Blog: 
Jan 2008
Dec
5% of Email is Not Spam
Nov
Oct
According to the editors over at ArsTechnica, an amazing
95% of email is spam. The numbers are a little suspect, since they come from
the spam fighting industry, but nevertheless they should depress anyone. Nineteen out of every
20 messages sent to you are pure junk. The rest may be from legitimate people you just don't like.
100 messages. All but five are junk.
The Current Problem: No Incremental Cost
Junk mail is an issue in the real world, but it's constrained by economics.
It does cost something to send a flyer to a potential customer, so not every
business will advertise via direct mail. We all encounter this during the holiday season,
when we send greetings by the postal service to everyone we know. Each card requires
a stamp; the more friends and family you have, the more you must spend on postage.
Email is not bound by incremental cost. Sending a generic electronic greeting to multiple
people on different continents does not require extra stickers, just a few extra keystrokes in
the CC field. It costs about the same to send an email to one person as it does
to a dozen---or a thousand, or a million.
The spam purveyor's logic takes off from here. Why target the campaign to a specific demographic when
you can hit everyone? No matter how small the rate of return, the spam economy will thrive.
A tiny percentage of success against a huge user base undoubtedly results in profit.
The Current Solution: Identify and Block
Countless experts spend their entire workday battling spam. While the details are insanely complex,
the principle is simple: analyze messages to determine if they are spam, and if so, prevent them from
reaching an inbox. And while this mostly works, we all pay the price. The cost of running the world's
email systems is roughly 20 times what it should be, if it weren't for junk mail.
The Future Problem: Less Obvious Spam
Right now, any human being can tell a spam message from a real one upon inspection, and computer
programs are therefore extremely accurate. But what happens when spam gets really sneaky,
when it starts to look like messages from our friends and co-workers? Suddenly the old identify and
block system won't work anymore, because we ourselves cannot tell the difference. Junk mail and
real mail will look the same.
The Future Solution: Cost and Identity
Someday, the spam problem will be solved, probably through a combination of technologies that mirror
the real world. There should be an incremental cost to sending unsolicited email. No one has a legitimate
reason to send bulk messages to people they do not know, and doing so should be much more expensive
than a single friendly email to a stranger. Second, you should not be able to send messages anonymously,
at least not as a standard practice. Communication must have a sensible cost structure and establish
the identity of the persons involved.
In practice implementing this solution is extremely difficult. The Internet is highly
decentralized, so there is no single gatekeeper who can enforce rules and levy fees. The protocols
for transmitting information between networks are designed for flexibility and reliability, not
security and authentication. It is hard to make a wholesale change to a system designed, on
purpose, without significant barriers to entry.
The greatest problem, though, is establishing identity. A great actor dressed in an impeccable
costume could not fool you into believing they are your mother. Conversely, anyone sitting
at your mom's computer, logged into her account could easily dupe you. Thus the humorous adage:
on the Internet, no one knows
you're a dog.
More Spam to Come
If only 5% of email is not spam, it is hard to imagine it getting worse. When does the cost of fighting junk
mail overwhelm the value of mail altogether? If we include
comment spam and the rise of
phishing websites, it seems the whole Internet is
being overrun with bogus data. Is there a singularity we are about to cross beyond which sharing
information just isn't worth it anymore? I believe so. Spam has got to go.
Further Reading:
Would you like to leave a comment?
Read this.
###
