Talk:Rasna

post-move[edit]

I disambiguated this after noticing it in January. The monthly page views[1] indicate the views for the soft drink have recovered to the old levels, and the clickstreams[2] now show:

With 321 views in February:

clickstream-enwiki-2024-02.tsv:

Rasna Rasna_(drink) link 67 (~20.9%)
Rasna Etruscan_civilization link 14 (~4.4%)
total: 81

With 353 views in March:

clickstream-enwiki-2024-03.tsv:

Rasna Rasna_(drink) link 82 (~23.2%)
Rasna Etruscan_civilization link 18 (~5.1%)
total: 100

Given the anonymization threshold it's possible we're missing some, both for the previously presumed primary topic and for other topics. Either way it seems this topic does not command the reader interest indicative of a primary topic by usage. --Joy (talk) 16:30, 7 April 2024 (UTC)[reply]

I don't follow your reasoning at all. All the clickstreams seem to indicate is that for many readers, the info that "Rasna" is an Indian soft drink, is sufficient. Which they learnt from the pre-move page as well. But looking at the page views, which dropped from 47 a day in December to 11 a day now, and seeing that Rasna (drink) has far more pgeviews than this page (or any of the other targets), it's hard to see how it is not the primary target[3]. I'm wondering what kind of numbers you would like to see to consider something to be the primary topic, if +80% of the outgoing pageviews isn't sufficient. Fram (talk) 09:05, 9 April 2024 (UTC)[reply]

@Fram the fundamental reasoning is: 70-80 readers a month is not many, but it's somewhat far from the anonymization threshold (<10 per source-destination pair) that we can assume that it's mostly accurate. Yet, compared to the total number of viewers of 320-360/month, it is comparatively small. If 300 people see a short list where the most popular item is at the very top, and in turn we find ourselves having no reason to believe that 200 wanted to click it, nor that they were somehow confused by our formatting or something like that, how would that really qualify as primary topic?

If we escalate that topic from the position of most popular to the position of primary, we risk astonishing a lot of those readers. Certainly not all of them, because it's statistically possible that some will necessarily be confused for reasons we can't grasp, some made a typo, some were just casually browsing and never wanted to read through, etc. But even if we try to liberally remove all those contingents of readers, it's still likely to be a favorable comparison with the ~20-25% who went for the drink because it starts from as high up as ~75-80%.

Likewise, the number of identified clicks for the drink is also not that larger than the 10-20 readers a month we know were looking for the Etruscans - at the same time those numbers are closer to the threshold so contingents of those viewers could be missing (IOW it could actually be e.g. 20-30 and we'd never know). That topic has kind of obvious long-term significance, and I don't see an inherent reason to disregard that in favor of on another topic based on just a 2:1 or 3:1 traffic comparison between them.

And in turn, we're not seeing the long tail - there could be another 5-15 people every month looking for a third or a fourth topic, another 5-10 for another couple, etc, but we're never going to see these in the clickstreams.

Note that I've been collecting data for a while now on how readers behave when presented with primary topics, primary redirects, and disambiguation lists with one very popular topic, it's in a few threads at WT:D now. I'd appreciate it if you could have a look and see if you can see any comparable cases that we could use to try to understand this better. --Joy (talk) 09:50, 10 April 2024 (UTC)[reply]

"Likewise, the number of identified clicks for the drink is also not that larger than the 10-20 readers a month we know were looking for the Etruscans" It's 5 times larger, that's quite a lot in my book. We are now inconveniencing 70 readers a month (who need to clickthrough to get to the drinks article) for the benefit of a few. I'll probably just move this back tothe previous situation, as this seems a clear case of WP:PT1 with 80% of the readers for the drink. Fram (talk) 10:14, 10 April 2024 (UTC)[reply]

But that's ignoring the fact that the lower the numbers go, the harder it is to avoid the anonymization threshold, therefore all the low numbers that we see are of dubious reliablity. You can't have a strict comparison between a variable that is clear and one that is unclear.

I would like to see any proof that these 70-odd readers are actually inconvenienced by having to click the first item in a short list to get to a topic of unclear long-term significance - in a sense, a citation is needed on that assertion, this is not clear at all.

In either case, if you want to move things further, please use the WP:RM process or WP:RFC or similar, so we get more community input. --Joy (talk) 16:20, 10 April 2024 (UTC)[reply]

I just moved it back to the long-term situation, it is up to you to use the processes if you want the status quo to be changed. Fram (talk) 16:30, 10 April 2024 (UTC)[reply]