Wikipedia talk:Pageviews and primary topics

Twice as many views prevents primary topic?

@Uanfala: Did you mean to write that if a page receives only twice as many views, then this is a clear signal that it is not the primary topic? I think that's a controversial opinion (which, of course, is perfectly proper in an essay). The page can still be much more likely than any other single topic, and more likely than all the other topics combined—to be the topic sought, making it a WP:PTOPIC. Absence of evidence might be a more plausible claim than evidence of absence. Moving the not, we might agree that "this is not a clear signal that it is the primary topic". Certes (talk) 20:49, 20 April 2020 (UTC)

I guess I should clarify that this applies to cases of an article already at the primary topic vs. the corresponding dab page. Let's say the article is Foo, it receives 100 views a day and it's got a hatnote to Foo (disambiguation), which receives 50 (This is an almost impossible distribution, for reasons that will become obvious by the end, but let's just assume it). Usually, most of the traffic for Foo will be coming from links, but let's assume there are no incoming links at all, and all 100 views come from readers who've typed "Foo" in the search box. If the dab page receives 50 views, this means that of those 100 readers, 50 have found what they had been looking for and stayed at Foo, whereas the other 50 have not been happy with Foo and so have clicked through to the dab page. This means that only half of people searching for "foo" are looking for Foo, so that article straight out fails the primary topic criterion of being more likely than all the other topics combined[..] to be the topic sought when a reader searches for that term".

Of course, the simplifying assumptions made in this example are that the dab page won't receive significant traffic coming from elsewhere (dab pages like this one don't normally have incoming links other than from other dab pages or "Distinguish" hatnotes at other articles, so this assumption is reasonable most of the time), and that readers who've been happy with Foo won't have gone on to visit Foo (disambiguation) (there are readers who will follow that link out of curiosity, but I can't imagine their number being significant unless the article is very prominent and the dab page extremely obscure).

Now, if you take into account views coming from links, then the situation moves even further away from there being a primary topic – say, half of the 100 views of Foo come from links, this means that only 50 views come from searches. But that's the same number as the dab page, so – subject to the limiting assumptions – all of the people who search for "foo" click through to the dab page, and the article at Foo was not what any single one of them was looking for!

So, the heavier the proportion of traffic coming from links, the greater the ratio will need to be for a primary topic. Obviously this varies across articles, but my experience has been that there's usually somewhere around an order-of-magnitude difference. Whenever I've seen articles like Foo with less than, say, 8x as many views as Foo (disambiguation), and have started RMs to move the dab page over the primary title, the RM's have almost always been successful (though largely for unrelated fundamental reasons, I don't normally go into that sort of technical analysis when proposing moves).

For situations that don't involve dab pages, but merely the two articles Foo and Foo (bar), then of course things are different – excluding traffic coming from searches, we've got one article receiving 2x as many views as the other, and of course that's open to debate, the central point being whether 2x counts as much more likely or not. – Uanfala (talk) 21:31, 20 April 2020 (UTC)

Ah, I'm with you now. So, taking a lucky dip in Special:RandomInCategory/All disambiguation pages until I got some called Foo (disambiguation), the primary redirect F150 appears to be on dodgy ground but Kinn looks safe. That makes more sense now that I understand it. Certes (talk) 22:15, 20 April 2020 (UTC)

See, on some days F150 (disambiguation) has had more views than F150, which suggests that not all of the traffic of the dab page is coming from searches. This sort of distortion is common when very prominent articles (F150 redirects to Ford F-Series) have hatnotes for very obscure dab pages. – Uanfala (talk) 22:34, 20 April 2020 (UTC)

Views by editors

Events which can cause spikes include... opening a Wikipedia discussion. Very true, especially for obscure pages. We have statistics on mobile vs desktop views. Do we have anything on logged-in editors vs anonymous visits? Although registered editors can consult WP whilst logged in and a few IPs contribute to RMs, that would still be useful for ironing out such spikes. Certes (talk) 10:12, 9 May 2020 (UTC)