[Foundation-l] Analysis of lists statistics: community in decline

[Robert Rohde]

Good work Milos.  I like the breadth and depth of your stats.

On the more general point though, I don't think this is mailing list
problem.  As Wikipedia matures, the community at the largest scales
has been contracting.  Hopefully we will stabilize at some reasonably
productive level, but in many ways the peak level of activity already
appears to be behind us.

-Robert Rohde


On Wed, Oct 29, 2008 at 10:27 PM, Milos Rancic <millosh at gmail.com> wrote:
> At the page [1] you may see complete data for a number of lists (of
> bigger projects). Below this email is my analysis of data.
>
> In brief, you may read inside of the analysis:
> * Almost all lists are in decrease.
> * Decrease varies between rare significant, but not high decreases and
> rare dead lists. In other words, the most of the lists are in
> significant decrease: Some of them since 2005, more of since
> 2007-2008.
>
> Positive and "positive" trends:
> * Technical lists show the smallest amount of decrease.
> * Two [of analyzed] lists -- textbook-l and wikija-l -- show
> *increase* of traffic! It would be good to analyze why it is so. Maybe
> they have the answer to our problem: increasing of list traffic
> usually means that community is increasing. (Or they are just in the
> earlier phase, which means that they will show decrease of traffic
> during the next year or two.)
>
> "Trivia":
> * Russian Wikipedians don't use WMF based lists for their
> communication. (Or they don't use mailing lists at all, which seems to
> me less possible.)
> * Portuguese list doesn't have extension "-l" in the name.
>
> So, some numbers are analyzed. Unlike simple claims like "foundation-l
> traffic decreased", we have now significant enough data: decrease is
> systematic, not only at one list and in amount of emails, but on
> almost all of [analyzed, bigger] lists and in amount of new and active
> participants. This shows very well that our community and our
> communities are not so alive like they had been in the past.
>
> And to be more clear. If we take a look at traffic at this list for
> Octobers 2006-2008, we may see that the approximation of decline is
> 20% for the first year and 50% for the second. If this trend
> continues, we will have ~175 emails during the next October, ~50
> during October 2010, ~10 during October 2011. According to the
> statistics of other lists -- 10 emails during October means no emails
> between June and September. This means that foundation-l will be in
> 2012 at the position where wikiquote-l is no (de facto dead list).
>
> And some good news:
> * We have enough time to change things.
> * If content projects would become history, MediaWiki would be alive
> for some more time.
>
> The question is: Do we have ideas how to make things better?
>
> [1] - http://meta.wikimedia.org/wiki/Problems/List/Low_activity_on_mailing_lists
>
> * * *
>
> == Analysis: Data interpretation ==
>
> === General lists ===
>
> * Communication on all general lists are decreasing. Technical lists
> (wikitech-l and mediawiki-l) are significantly better than general
> non-technical list foundation-l. Out of other particular cases,
> technical lists stays the best.
>
> === Project lists ===
>
> * commons-l has very high decrease in traffic, while decreases in
> number of new and active participants are significant, while not so
> high.
> * textbook-l (Wikibooks list) is one of the rare examples which is not
> in decreasing!
> * wikimediameta-l increases, while it is a very young list, which
> should take community part of discussion from foundation-l list.
> However, increase on this list is not significant enough to cover (at
> least, partially) decrease at foundation-l.
> * wikinews-l: Similarly to foundation-l -- significant decrease.
> * wikipedia-l: It shows very high decrease in all aspects. During the
> first years it was used as the main list, including for Wikipedia in
> English. Decrease 2002-2003 shows moving Wikipedia in English issues
> to the language specific list -- wikien-l. Decrease 2005-2006 probably
> shows moving general issues to foundation-l. However, 2006-2008
> doesn't have any obvious reason and it follows the similar decrease on
> the foundation-l list.
> * wikiquote-l: While it was not a very active list ever, this list is
> de facto dead from June 2008.
> * wikisource-l: While it was not a very active list ever, it shows
> ''not'' significant decrease during 2008.
> * wikispecies-l: This list was not significantly active ever and
> conclusions about time line of its activity can't be made.
> * wiktionary-l: The list was in decrease during 2006 and again between
> mid-2007 and present (October 2008).
>
> === Per language Wikipedias ===
>
> * wikide-l: In constant decrease since the first half of 2005.
> * wikien-l: In decrease since the beginning of 2008.
> * wikies-l: In increase from the second part of 2006 and during 2007,
> but in decrease during 2008; the second part 2008 has less traffic
> than the second part of 2006.
> * wikifr-l: In constant decrease since the end of 2005.
> * wikiit-l: In constant decrease since the beginning of 2007.
> * wikija-l: The only analyzed Wikipedia list which shows increase of
> traffic -- since the beginning of 2006.
> * wikipl-l: In constant decrease since the second half of 2005.
> * wikipt: While it was never a very active list, it is almost dead (3
> emails for October 2008, no emails between June and September).
> * wikiru-l: Data shows that the community around Wikipedia in Russian
> doesn't use wikiru-l as their mailing list.
> * wikizh-l: In constant decrease since the second half of 2006.
>
> _______________________________________________
> foundation-l mailing list
> foundation-l at lists.wikimedia.org
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
>