Hacker News Reader

Comment by angersock | Hacker News Reader

I'm rather surprised--which is to say, calling bullshit--that no porn sites (any of the $TUBE network) shows up on this.

View on HN

Probably because most of the data comes from comScore who, like Alexa, gets their data from a self-selected sample size. We're ever going to get truly accurate results this way.

Of course, I'm sure if the NSA put its mind to it (not happening), they could get a fairly accurate readout of the most popular sites for the US...

I wonder how you'd collect that...log every DNS lookup at the ISP level? Surely they already do this.

Probably count tcp/http connection to sites at ISP? Of course it's not easy or accurate for us. Not NSA ;)

Another option for conspiracy theorists is to just get data from apache logs etc., because, you know, NSA has access to everything.

I don't have an answer for you, but a simple DNS lookup at the ISP wouldn't get you the whole picture. Lot's of local/LAN DNS caching.

You don't need the NSA to get that info. There are plenty of companies who aggregate data from ISPs.

Really? Does that data get made available in market reports?

It seems you've overlooked "Penthouse" near the bottom of the list in 1996.

It's likely they clean these sites out. I know a company doing a similar report did this. In reality 3 or 4 of the top 10 sites were porn.

Also I found interesting gay porn seemed to be very over represented on mobile internet, about 20%-ish of top porn sites. I found this interesting as in a male world you would expect this to be about 10% so it seemed either gay males watch much more porn or as I assumed, guys would have a look on their phones as this was more private to have a curiosity look...??

Reply on news.ycombinator.com