Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Speaking as a former reddit admin who had direct access to the server logs (and "wc -l"): Quantcast (and Alexa, and Comscore, and all the others) are terrible at estimating traffic. They make wildly inaccurate guesses using low-quality source data and all the ad execs just gobble up their results as if they're fresh off God's own LaserJet. It's very frustrating to watch.

The most accurate results are visible at Google AdPlanner, because it's reporting actual Google Analytics data from a bug embedded on every reddit page:

https://www.google.com/adplanner/planning/site_profile?hl=en...

Google Trends is also pretty good -- again, because they have access to actual, real search data which they're presenting raw and unvarnished:

http://www.google.com/trends?q=reddit%2C+digg&ctab=0&...

To my eye, the hockey stick begins in Q1 2010, well before the mid-Q3 spike when Digg v4 was launched.

TLDR: Digg's implosion may have accelerated Reddit's traffic growth, but it was already doubling yearly, a trend which goes back to the site's launch in 2005.



Thanks for this site, I never knew Google publish analytic traffic and their estimated traffic.

Something interesting I have noticed, if you are looking at Hacker News' traffic data, there seems to be a steep decline somewhere around April [1]. I wonder what whether it is correct, and whether there was a fundamental change in this site / ycombinator around that time.

[1] https://www.google.com/adplanner/planning/site_profile?hl=en...


HN either doesn't use Google Analytics or didn't check the box to make their traffic numbers public. That's why the graph you link to has a dotted line, rather than a solid one. The accuracy is much less certain in their case.


(Sarcasm noted :-) For the uninitiated, this was around the time comment scores were hidden. Only PG knows whether this had any real effect on traffic; and even if so, whether that's a good or bad thing is subjective (less poor quality traffic / comments is good, and since HN is not directly commercial, lower traffic is not necessarily a bad thing).

Loss of value to everyone else from lack of comment scores is another question entirely...


If we have that much evidence that having no visible comment scores reduces poor quality traffic, I think I may have just switched sides in that particular debate.

I'm not saying that is the case, but it could be a very interesting revelation.


And in case it helps, I was about to write the exact same comment. They are using Alexa numbers which are, basically, unsubstantiated guesses.

Heck, they didn't even bother to read the numbers reddit actually publishes.


Any thoughts on why such a difference in the search volume graph which you posted, and Google's estimate of actual traffic? http://trends.google.com/websites?q=reddit.com%2C+digg.com&#...


One's based on raw, direct data, the other is an estimate based on imperfect, partially-related data.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: