How much of all Internet traffic is pornography?
How much of all Internet traffic is pornography? I'm talking Web pages, peer-to-peer transfers, and so on--the whole kit and caboodle. My friend claims that by far the majority of Internet traffic is porn. I'd say that while it makes up a large percentage, it's probably less than half. Who's right?
Not an easy question to answer. But that's why people write to me.
First let's dispense with the myth your buddy is flogging, namely that most online traffic is smut. This belief stems from a 1995 study published in the Georgetown Law Journal by Martin Rimm, a 30-ish electrical engineering student at Carnegie Mellon University. Rimm and his colleagues studied X-rated images on the alt.binaries forums of Usenet. Apparent conclusion: 83.5 percent of what was out there was porn. Rimm, no fool, gave Time an advance peek at his article, and the magazine splashed the story on its cover with the war-type headline CYBERPORN.
Livid critics went to town, pointing out so many flaws in Rimm's work that Time printed a partial retraction. No need to rehash the arguments--they're meaningless now. Usenet? Binaries? More important, U.S. "backbone" Internet traffic grew from 16 terabytes a month in 1994 to 80,000-140,000 terabytes a month in 2002. Even if Rimm's claims were spot-on in 1995 (which they weren't), today they wouldn't mean squat.
So what percentage of Internet traffic is smut now? No idea. Among the wild cards: Untrackable peer-to-peer file sharing, made famous by Napster but equally suited to porn exchange.
Come now, you say. Even if we can't put a number on it, surely we can get a handle on Internet smut volume. The closest thing the Web has to a central registry is the databases amassed by search engines such as Google and Yahoo, which have cataloged billions of Web pages. A few nanoseconds of googling and we'll have a ballpark idea how much porn's out there, right?
Sorta. The Internet traffic equation has two sides, supply and demand. Supply is easy to gauge. Google "sex" and we come up with 214 million pages--an impressive number, but as of September 20, 2005, Google had cataloged 8.2 billion pages. What's more, we're casting a wide net--"sex" takes in pages about sex discrimination and other sober subjects. Better to pick the keywords pornmongers use to advertise their wares. "Porn" (duh) summons 35 million pages, "XXX" 32 million--numbers so close you can't help but think we've hit pay dirt. Conclusion: Four-tenths of one percent of searchable Web pages consist of hard-core porn.
But that's just what's for sale. How many are buying? To get an idea we need to know the most popular search terms used by Web surfers. That's not as easy as it sounds. While many search engine sites publish "hot search" lists, most such collections have been pruned of potentially offensive terms. Lycos, for instance, claims that the most popular search term of the past ten years is "Pamela Anderson." Get outta here. Nosing about, we discovered a list of the top 20 search terms as of 1999-2001 compiled by the Web research firm Alexa. The number one term, "sex," was used in 0.3 percent of searches, or one in 300. Number four was "porn," number thirteen was "nudes," number fifteen was "XXX," and number nineteen was "Playboy."
That's more like it. But in Internet terms, 2001 is the equivalent of the French and Indian War. What about now? Since nobody's volunteering, I've taken matters into my own hands. Overture, a Yahoo subsidiary, has a search tool that tells how many times a given word or phrase was requested in the previous month, in this case August 2005. (To be precise, Overture lists the 100 most frequent search requests in which the term appears.) Plugging in the terms on the Alexa list plus all the other likely candidates I could think of, I came up with the following big--I'm not saying top--ten: (1) music, 39 million; (2) Google, 37 million; (3) Yahoo, 35 million; (4) travel, 33 million; (5) free, 29 million (includes "free porn" and "free sex" plus many innocuous variants); (6) sex, 26 million (includes "teen sex," "cyber sex," etc); (7) games, 24 million; (8) Internet, 19 million (actually, the lead term in this category is "Internet Explorer"); (9) eBay, 17 million; (10) download, 16 million.
Discounting "Google," "Yahoo," "Internet," and "download" as mere means to an end and "free" as too diffuse, we find that the principal destinations on the Web are music, travel, sex, games, and eBay. Proof of porn's preeminence? My arse. If we consider "eBay" a stand-in for "treasured possessions" and chalk up the absence of "food" to the inability, so far, to digitize taste and smell, I'd say we've got an unremarkable list of life's little pleasures, whether online or off.