While I try to keep my posts mostly related to learning Japanese or translations, because of the amount of time I spend thinking about blog statistics I like to write articles about stats once in a while.
By no means do I consider myself an expert at statistics, but I know they are a tricky double-edged sword that can alternate from providing meaningless, even counter-productive information to critical insights that can be used to improve a variety of processes and results. It depends on what assumptions you have, how you explain them, how you analyze them, and how you use them.
While I, like most bloggers out there, value the number of hits to articles, there are limitations to this metric. One of them is the (apparently) large number of bots related to spam, search, and other things. While WordPress (and related plugins) uses some methods to try and detect these, they are not perfect. Furthermore, I have seen issues with inconsistent statistics and other anomalies which throw more doubt on the accuracy of these numbers.
Another statistic which is useful is the number of likes, but and while I think the raw count of these may be more accurate than page views (there is less of a reason for a bot to artificially like posts, though I can think of one or two), the interpretation is even more murky. Personally I know when I ‘like’ someone’s blog it can mean everything from “this post was great”, to “this post was so-so but I want to encourage the poster to write more on the same topic.” Also, I think many people who generally surf the web do not have wordpress accounts and hence cannot ‘like’ posts.
So, given that neither of these numbers is an ideal indicator to judge quality and interest of posts, I am always looking to other sources of information.
One nice piece of data is the number of people who have joined reading lists on sites like Novel Updates, These show that there is interest in reading future chapters of a specific story. While this metric only applies to translations and not informational articles like the one you are reading now, the great thing about them is that they are selfish, in the sense that when someone adds a story to their reading list it is clearly for their own benefit. Also, since there is no link back to these people’s names (at least no obvious one I know of), there is little reason to add a story to a reading list for unrelated things like “hey, check out my blog!”. So I feel these such statistics are more pure, and hence have potentially more value as compared to metrics such as likes and raw page hits. They show that, without a doubt, someone read and enjoyed a post.
Of course, there are other elements which effect reading list statistics. Using Novel Updates again as an example, the number of translations posted around the same time will effect how long a certain translation stays near the top of the list, not to mention the quality of the synopsis. Also, a story that is popular should be seen in the context of the demographics of the site in question. I tend to choose works that are a little less ‘light’ (more serious), which I believe is a little different than the average story on that site. Nevertheless, I am happy that some people there have expressed interest in reading a few of my translations.
Once you have total hits per article plus the number of new reading list additions you can even calculate derived stats, like percentage of people who want to read more after actually reading that chapter. This re-introduces the uncertainty of page hit count, but has the potential to help factor out some elements such as the quality of the synopsis. When I compare my translated stories, sometimes I will divide the total number of people on its reading list by the number of chapters posted so far in order to smooth out various random factors.