David MacQuigg <macquigg(_at_)gmail(_dot_)com> writes:
I have been using gmail for the last few years. No connection other
than as a client. Spam in my inbox is maybe one in 200. I have noticed
one or two lost emails in the last year. Before that I was not aware of
any. Note: I don't count as spam any that appear on a separate tab
labeled "Promotions". I don't count as "lost" any that appear in my
spam bucket. How do they do it?
Gmail sees something like 20-30% of all email on the Internet and has
millions of users manually classifying false positives and false negatives
for them in real time in their application.
There are obviously a whole bunch of other technical controls, resources
to build large databases of vendors, and so on that are hard to reproduce
economically, but I think the first point is the critical one. Spam is a
classification problem, your classification engine gets better and better
the more data you have (assuming that you have the engineering resources
required to do something with that data, which Google does), and Gmail has
by far the best source of data of anyone doing email.
The things that many people dislike about Gmail (central concentration of
email clients and senders into one powerful entity, vast visibility into
everyone else's business, the ability to gather surveillance information
through network effects) are exactly what makes them so good at spam
filtering.
--
Russ Allbery (eagle(_at_)eyrie(_dot_)org)
<http://www.eyrie.org/~eagle/>
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp