How Much of the Internet is Actually for Porn

Starting September 7, .XXX domains are for sale. (That is, if you own a porn site or a trademark to protect – the rest of us have to wait till December 6 and hope nobody scoops our name first).

It’s like the internet is growing up a bit. There’s certainly enough porn to justify .XXX, if .biz gets its own extension.

But the arrival of .XXX begs the question:

How much of the internet is actually for porn?

Plug the question into Google and you reach an abandoned corner of the net: an article from 2005, a mess of answers from 37-80% and the Avenue Q Wikipedia page.

Surely there must be a better way.

So I asked Ogi Ogas, one of the amazingly nerdy neuroscientists behind A Billion Wicked Thoughts, who says he and co-author Sai Gaddam are sitting on what they think is “the most comprehensive collection of porn-use stats on the web.”

(At least, till they stopped counting  last July. Now they’re busy with their book tour.)

So Ogi: How much of the Internet is actually for porn?

There are a couple ways of thinking about the proportion of the Internet that is porn:

●      In 2010, out of the million most popular (most trafficked) websites in the world, 42,337 were sex-related sites. That’s about 4% of sites.

●      From July 2009 to July 2010, about 13% of Web searches were for erotic content.
Both of these are from our research in Billion Wicked Thoughts. We consider our data the best available. It’s an impossible task to say exactly what % of *ALL* websites are pornographic or anything else, because the web is both so enormous and so dynamic; looking at the million most popular sites is a very reasonable sample.

You could also look at the number of “adult sites” that are blocked by various parental filtering software programs–for example, CYBERsitter claims to block 2.5 million adult Web sites–though it’s not clear how they’re deciding a site is an adult site; the sites may not necessarily be sexual sites, and they may be exaggerating their numbers.

Only 4% of the top million websites? That feels… underwhelming. I was thinking at least half.

There have been a bunch of false and ultimately mythic stats floating around for years that say half the Internet is porn or one third of the Internet is porn, though this has never been remotely true.

In the early days of the Web, when the vast majority of users were guys, there was a much greater proportion of searches for porn–I think in 1999 that 4 or 5 of the top 10 searches on the Web were for porn–though the % of websites that were porn may actually have been lower than today.

Web filtering companies used to always release competing figures on the number of porn sites they blocked, but these numbers were almost certainly boosted to get sensationalist headlines  and to seem competitive with other filtering companies that filtered “less” adult sites. For example, N2H2 claimed there were 260 million porn sites (ed. correction: pages) –haha, one for every American citizen! :) Conservative groups are always coming up with porn figures that are crazy high, too, especially with regard to children’s exposure to porn.

You and Sai Gaddam are sitting on arguably the largest collection of porn data in the history of mankind.  Was this the first burning question on your minds?

This was one of the first questions we tackled while working on Billion Wicked Thoughts. We did our best to locate every previous measure and quickly realized that most available stats were completely fabricated or bogus.

The only systematic scientific attempt to determine how much of the Web was porn was Berkeley professor of statistics Philip Stark’s 2006 study carried out at the behest of the US Department of Justice under Bush.

Stark found that about 6% of searches were for sexual content – he must have used crazy stringent definitions of “sexual content” because all the numbers we found in our own data sets was higher. Stark also found that about 1.5% of all web sites were porn sites. We took issue with his method of random sampling sites; we looked at the million most popular websites in the world, since we figured that would give a more realistic indication of a random person’s experience on the Internet.

So how do you go about measuring how much of the internet is for porn?

I think the two best metrics for analyzing the level of human interest in sex on the Internet is: (1) how often do people freely search for it and (2) what amount of web traffic goes to porn sites.

The best way to evaluate (1) is by counting porn searches on Internet search engines; the numbers are usually from 10-15% these days (higher as you go back in time to more men and less women online).

The best way to evaluate (2) is by looking at the number of sites out of the million most highly trafficked websites and seeing how many are porn sites (about 4% according to us, higher than Philip Stark’s random-sample 1.5%).

You could even count the monthly traffic to all of these porn sites, which I admit we didn’t do, though it would be pretty straightforward. In our book we have a table of porn traffic to the five most popular porn sites, varying from about 7 million to 16 million visitors a month.

What about porn downloads? Isn’t that a lot to account for?

The only reliable source I know of that can measure “porn downloads” is Nielsen, but I’m not sure how they measure this. We actually spent some time trying to get Nielsen to share data with us, but they never did, so I have to be skeptical of their calculations. I’m sure they’re not considering torrents or other P2P sources of downloads.