Wikipedia talk:Wikipedia Signpost/2016-09-06/Traffic report

Discuss this story

Influence of Reddit on Wikipedia pageviews

Reddit is part of the tiny "other referrers" rectangle in this chart

Regarding "Reddit, which bills itself as 'the front page of the Internet' because Wikipedia doesn't, has been a major factor in driving traffic here":

That's probably true if "here" means the Top 25 pages, and in any case there is no doubt that the traffic of an individual page linked from a popular Reddit can spike considerably. (By the way, there is an academic paper about this, which we haven't yet covered in the "recent research" section - if anyone is interested in writing a review, let me know; otherwise I might possibly do it myself in our next issue a month from now.)

However, before we get too excited (or worried) about Reddit's "role in aiding Wikipedia", let's not forget that the top 25 articles receive only a tiny, tiny sliver of Wikipedia's pageviews overall, where the ratio of Reddit referrals is so small that it was not even called out separately in the above chart (from this 2015 research). Of course I absolutely agree that it's worth thinking about how to better draw people's attention to the information on Wikipedia (there has already been quite a bit of work on this by editors, the Foundation and other Wikimedia organizations, but there may be many more opportunities that we have not made use of fully yet).

Regards, Tbayer (WMF) (talk) 03:03, 7 September 2016 (UTC)[reply]

PS: Out of curiosity, I ran the actual numbers (for August 24, the day after the TIL for Tic Tac was posted). On that day, 0.5% of pageviews on the English Wikipedia had a referrer from Reddit - so very much not dominant, although still twice as large as the traffic coming from Facebook. The Tic Tac article itself though had 81% of its pageviews coming directly from Reddit.

Extended content

Data source:

SELECT SUM(IF(referer LIKE '%google.%',1,0))/SUM(1) AS google,
SUM(IF(referer LIKE '%.facebook.com/%',1,0))/SUM(1) AS facebook,
SUM(IF(referer LIKE '%.bing.com/%',1,0))/SUM(1) AS bing,
SUM(IF(referer LIKE '%.yahoo.com/%',1,0))/SUM(1) AS yahoocom,
SUM(IF(referer LIKE '%.reddit.com/%',1,0))/SUM(1) AS reddit,
SUM(1) AS allviews
FROM wmf.webrequest WHERE year = 2016 AND month = 8 AND day = 24
AND is_pageview 
AND agent_type  = 'user'
AND normalized_host.project = 'en'
AND normalized_host.project_class = 'wikipedia';
google	facebook	bing	yahoocom	reddit	allviews
0.3600752916003187	0.0023931338194005474	0.014221534826641381	0.008654790611702571	0.00523757841064597	254312565
1 row selected (396.326 seconds)

SELECT SUM(IF(referer LIKE '%.reddit.com/%',1,0)) AS redditrefs,
SUM(1) AS allviews
FROM wmf.webrequest WHERE year = 2016 AND month = 8 AND day = 24
AND agent_type  = 'user'
AND normalized_host.project = 'en'
AND normalized_host.project_class = 'wikipedia'
AND pageview_info['page_title'] = 'Tic_Tac';
redditrefs	allviews
263560	324843
1 row selected (214.299 seconds)

The webrequest database used in this query is not public; for privacy reasons only some WMF staff under an NDA have access. But if someone is interested in related data and takes it on themselves to modify the above queries accordingly, I might be able to run it for them if they post a request as described here.

Regards, Tbayer (WMF) (talk) 07:21, 8 September 2016 (UTC)[reply]

Tbayer (WMF), that paper does look interesting, I'll give it closer read and see if I can't write something up about it. And yes you are correct that the commentary was referring to articles making the Top 25. Redditors on TIL clearly use wikipedia's vast store of articles to find interesting things to highlight, and thus random articles will make the Top 25 as a result. I don't think this is bad, if anything it probably influences why reddit seems to like wikipedia so much.--Milowent • ^hasspoken 03:08, 10 September 2016 (UTC)[reply]

Can we talk about how the Signpost allowed a biased and factually-incorrect rant be published for all the world to see? Also, wouldn't Wikipedia trying to push popular articles force the site to cater to clicks and go down the road of clickbait that has befallen many other websites? 24.113.234.93 (talk) 20:59, 8 September 2016 (UTC)[reply]

@24.113.234.93: No and no. The folks that do this report always do a bang-up job. You're welcome. Chris Troutman (talk) 01:51, 10 September 2016 (UTC)[reply]

The commentary above that you complain about is not really about articles that make the Top 25, but about other well known issues with Reddit. I am on reddit and aware of these issues, as are many of us. Don't take it personally.--Milowent • ^hasspoken 03:01, 10 September 2016 (UTC)[reply]