Being mindful of our site metrics, I was puzzled why our viewcount seemed so low given our total number of site views:
For example, on 5th Feb we had 137 visitors per day and the site has been active for 17 days.
If you take the incubation and leg-jiggling questions, we have around 3,200 views, that makes 188 views per day (3200/17) on their own.
Apparently, as econ.se found out, visitors per day is calculated as the median visitors per day over the previous two weeks.
I.e., the median of 10 days of 100 visitors and 4 days of 2000 visitors is still 100, whereas the mean is 643.
My basic mental model of site traffic goes like this:
When a site is new, most of its traffic comes from users. However, if links are shared, viral spikes can be much greater than the background hum. That said, such spikes often have a short half-life or around one day in the case of Reddit.
Traffic from Google grows gradually as content increases and Google learns that the answers provided are valued by searchers.
Also, if a site is well-established, even viral spikes provide a minimal effect to the background hum.
Unless there are more than around 4 or 5 viral spikes a fortnight, they are not likely to have a huge effect on visits per day statistics. And even then, once their effect is over, the site statistic will return to users + google. So, if we want to grow our views per day statistic, we need to increase the amount of great questions and answers indexed by google, and grow our number of users asking and answering questions. A quick glance at this page listing questions and views per day by site suggests to me that views per day is often somewhere between one and two times the number of questions on the site.
That said, there are many great reasons to share links to great questions and answers with communities that enjoy the content. Beyond altruistic reasons for sharing great content, it seems to be one of the best ways to grow the user base.
