x-mailer (you can try with SendBlaster Free Bulk Emailer) text and spam check your mail, then leave the original x-mailer text, and spam check your mail again. Then take a clean Outlook Express email and spam check it, you will notice heavy spam points decreasing.

Bookmark to:

Gmail spam filter Vs. Spamassassin

September 20th, 2007
By Simon

The official Google website says:

Google continues to work on spam filters to keep unwanted messages out of your inbox. However, spammers are deviously clever and adaptive. Defeating them will require your help. If you see a spam message, use the Report Spam feature to squash it. The spam will automatically be removed from your inbox and information about it will be reported to our quality team.

This means that the Gmail Spam filter is constantly learning the new spam behaviors, it stores each user click on the check this as spam button. But what if my mail gets caught by mistake by the Gmail spam filter? This issue is known as the “false positive” spam filter mistake. Gmail for the moment is not offering a I’m not a spammer button. Reading non-official sources today I discovered that SpamAssassin seems to be included inside Gmail’s master spam filter. How incredible, I know perfectly each spam rule inside SpamAssassin since I check my emails before each send with my colleague Luca’s homemade software* that incorporates Spamassassin rules set; and Gmail spam filter rates are not clear as spamassassin rates. Gmail spam filter has two options: it displays your emails in inbox… or spams them (they go straight to junk folder).

All the complex Gmail spam rules rate generation is invisible to our eyes

Well, most of the spam rules are based on common spammers’ behavior… but the spammers behavior is based on eluding each spam rule! So how can I get out of this recursive maze? Some clever guy uses Gmail as a spam filter check. It’s possible to send to our gmail account the email that we want to check, and then forward to our email client each email with gmail’s pop forwarding feature. So if the email gets back to our sending client, this means that Gmail is not seeing that email as spam.

But wait: Gmail spam filter is not identical for each client, since it must contain a different white list for each user (the email address of your friends for example… you can try sending a dangerous topic email to your friends, your email will be forwarded, since you are in their Gmail’s trusted network) .

So, who’s the winner? SpamAssassin or Gmail?

This Gmail homemade spam check is powerful but dangerous because of the differences between each gmail user settings; while SpamAssassin rules set is clear and has a fast integration on your machine without wasting time (*yes, check our download section if you want to try our free spam rate validator software).

What can I do to avoid gmail spam filter?

The best practice is having your email marked with a “not spam” action button in your custormer’s inbox.

Bookmark to:

Slicing and filtering: the mailing list in meaningful parts

September 11th, 2007
By Simon

Last month we learned how to ensure or improve email deliverability, how to check our email with a SpamAssassin spam checker rating softwares, including an email text version and using proper email html coding.

At this point we should send to our customers a light newsletter with interesting content, based on their needs. Did you put an opt-in webform in your website that asks for your customers interests, right? The opt-in webform lets you ask to your customer some important informations. Upon these informations you will divide your mailing list in sliced sectors.

A sliced mailing list can be named by targets, and gives you the opportunity to send just the emails to the most prolific customers in each sector.

You can tag each separate mailing list, using a tagname that can identify each mailing list sending session. Here are some examples of segmented mailing lists:

• last purchases
• most recent subscription
• top 10 customers
• customers active in September
• and of course, a single tag for every product.

Sending emails to people that reads them is the cheapest mail marketing solution, for you and for them. (… yes again, SendBlaster can do mailing list slicing and filtering)

Bookmark to:

Building a solid Html mailing list: subject matters

September 07th, 2007
By Simon

In the last few weeks we saw how to build a clean spam free mailing, following some simple steps and looking to the common spam rules that block our mails during the server side spam checking. Please remember small is better: few images and more text for your e-mails.

You want your subscribers read your email, right? But being checked as spam free is just the first step. Once the server machine decided that you are a clean mailer, comes the difficult part: your customers too have to think that your informations are useful for them. Spam it’s a matter of software, but spam can be a matter of context too. If you are out of your customers’ interest field, they will never open your email.

So, here are some advices based on our sending experience:

• try to put the name or the surname of your customer in the email subject field.
• try to send personalized informations based on your customers interest field.

Spam is all about requested information. Send only when asked to send.

(yes, as usual our bulk email software SendBlaster has the features that your email needs)

Bookmark to:

Spam score filter failure

August 22nd, 2007
By Simon

Today I’ve found an interesting article about spam filters failure. I post it here because we can use some of these methods to send our clean opt-in emails through the server spam filters.

In early July of 2006 there was a prominent stream of spam messages that simply quoted three random lines of the book ‘the hobbit’, with a subject header of 6 random letters. The popular belief is that this was a ’script kiddie’ who had got hold of spam suite and was inept in it’s use. This belief was reinforced shortly after when the same messages began appearing with an image overlaid, a popular spammer trick(The seemingly innocent words allow the message to bypass filters but all that is displayed on opening the message is the picture).

It’s nice to put seemingly random and pointless messages down to inept practice on the part of the spammer, but my personal belief is this is not the case. To me this seems to look like a deliberate attempt to corrupt the improving Bayesian Filter technology. Bayesian filters work by assigning a spam score to words that are found in spam e-mails. The more regularly they appear in a spam message, the higher the spam score and the more likely a message contianing those words is to be marked as spam.

Given that piece of knowledge, imagine the implications of a concerted campaign of spammed messages that contains a short message of commonly used words. The ’spam score’ of these words is elevated, the effectiveness of the bayesian filter is diminished and when the real spam message is sent through the defenses are lowered, or indeed have been removed having provided too many false positives.

The obvious clue to me is in the way these messages are sent. Firstly the title is randomised. Many Bayesian filters treat nonsense as ‘high spam’ score. Furthermore title text is usually given a higher priority than body text. Thus a nonsense title may be enough to get a message banned by itself, and combined with a common spammer trick such as the picture overlay, it seems the spammer wanted these messages to be caught.

As more people come to rely on Bayesian filtering, this will become a more and more serious problem. We already know spammers are prepared to send out millions of messages just to get one sale. Now it appears that they are also prepared to make multiple mailings to those millions of addresses just to soften up anti-spam defenses for their one commercial mailing.

Article source here.

Bookmark to: