Delicious did in the BILLIONS of monthly PVs when I left - I really doubt HN throws more traffic in aggregate.
HN gets about 60k unique visitors on weekdays. We used to have about 750k pageviews on weekdays, but I just cracked down on crawlers that were killing performance, and the number is closer to 650k now. I've heard from several people that HN is a source of referral traffic out of proportion to its size, presumably because its concentration means more articles are interesting to more readers. But it would have to be a traffic source way out of proportion to its size to get the kind of numbers they're quoting in this study.
That's attributable to the fact that I'm active in this community, so I was surprised to see it accounting for so much traffic on a larger sample, but there's certainly evidence that even links that get no votes and no comments still drive traffic (though not as much as links that add value and create discussion).
Look at the numbers, HN is a top 3,000 site, reddit is a top 300 site.
A huge chunk of reddit links are back to reddit, to imgur and to things like major newspapers and youtube. If only a few of these are not actually tracked by Woopra, the results would get skewed significantly.
Top-level front-page reddit is pushing 100k.
Here's an example of the traffic of a recent article that got both HN and proggit: http://imgur.com/TpQ5T Obviously, n=1 here, but this pattern repeats all the time.
I mean, how is it useful as bookmarks, social or otherwise? Are you able to save something so you can return to it later?
There are lots of bookmark lets that have nothing to do with bookmarking. It's just a way to execute some JavaScript in the context of a page.
Most of the sites there aren't social bookmarking, either. SU has the functionality, though.
For what it's worth, I'm the one who coined the term.
I know that I personally hardly read it anymore, but I used to read it many times per day.
So it doesn't drive as much traffic as you might expect, but when your site does get mentioned on /. it gets much more traffic than a HN posting.
Also, there probably is very little overlap between the typical /. covered site and the sites that woopra covers, /. is not really a start-up related site.
If you were to re-run the numbers with all of google analytics covered sites (impossible to get the data, unfortunately) you'd likely get a different picture.
The barrier to entry for a successful HN posting is much lower than it is on /., the volume of 'posted' stories on /. is much lower too.
All in all less surprising than you'd think at first sight.
Regarding the seeming discrepancy in data for Google's share, the reasons are probably: (i) more of Bing/Yahoo's traffic stays within their portals than Google's does, and (ii) Google is especially dominant among tech-savvy users, and the websites that were analyzed have proportionally more of such users (which would also explain why HN shows up).
