Earlier tonight I was studying up on analytics for Web Software Architecture (class) and I decided to dig into Talk Amongst Yourselves' public analytics. Naturally, this is pretty off-topic, but I started it as an exercise in analytics and it felt like a waste to just delete everything, so I'll leave it here.
First off, the questionable data. Our estimated demographics information may be hopelessly skewed because of our emphasis on anime, Japan, and the like. It may over-label many of our visitors as "Asian" (it says we have 10%) because of this emphasis we have, particularly as of late. Of course, it easily could be true too, I'm not sure. I'm actually going to back the Hispanic data point though. We seem to (especially on the Ani-TAY tag) get quite a bit of through-traffic from South America and Mexico, so I'd guess we have a sizable proportion in the States too.
That's right, TAY people tend to like video games. I'm a bit more concerned with "Adult" showing up though. Come on guys, at least use incognito so you don't store these tracking cookies!
The top shows the traffic analytics as a function of uniques. On a global scale, 79% of the people that visit Talk Amongst Yourselves are just random people that find us online or link in to the site on a non-regular basis. These might be Kotaku main-pagers or Google link followers. These same non-regulars make up 49% of all our page hits. 21% of visitors are "regulars", namely people that seem to return on a normal basis. They make up 46% of our page hits. "<1%" of the viewers are addicts. I'd guess that means people that visit everyday many times (which likely makes me part of this grouping) and they make up 5% of our hits.
It would also seem that 73% of our hits come from computer-based browsers and 27% of them come from mobile sources.
It's important to note that TAY's hits (as you'll see in a second) go through intense jolts, which is why the percentages seem so huge. It's not as bad as it looks though. We get 259,903 global uniques per month and 371,488 visits per month (this equates to 1.62 visits per person every month, a rather good retention rate and just under half of Kotaku's same number).
Similarly, we get 577,159 page views a month globally and 375,411 from the United States. That means about 35% of all our page hits come from elsewhere in the world.
But really, the most interesting part of the analytics is one of the most revealing:
It took some quick research, but there's two events that occurred in the last six months that resulted in more uniques. Bobsplosion's PokeBank article that was cross-posted to Kotaku generated a truly disproportionate number of tay.kotaku.com uniques on Christmas Day. After that, numbers seemed to begin dropping to normal levels again until Bonny John gave his open letter to the Kotaku staff about changing up the roundup styling.
We already knew the new roundups were a success from the get-go. They increased the viewership by making the articles more visible and by adding an image to the post, a sure fire way to attract viewers. These analytics tell us even more. In addition to increasing viewership spikes (we consistently hit high-uniques that were unheard of before the turn of the year), our little sub-blog also has more uniques in general. Our previous average uniques count has risen from relatively low numbers (think 3000 or lower per day) to about 9000 unique viewers per day in this last month.
I could go into far more details here, but this has already gone on a bit too long. I originally was going to keep this short, but I ended going for this long. Analytics are fascinating.
My takeaway from all this is to give a hats off to everyone for their good work and of course to Bonny John for getting the ball rolling on the roundups. I think that had a great deal to do with the uptick in average viewership since that day.
If anyone wants to explore this data yourself, Quantcast provides it free: https://www.quantcast.com/tay.kotaku.com