Our referral traffic for 10/13 - 10/14 months follows. It's heavily weighted towards seach. There are some industry wide issues about identifying referrals from Facebook and Twitter's mobile apps which probably underestimate the numbers but it seems like we have some room to improve here.
According to this report from Buzzfeed[1], social and search referrals (the way the industry thinks about referrals) are about even (around a third of traffic each) in 2014. When I was at Yahoo a few years ago, social was just starting to approach search so this seems reasonable.
-Toby
Wikipedia Referral Traffic (10/13 - 10/14)[2]
Other
84,951,586,000.00
78,603,395,000.00
Internal
73,696,896,000.00
Yahoo
4,756,204,000.00
Yandex
1,951,328,000.00
Bing
1,939,456,000.00
Baidu
569,554,000.00
328,492,000.00
Naver
319,629,000.00
Ask
271,658,000.00
260,840,000.00
190,642,000.00
Sogou
101,626,000.00
Seznam
97,720,000.00
DuckDuckGo
80,658,000.00
Daum
69,569,000.00
AOL
67,526,000.00
Startpage
56,814,000.00
[1] http://insights.buzzfeed.com/industry-trends-2014/ [2] data assembled for readership report, 12/14 from http://pentaho.wmflabs.org/pentaho/Home
On Fri, Jan 9, 2015 at 3:26 PM, Jared Zimmerman < jared.zimmerman@wikimedia.org> wrote:
I'd be really interested knowing how our inbound referral traffic from social sites differs from that from inbound traffic to social and news sites from social referral traffic. When we talk about reader decline, we rarely talk about how much a small increase in social referrals could offset that.
*Jared Zimmerman * \ Director of User Experience \ Wikimedia Foundation
M +1 415 609 4043 \ @jaredzimmerman http://loo.ms/g0
On Fri, Jan 9, 2015 at 3:15 PM, Kevin Wayne Williams < kwwilliams@kwwilliams.com> wrote:
Max Semenik schreef op 2015/01/09 om 16:01:
As always, if there is a way to do something, there will be a way to
abuse
it. Remember when we enabled IPv6 support some people started moaning
that
new style IPs are vandalising even though the rate of vandalism wasn't different between IPv4 and IPv6 anons? This is the same situation: to
your
example one can always provide a counterexample, "OMG the article about our favorite singer is so crappy, let's all help make it awesome!" Is that bad? Even someone as hating social networks as me has to agree that by now, there's no rational reason not to add social sharing buttons.
Not sure where to reply to a top-post to a bottom posted thread, so I
will
shoot for the middle and hope people can keep track of this knot. Your counterexample (which can be manually done today, so I've got experience with it) invariably winds up with a fan-flood of inexperienced editors
and
we wind up semi-protecting the article to keep them from damaging it.
KWW
On Fri, Jan 9, 2015 at 2:25 PM, Kevin Wayne Williams < kwwilliams@kwwilliams.com> wrote:
Brian Wolff schreef op 2015/01/09 om 15:17:
I think its important to separate two types of social media interaction:
*allowing people to post their favourite article (share this links) *meta level interaction (stuff about the community)
Nobody objects to the second afaik. The first is like proposing nsfw filters on commmons (ie get ready for the pitchforks).
You missed the worst part: "Some evil administrator won't let me post
that Mariah Carey/Iggy Azalea/pop singer of the week sold 50 bajillion
copies
of her latest album! Fans Unite, and make sure that Wikipedia has the TRUTH!" accompanied by an "edit this article" link to the singer's article. The last thing we need to do is make that kind of crap easier.
KWW
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l