Hello,
I watch pageview traffic to about 100 popular Wikipedia health articles. In January 2014 as compared to January 2013, stats.grok.se reports about a 50% drop in pageviews.
Does someone have a ready explanation for this? The obvious explanation would be that Wikipedia's traffic has halved, which does not seem right to me, but it could be so.
Can someone direct me to previous or current conversations about this?
yours,
Lane Rasberry, 06/03/2014 20:23:
I watch pageview traffic to about 100 popular Wikipedia health articles. In January 2014 as compared to January 2013, stats.grok.se http://stats.grok.se reports about a 50% drop in pageviews.
Monitor how? Do you manage to use http://tools.wmflabs.org/glamtools/treeviews/ despite https://bugzilla.wikimedia.org/show_bug.cgi?id=42259 ?
Does someone have a ready explanation for this? The obvious explanation would be that Wikipedia's traffic has halved, which does not seem right to me, but it could be so.
Can someone direct me to previous or current conversations about this?
Of course for general stats you can always see http://stats.wikimedia.org/EN/TablesPageViewsMonthlyCombined.htm and there was some research lately https://meta.wikimedia.org/wiki/File:2013_Wikimedia_traffic_trends.pdf . The rest is just speculation; in particular when looking at narrow portions of the traffic, like a small subset of pages or a "small" project, the noise (i.e. bots) is often more than the signal, so you can only trust multi-year trends.
Nemo
Hi,
On Thu, Mar 06, 2014 at 09:37:47PM +0100, Federico Leva (Nemo) wrote:
Lane Rasberry, 06/03/2014 20:23:
I watch pageview traffic to about 100 popular Wikipedia health articles. In January 2014 as compared to January 2013, stats.grok.se http://stats.grok.se reports about a 50% drop in pageviews.
[...]
Of course for general stats you can always see http://stats.wikimedia.org/EN/TablesPageViewsMonthlyCombined.htm and there was some research lately https://meta.wikimedia.org/wiki/File:2013_Wikimedia_traffic_trends.pdf . The rest is just speculation; [...]
Full ACK.
But as Lane explicitly mentioned stats.grok.se ... Parts of the overall traffic seems to be shifting to mobile. However, as far as I know stats.grok.se's per page numbers do not contain mobile traffic. That's not stats.grok.se's fault, but ours: We do not have per page numbers for mobile traffic. Nevertheless, it explains at least part of the drops that you might see on per page numbers of stats.grok.se.
As specifically January 2014 data got compared: There were hardware problems on our end during the first weekend of January 2014, which results in no data being available for 2014-01-05 and 2014-01-06 on stats.grok.se. That explains another ~5% drop between January 2013->2014.
Best regards, Christian