The Wikipedia Library/1Lib1Ref/Lessons/2018

The Wikipedia Library
#1Lib1Ref Year 3
2018 Report
by Jessamyn West with assistance from The Wikipedia Library

Executive summary edit

2018's #1Lib1Ref campaign built upon the scaffolding of previous campaigns, the ease-of-use of new and improved tools, and a coordinated communication effort. 824 editors made over 6500 #1Lib1Ref edits to Wikipedia in 22 languages, more edits than the first two years combined. Improved data collecting methods mean that we also have access to more and better statistics, including:

Social: nearly 4400 tweets from 40 countries, reaching 5.8 million people. Facebook group experienced 22% growth
Retention: More newly-registered editors returned from 2017 to edit this year than in years previous
Gender parity: Facebook Wikipedia + Libraries group is 45% women, 54% men (by facebook gender ID stats)

Website edit

The website at 1Lib1Ref.org is great for being a Wikipedia-Library branded, multilingual portal to the various aspects of the campaign and even better at being a linkage between the #1Lib1Ref campaign and other Wikimedia projects and information. Though a focus of earlier campaigns, it was not as central to the organizing of the campaign as it had been in years past. More day-to-day discussion happened on the Wikipedia + Libraries Facebook group.

Room for improvement: Monitoring the talk pages of the 1Lib1Ref website in future campaigns to make sure queries don't go unanswered.

Tools edit

Citation Hunt edit

Most people cited the Citation Hunt tool as being very useful in helping them locate articles needing citations. Increased effort was made to highlight CH's search features to enable people to find articles on their topics of interest and languages of interest. New addition this year of a Leaderboard added to friendly rivalry between participating editors. Citation Hunt's code is on Github.

Room for improvement: One item on the CH "to do" list is the addition of worklists whereby people can use CH as a list-generating tool for making list of thematically linked articles needing attention. This has now been added to Phabricator and should be done by next year's campaign.

Wikipedia social search edit

This was a powerful tool for gathering stats in almost realtime and improved dramatically during the campaign with the addition of a date limiting feature. A secondary benefit of this tool is that individual edits are linked making it a simple way for people to go thank people for edits or, if needed, review #1Lib1Ref edits being made.

This search tool also allows for easy quantification of the content produced by the campaign. For example, we developed maps showing countries associated with articles improved and also more specifically birthplaces of women whose biographies were edited.

Room for improvement: Having the date limiter be ready from day one will be a big help as well as having a way to run stats for a single campaign. Sometimes the search is slow, so there may be room for optimization of the query. It would be nice to have a way to search for multiple hashtags at once to include all language variants.

Slack edit

For the first time, the entire team had a Slack channel where they could communicate asynchronously, accessible via mobile and desktop. This was useful for discussing issues that came up, keeping team morale high, and "pinning" working documents so that they were easily findable, searchable and discussable. It was an invaluable aid for troubleshooting, workshopping, and bringing in other members of the Wikimedia staff for questions and concerns.

Room for improvement: We're still learning how to use the tool to its best advantage and it's not yet a part of everyone's workflow.

Stats edit

On English Wikipedia, an analysis of editor retention shows that it's improving. First off, we have more editors:

  • Number of editors in 2016: 272
  • Number of editors in 2017: 667
  • Number of editors in 2018: 824

Second, this is a little complex to parse, but we seem to be hanging on to them, though maybe not as well as we'd like to.

  • Number of editors from 2016 who returned in 2017: 60 (22%)
  • Number of editors from 2017 who returned in 2018: 129 (19%)
  • Number of editors from 2016 who returned in 2018: 51 (18%)

When we look at newly registered editors, ones who signed up within a week of the beginning of the campaign, they're sticking around but that effect decreases over time. Only 35 editors have stuck around for all three years.

  • Number of new registered editors (en) from 2016 who returned in 2017: 4 (4%)
  • Number of new registered editors (en) from 2016 who returned in 2018: 1 (1%)
  • Number of new registered editors (en) from 2017 who returned in 2018: 15 (7%)

Writing edit

More content means more things to talk about and use for publicity/sharing. The core team had a more coordinated communication effort this year, particularly in trying to lay the groundwork in advance of the actual campaign.

Two emails were sent to community organizers, and another to various listservs introducing the campaign and pointing people to various helper documents and sites for getting oriented. Jessamyn wrote an introductory email which she sent to representatives from all 50 state library associations. This spawned a lot of helpful follow-up conversations to help with strategy specific to libraries. The team also kept #1Lib1Ref in the front of wiki-conversations, generally helping with related projects like this Wikimedia Commons tutorial.

Three blog posts during the campaign ([1][2][3]) targeted different groups and showcased different aspects of the campaign

We also created the Coffee Kit, translated into 13 languages, a step by step guide for people new to Wikipedia to create a 45 to 90 minute workshop type event to work on #1Lib1Ref which included a sample flyer.

Wrap-up reports and selected writing from other organizations (full list):

Room for improvement: Blog posts on the WMF blog don't meet general standards for accessibility and should make use of consistent application of ALT text (or captions if for some reason ALT text is impossible).

Design edit

Graphics edit

By the time the campaign was underway we had a graphic identity which we could use for social headers etc.

Handouts edit

Coffee Kit handouts were made available via the #1Lib1Ref Coffee Kit page. A basic #1Lib1Ref handout was distributed to library associations in all 50 US states. Wikipedia volunteers translated them into French and Spanish. These were then shared out via Dropbox. It would be good to have a better and more Wikipedia-branded way of being able to share documents that users could maintain themselves.

Video edit

Felix took the lead on creating a video using SimpleShow which outlined the importance of the #1Lib1Ref campaign in a friendly, accessible manner.

This was a popular early morale booster which was an easy introduction to the topic for beginners.

Room for improvement: Video format not accessible in all modern browsers, should transcode into alternative formats for maximum accessibility.

Social & community engagement edit

Twitter and Facebook were active channels for user engagement. All members of the #1Lib1Ref team were accessible via multiple social media platforms at various times of day which allowed for our social media presence to be a great way to find a team member to ask a question or amplify signal on a local event.

Twitter edit

Two Tweetstorms ran at various stages of the event. We also created two Twitter moments collating other #1lib1ref activity.

Facebook edit

We celebrated our 10,000th #1Lib1Ref edit, from Serbia, which was an opportunity for Facebook platform engagement.

One of the aspects we did not use as much as last year was the Facebook Events option (they do not allow events going over two weeks). Most Facebook activity was in the Wikipedia + Libraries group which just passed 1500 members. Membership to this group increased 23% during the #1Lib1Ref campaign this year and 26% since the start of 2018. Nearly half these members were active in the group during the campaign, and all of the engagement metrics for the group were increased.

Instagram edit

While we did not make a concerted effort at outreach or engagement via Instagram, librarians were still active on the platform and it's an area ripe for buildout next year.

Coordinators edit

We were able to draw feedback from coordinators covering six countries, out of which five were emerging communities. The predominant concerns were the lack of translated marketing materials (i.e. prefilled announcement for translation) that could have been distributed to librarians and Wikipedians before the campaign. Coordinators seem to believe that this would've considerably multiplied their opportunities to reach out to their local communities and friends.

A majority of respondents from the African continent also shared their limitations in accessing local sources even when they were excited to take part in the vision of the campaign, to improve articles in their languages. One other comment from a respondent that seems to strike our thoughts about 1Lib1Ref, "I guess it could be a good idea, for the future, to transform 1lib1ref into something more generic. The reference problem is real, but the solution is more broad and inviting, just librarians will not fix the problem in general, as well as spreading an idea that only librarians can fix it".

Future plans edit

While there was some call to keep #1Lib1Ref going year-round, there is also an appeal to a targeted date-limited campaign for focusing people's efforts and encouraging them to stretch to meet goals. While #1Lib1Ref is timed to coincide with Wikipedia's birthday, antipodal Wikipedians would prefer a campaign that is during their wintertime and not their summer break. A big addition will be 1Lib1Ref South happening in May.

Opportunities also abound for promoting other ways of Wikipedia engagement and interactivity including

Links edit