Comment Microsoft is still buying crap research. (Score 4, Interesting) 182
Calling this "independent" is hogwash. It's a scam MS has been pulling for well over a decade, paying for "independent" competitive studies whose design and publication they control, and then trumpeting the results of the ones that say things they like.
In this case, the methodology was designed in a way that only exposed the test addresses to a narrow subclass of spam and which helped rationalize the fact that the study is completely blind to false positives. It cannot be accidental that the most widespread criticism of Hotmail and Microsoft's other hosted mail services by outsiders who work with mail servers and spam control is not that they deliver or emit spam, but that they have massive chronic false positive problems, not just with mis-filing into "Spam" or rejecting in SMTP for no good reason, but with mail being accepted for delivery and vanishing without a trace, in large volumes. It's a mess and I am 100% certain that MS knows about internally, at least at senior mail geek levels. It is a spectacular display of chutzpah for MS to be applauding themselves for a study in which they would have been beaten by a email system with no Internet connectivity.
And as someone who has been dealing with spam filtering and prevention since before anyone at MS knew that "spam" wasn't just a Hormel product, I should add that a methodologically sound study of the filtering systems of the big freemailers is probably not possible in the real world. Different people get significantly different types of spam and non-spam based on the history of their addresses and how they use them, and you really can't say anything meaningful about an 'average' mail stream because no real address has one. The big freemail providers have a very hard job because of the scale and diversity of their user base and pathological business models, but that can't justify promotion of a study which ultimately is worthless.