RBL lists, no matter how accurate the CAN be should never be used as a sole spam source. I have studied this very extensively over a 8 month period where I studies tested and evaluated almost all SPAM software on the market I could find 8-10 hrs a day (what a pain that was). ALL, and I mean ALLm, cheap solutions are left to be mainly dependant on RBL lists and are just not reliable for busines use. A FAIR system will use RBL AND some keywording, a DECENT one will also check message body text and the most robust systems have 6 or 7 different types of scan they perform in order to scan SPAM effectively, EVEN then , the lists need to be SPAM vs HAM trained (another benefit of GOOD software).
So I stick to my oringinal comments in that it is most likely due to the recipient having cheesy or low end spam software installed that is simply STOPPING anything found in what could be 20 different RBL checks.
If you want REAL and PROVEN spam protection for business, be well prepared to spend from $1000.00 to $6000.00 in software and hardwae costs (obviously the low end not including new hardware).
THis is something I studied VERY closely for some time befor espending 6 month in serious test envronments and testing near 40 different busines targeted solutions. I actually did write a white paper on it and was asked by several companies,either by phone or email, to evaluate their systems and make recommendations.
RBL lists CANNOT AT ANY TIME be considered an deffective way to stop spam, they simply are ot monitored nor accurate enough.
How does an RBL list understand what YOUR company's acceptable and unacceptable mail is?
Just because a customer has an open relay, I still want them to be able to contact me. I wasn't worried about open relays at my office, GroupWise is a little more secure than most would think.
I took a company getting nearly 1500 spam a day to less than 5 per week between MANY users, four locations and without a SINGLE whitelisted address, yes not ONE. It did take about 4 months training expensive software, sorry but SPAM Assasin, Spam Cop and even the few thousand dollars for GW Guardian wasn't as effective.
What if company B wants to order aproduct from you and because of your system scanning 4 RBL lists it is then blocked and you lose the business, this costs many thousands each year in lost business.
RBL was the first and lowest weihgted score in the system I built. It then was scanned for keywords, then a FULL heuristic scan against a 45000 entry databse, then for address blocks, then against the HAM list, then scanned for viruses and trojans, then the attachments wer eal lscanned, this is before it enters the company's MTA, and long befor it got to the desktop. Yet it all took place in mere seconds and using the company's existing AV software. The footprint was a little under 100MB and it ran on the existing server while raising resource use only a few percent(2-4%).
Another system I tested did similar, ye required a dual processor server running Windows 2003 (for a Novell targeted product that works with GroupWise?).
There's a lot of garbage out there yes, there are lot in fact THOUSAND of incorrectly blacklisted companies, many not even knowing it.
RBL's should be used ONLY as a preliminary scan and always weighted against a threshold that it a result of multiple scans, but NEVER as a final decision for spam, they just aen't reliable enough nor managed well enough, and yes I have spoken personally with many of the list providers regarding their update and management measures, which for the most part are quite pathetic.