On Monday morning I added a link on HackerNews to a recent post on this blog. Happily, a short while later it was on the front page, where it sat in the top 30 for about a day and a half until it dropped off. Using Google Analytics, its possible to do a little analysis of what kind of traffic you can reasonably expect from a popular HN submission. This blog gets only a handful of visitors on a normal day, so the results sure be mostly pure.
It’s important to note upfront that it’s hard to draw too many conclusions from the data. What time you post, what day of the week, the provocativeness of the link title, how high it gets in the rankings, how long it sits there, the comment to point ratio, your nick’s reputation, and a lot more could drastically change how many people click the link and make it to your site.
Without further adieu, here are a few traffic stats for this blog from Monday and Tuesday:
All Traffic Sources
Top Content
(the rest were irrelevant)
Region
And for those interested in California…
Browser and Speed
So…
- San Fransisco, Los Angeles, New York City, and Cambridge are good places to be. 40% of the visits were from outside the United States; Toronto, Bangalore, London, and Sydney were the popular foreign locations.
- news.ycombinator.org?
- It’s not clear where all the direct traffic is coming from. I thought direct traffic was made up of people typing in the URL, which wouldn’t be reasonable because it was so long.
- 24K uniques/day? I’m not sure how to interpret the numbers. It would seem that only about 10% of the visitors to HackerNews clicked on the link to this blog. It’s actually even less because these numbers are over a two day period whereas the 24K number reflects HackerNews’s traffic in a single day. As noted above, a lot of factors could change the click through rate, but I’d still expect it to be a bit higher.
- Why is it under Top Content the article has an average time of 7:33 (about what you’d expect for a post of this length), but under All Traffic Sources it says that the average time on the site was 55 seconds? I wish Google let you view the distribution so you could see, for example, how many people were on the site for less than a minute, how many 1-2 minutes, etc.
- People still use Dialup and Internet Explorer?
If I missed any other stats that you’d like to see, let me know.
Matt Mazur 