If this idea is a rehash, please accept my apology and point me to the former
discussion.
What if we turned the spam problem upside-down? Seems like most energy is
focused on stopping or blocking spam. What if we facilitated it's delivery?
Sounds crazy right! Please read on.
I got off to pondering about how spam is basically an economic
phenomenon...increase it's cost and it decreases (as evident by anti-spam
results), but more important perhaps that by decreasing it's cost, it (the
actual reading of it) might actually increase and that might be good thing.
In short, my idea is we propose a standard for attaching keywords in the header
of an email, which summarize the targetted content of the email. Think of
bidded search terms for PPC advertising (e.g. Overture.com). Then receivers of
email can easily install filters to prioritize email on content they are
interested in and de-prioritize (delete, delay, relegate to spam folder, etc)
the rest.
The spammers typical motivation is economic. Spammers also compete for
resources between themselves, just like any economic system. If some spammers
make more $ by targetting their spam, they squeeze out economically those who
use inferior methods. Right now, it seems we have the opposite, where spammers
(the inferior targetting) squeezes out the relevant email (even email from
marketers I wish to hear from). Overture proved that highly targetted
advertising tromps other advertising, and PPC continues gooble up advertising
resources across the internet. The difference is that spammers have no
incentive to pay per click, but I will get to issue of reputation capital
below...
It seems to me that what makes spam such a monumental problem is that currently
the spammer has no global way to target his message other than to send a copy
to everyone, as evident by the microscopic response rates to typical spam. If
a spammer could yield a 100 or 1000 times greater response rate by merely
attaching keyboards, it seems to me the spammer would have an economic
incentive to do so.
Then the problem becomes how to prioritize amongst the messages with similar
keywords, and how to detect deception. Basically it becomes a reputation
problem. If we can indeed help spammers make more $ by helping them target
their messages, then my companion idea is build in a trust mechanism that
spammers can opt into. Spammers then can build reputation for their trust
signature. That reputation could be qualified on different levels, such as my
local actions (which message I read/keep and which I discard) and on shared
reputation totals. In short, a system where spammers compete with each other
to find the best keywords and build best reputations for those keywords.
Instead of paying with currency for ranking (ala Overture.com), they pay with
their reputation (ala Google.com).
It seems to me that some people want what spammers are selling. If we help
both parties (sender and receiver) become more efficient, then both parties
have an incentive to opt into that framework.
In terms of creating a global standard, this idea does not require an overhaul
of existing protocols and doesn't suffer from "chicken and egg" delimma, i.e.
is compatible with an evolutionary adoption rate.
This won't completely stop some fringe spammers from attempting to bypass the
reputation system and send untargetted email, but the realities of economics
are that those who provide resources want to maximize profit. Resources are
not infinite, else spam would already be 99.999999999999% of all email. I
think my idea will may spammer resources more expensive, and thus the spammer
who has a choice of sending an ad about viagra to everyone, or sending it only
to people who are interested in viagra, will find it more profitable to enter
the reputation system.
I admit I haven't thought through this idea very far, especially into all the
implementation issues. It is just an idea that bounced off top of my head and
I thought I would share it here, FWIW. No need to flame me if you think it is
bad idea, just let it keep bouncing right off into neverneverland :-)
P.S. The idea has nothing to do with algorithms at AccuSpam, so please do not
involve AccuSpam in the discussion. Thanks.
Shelby Moore
http://accuspam.com
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg