We are pleased to present our progress against the Wikimedia Foundation FY16-17 annual plan. Our programs were directed by the three areas of focus determined during the early 2016 strategy process. Those areas are readership and reach, volunteer retention and engagement, and support for knowledge creation (Reach, Communities, and Knowledge). We have also worked to build and rebuild core capacity, improve overall organizational effectiveness, increase communications and transparency, and prepare for a long-term movement strategy.
Some selected highlights from the year so far include:
The Advancement department improved the fundraising performance of all channels in the 2016 English campaign countries (United States, United Kingdom, Canada, Australia, New Zealand, and Ireland), and hit the yearly goal in record time. In response to this, we ended this banner campaign a week early.
The yearly audit of the Wikimedia Foundation was clean, finding no issues with the accounting of our financial records, and no deficiencies or weaknesses of significant importance in Foundation processes supporting our financial transactions. The audit report and frequently asked questions for the fiscal year ending 30 June 2016 are available on Foundation wiki.
Talent & Culture launched a new Human Resource Information System, streamlining information about benefits, payroll, and other essential information into a single coherent system.
People all over the world have yet to discover Wikipedia, we went to them to find out why. The New Readers team went to India, Nigeria, and Mexico to reach young mobile users and found that Wikipedia name/brand recognition and awareness was less than 50% of surveyed participants in all three countries. In India and Nigeria, greater than 75% of participants said they had never heard of Wikipedia. These findings are being integrated into Foundation work plans to better inform product design, awareness efforts, partnership development, and volunteer leadership capacity building.
The Desktop and Mobile Web team helped readers save money by decreasing the amount of data needed to open a page. Photos are a ubiquitous element of Wikipedia’s most popular and highest quality articles: this change means that pages will only load images as a user scrolls through the article, saving data and money.
The Community Engagement department has been actively working to implement solutions to key challenges exposed in the Anti-Harassment research. The primary focus for Support and Safety is creating training modules to support community functionaries. This platform is meant to be expandable, but our initial focus is on anti-harassment training for online functionaries and event coordinators.
In 2016, the Research team started a formal collaboration with Jigsaw, aiming to develop algorithmic models to identify and measure personal attacks and aggressive tone on talk pages. A preliminary framework for this approach has been designed and evaluated and is being further developed.
The 2016 Community Wishlist Survey has completed the voting stage. There were 265 proposals, 1132 contributors, 5037 support votes this year. The proposals will be reviewed by all Product teams for inclusion into their goals.
The Legal team successfully defended a community member facing criminal defamation charges in France, and set strong precedent safeguarding the Wikipedia trademark in Spain.
Community Engagement partnered with community and the Internet Archive Library to repair one million broken links. This pilot was an effort to preserve copies of external pages used as sources on English Wikipedia and, when links are broken, provide links to archived versions. Next steps will include expansion to other languages.
Strategic direction for the Wikimedia movementEdit
Since then, we have created a core support team that includes Guillaume Paumier, Suzie Nussel, and Ed Bland and Shannon Keith from strategy consulting firm williamsworks. Working with a steering committee of community members, they are developing an extensive and inclusive process to determine a strategy direction for the Wikimedia movement.
For the period July 2016 to December 2016, the Foundation experienced a $11M or 22% increase in revenue and a $2.7M or 10% growth in total operating spending year over year. We incurred the most significant growth in spending in support of our grant making programs; this growth in spending represents approximately 85% of our overall growth in spending. The year over year growth in grant spending resulted from changes in both the structure and timing of our grant making programs. When normalizing for this material increase in grant spending, all other operating spending was up $0.4M or 1.6%.
WMF Total by Program
Data Center Expense
Outside Contract Services
Travel & Conferences
Sub Total Program Expenses
Non Program Expenses
Donation Processing Fees
Outside Contract Services
Travel & Conferences
Other expenses (Facilities, Insurance, FF&E, All Hands, Recruiting, Staff Development, Personal Property Taxes, and Etc.)
The Advancement department has exceeded their revenue projections considerably despite underspending in contract fundraising services for the first half of the year. Several major travel expenses related to fundraising events were delayed to the second half of the year. Success of the fall fundraising campaigns led to a proportional increase in donation processing fees that is expected to continue to be over budget for the year.
The Community Engagement department is on track to meet the majority of their annual objectives but with some delays. The Community Resources team has allocated all of their budgeted grants so far this year. However, $1.7 million in FDC Annual Plan Grants that was budgeted to be expended in Dec was delayed until Jan. Therefore it is not included in the below reported actual expenditures. The Community Engagement Department is underspending in a few other budget areas caused primarily by a decision to postpone the hiring of 2 positions, an executive and management position, as well as less travel than originally planned.
The Product department is on schedule with nearly all of it's Annual Plan objectives. The Product department has small underspending in a few categories. Personnel costs were reduced by delays in hiring two staff positions. Contract Services were underspent in the first half of the year with some costs postponed to the second half of the year. An significant unbudgeted investment in Wikidata has increased other costs and has offsets some underspending in other areas.
The Technology department is on schedule to meet most of its objectives, but it has experienced some delays with a few. See introduction above and descriptions below for more information. The Technology department has underspent in all major categories. The security program was paused while a Chief Technology Officer was recruited. This led to the substantial underspending of both personnel related expenses and contract services associated with security. Underspending in Data Center Expenses are temporary. The WMF's servers have been managed so that server refreshes could be shifted to the second half of the year which has will lead to much of the data center expenses happening in Q3 and Q4.
The Governance department has an on-going underspend in non-salary personnel expenses and had a reduced need for contract services expenses in the first half of the year. This is being somewhat offset by an increased amount of travel related expenses.
The Communications department is on track to meet nearly all of it's objectives on time, despite underspending for the first half of the year in personnel and contract services. Hiring for the executive level position on the Communications team was postponed while the foundation focused on hiring a Chief Technology Officer. An additional management position will be hired later in the year than expected. The Communications team has also negotiated some cost savings in their largest contract service expenses.
The Legal department is on schedule with their programmatic activities in the FY 16-17 annual plan despite underspending in key spending areas, including staffing and legal costs. Hiring of the General Counsel position was delayed while the Foundation focused on hiring a Chief Technology Officer. Our legal fees were lower than expected for two primary reasons - new litigation filed against our contributors in the first half of the year produced lower costs than plan and an in-house staff member was hired to resolve routine day to day staffing activities.
The Finance & Administration is on schedule with a majority of their annual plan objectives despite underspending in outside contract services and general & administrative expenses. The majority of the underspend was due to fewer resources required to achieve certain objectives, and the department has repurposed over 50% of its underspend towards the Movement Strategy initiative.
Financials by department: Finance & Administration
The Talent and Culture department is on schedule to meet their Annual Plan objectives and to meet spending expectations with the exception of staffing. Staffing was underspent as a result of the delay of hiring the Chief of Talent and Culture Officer while the priority was placed on hiring the Chief Technology Officer.
On track with over 300 leaders engaged across nearly 600 engagements through outreach and conference events in our first two quarters and preparing for high engagement at Wikimedia Conference in Berlin.
We are on track to exceed our target of 60 community leaders engaged in Learning Day events with our pre-conference convening at Wikimedia Conference.
We are on track to reach the target of engaging at least 20 community leaders as trainers and mentors in their region or thematic area at Wikimedia Conference in Berlin as well as to reach our target to deliver program capacity and learning & evaluation session content at 4 or more regional/thematic events to support at least 150 program leaders in developing their capacity for programs and learning. So far we have reached 107 at the mid-year point through events at WikiConference India, CEE Meeting, and WikiConference North America.
Boost peer mentoring engagement
We are on track with this objective with our responsive workshop support at regional events in Q1 and Q2, we will set this target also for Wikimedia Conference in Q3.
Scale learning and evaluation teaching and best practices through community leadership
To date we have engaged about 20 leaders in learning about community listening at the CEE meeting in August. We are looking forward to the Wikimedia Conference to be able to reach more people.
For the survey and processes for annual surveys, see Program 10 Goal 2
Program 2: Introduce training modules for functionaries and event organizers regarding issues managementEdit
On schedule to have the full draft finalized before the end of Q3 in order to use the Berlin Wikimedia Conference to present the module and get substantial qualitative feedback to leverage for Objective 3.
Expand the module's reach
On schedule to begin work after the module is complete.
Program 3: Develop capacity-building tools and resources for core Wikimedia programsEdit
Create minimum viable product for the programs and events dashboard
We successfully released a version of the Program and Events Dashboard that is beyond the minimally viable version. The current version works with any Wikimedia language project and contains a campaign feature that tracks groups of programs and events. The Dashboard is being used on over 10 languages, including at least 3 emerging communities. It currently includes over 280 programs and 4,000 participants in those programs.
We are grateful for the ongoing partnership with the WikiEdu Foundation on this project. In addition, the Dashboard will be used for other fruitful projects, including SuSa's anti-harassment and event safety trainings (part of Program 2 above.)
Develop a Wikimedia resource center
In the first phase of development for this platform, the exercise of standardizing criteria to navigate resources and represent topics across teams and departments at the foundation brought along a broader topic: how do teams represent themselves online on community-facing platforms? This is why, in the second phase of the WRC, we will work to develop team templates that will not only give standards for teams to create their page, they will also facilitate embedding of content on the WRC.
Maintain infrastructure and content supports
Wikipedia Library Card platform released live phase 1 signup website. Phase 2 proxy access should be complete by end of fiscal. Phase 3 search should be complete by end of 2017 calendar year.
Program 4: Allocate resources to effective community-led projectsEdit
Meet community needs during the restructuring transition
196 grantees were supported in Q1 & Q2 (169 & 27 grantees, respectively) through 2 Project Grant rounds, 1 FDC APG rounds, our other grant programs (Simple APG + Conference + Rapid Grants.). 48% of those grants were awarded to emerging communities, amounting to 16% of the total amount of money awarded in Q1 and Q2. From the grant reports that were submitted in Q1 and Q2, we also know that about 62,500 people were involved in grant related work.
Improve WMF’s support system for grantees
We successfully transitioned all grantees, no remaining open PEG requests! All annual plan grantees on track to transition to APG by 2018, after the end of the pilot period.
Evaluate and adapt
As planned, our report on the Gender Gap Inspire Campaign was published on Meta in December 2016. This is first report on CR's thematic-focused Inspire campaigns, the first evaluation report ot focus on how the grants we give are making an impact in thematic areas, such as gender diversity and the gender gap. This was report was also a successful pilot of a new approach to grants evaluation, one which integrates qualitative and quantitative information to showcase impact in a way that is more digestible and relatable to both Wikimedians and non-Wikimedians. This report has received positive feedback from grantees, community members, and non-Wikimedians. The Evaluation of all grant programs is scoped to begin in Q4, but will likely not be complete by the end of the fiscal year as it is a multi-month project with a long data aggregation and synthesis phase. Priority for quarters 3 & 4 is the scoping of a grants metric tool, in collaboration with Community Tech.
Program 5: Give more support to GLAM partnershipsEdit
At time of annual plan, we imagined formal technical consultation; however, early evaluation found a critical ready-to-fix technical issue: Commons workflows. Structured data on Commons addresses many of these concerns and opens a venue/route for addressing other concerns; we accordingly dedicated time to contributing to developing a proposal to support structured data on Commons. Additionally, we identified venues for collecting other technical needs (Community Tech Wishlist and European Coordinators meeting). Formal consultation may be de-prioritized this fiscal in favor of the more informal avenues.
Develop resources on best practices
Beginning the process, and have identified plan of action. Resources may not be finished by end of fiscal, but long term plan in place for documentation prioritization.
Develop and maintain responsive staff support systems
Have developed awareness and connections across the organization and GLAM-Wiki community to support conversations. Have a regular workflow for assessing outside requests, and needs from various departments.
Program 6: Support Affiliations Committee and affiliate partnershipsEdit
In Q2 we built up a backlog near the end with affiliate report review queue after big surge of submissions; we also struggled to maintain response times to user group inquiries due to additional focus needed to get the Community Engagement Insights surveys launched.
Maintain mentoring and consultation support
In Q1 we completed SMART target consultation and in Q2 proposal review for Round 1 FDC annual grant applications and continued to address technical needs for developing grants metrics library and tools including outreach and needs mapping support for program tools in the Q2 Community Wishlist.
Increase AffCom engagement and support to affiliate partnerships
AffCom have begun to accept chapter applications again and were able to lend financial support to enhance the upcoming Wikimedia Conference pre-conference Boards Training in order to support smaller chapters and affiliate groups who do not have budget to send their board members the two days early.
It is expected that the group continue to take on more leadership at engaging affiliates in upcoming strategy conversations as we move beyond the most high level movement strategy to the more specific on affiliates strategy.
Program 7: Improve collaboration with communities in product developmentEdit
To date, all requests to support Product goals have been resourced adequately.
Participate in audience planning
Different planning processes between Product teams make our participation under common standards challenging. We consider our participation in planning processes good in the case of Community Tech and Editing; overall good but sometimes challenging with Discovery; and in need of improvement with Reading. The amount, average size and average length of projects in Reading and Discovery requires us to approach our participation in a different way, which we have started to discuss.
Facilitate all major product conversations
We have facilitated or are facilitating all the major discussions related to Product. The list of related projects includes the Community Wishlist, the Flow satisfaction survey, contributions from Android app users, the 2017 wikitext editor, the Save/Publish button, Page Previews, interwiki search, Wikidata descriptions in mobile web, Language button in mobile web, and Design's statement of purpose.
Initiate a guideline for technical collaboration between WMF teams and the communities
A first draft for review was completed on time, but the review is taking a bit longer than initially planned. However, we believe that taking this extra time to listen to more voices better will ultimately help us delivering a better first version of the Technical Collaboration Guidance on time.
Facilitate wide adoption of the community engagement guideline
Too soon to tell. A lesson learned is that the TCG can only be fully applied to new projects, since the beginning, in order to become an effective tool. We are planning to start a new pilot in the beginning of 2017.
Integrate community engagement more closely with early Foundation product development stages
Too soon to tell. The first responses to the TCG both from community members and Foundation teams show a common understanding of the need to have the important discussions early in the development processes. Still, we need to see how this works in practice, and FY2016-17 might be too soon to have more than anecdotal evidence.
Program 8: Connect volunteer developer work with community technical needsEdit
Encourage volunteer focus on addressing community priorities
We are confident about reaching the goal of 10 Community Wishlist projects completed by volunteers, although a lesson we learned is that from dozens of requests, only a minority are projects really fit for a regular volunteer project.
Encourage hackathon and outreach work into community requests
Already now the Community Wishlist has become a main reference for our main developer events and developer outreach programs. In fact, the success of this model moved us to support a community request to organize a Developer Wishlist Survey with similar goals.
Better connect developers with tasks and mentors
Too soon to tell. We have made progress discussing ways to improve the onboarding of new volunteer developers, connecting them more efficiently to suitable tasks and mentors, but actual changes still have to be implemented. Instead of organizing an own technical community health survey, we agreed to join efforts with CE Insights, and it is too soon to tell whether this model will provide the data we need to mesure progress.
Program 9: Identify support gaps and provide targeted resourcesEdit
One Inspire campaign launched Jan. 30th on outreach to outside knowledge networks. Will track [# of ideas submitted], [# of participants], [# of ideas actively supported for implementation through grants or community consensus]. Changes made & expectations for Campaign #2.
Improve community capacities through pilot programs
Community Capacity Development Program evaluation in process, but 10 communities not likely this fiscal.
Based on last Inspire campaign on addressing harassment, many ideas do not require grant funding and need staff support to be implemented through local community consensus. We have supported many community members as they turn their ideas into action, e.g. semi-protecting User: namespace to reduce harassment. CCD has seen early success annecdotally, particularly on the Brazil/PT & Tamil pilots, but the data is still forthcoming; however, CCD program officer is likely to be deployed to the movement strategy work, which was not built into the annual plan.
Program 10: Maintain quick public response time and expand community surveysEdit
Answers@ and business@ KPIs maintained through Q1 & Q2
Conduct follow-up surveys
Support & Safety team surveys of staff and supported community groups launched in Q1 and Q2 and areas of improvement identified; plan created to better serve, where possible, and also to help groups with internal challenges.
The survey translation was completed in December and the survey launched in early January in 12 languages to 4 audiences, which include a sample of 16,000 editors who are either active (1-99 edits/month) and very active (100+ edits/month), 400 community leaders (e.g. affiliates, organizers), and an estimated 200 technical contributors. The survey included about 300 questions to learn about community experiences, gather feedback from communities, and measure the outcomes of Wikimedia Foundation programs. As of 31 Jan we have about 2,600 responses.
Program 11: Protect our users, staff, and publicEdit
Over Q1 and Q2, SuSa created temporary systems to track contacts unrelated to emergency systems and track the speed with which we respond as well as the speed with which non-emergency cases are actioned and closed. We are still working to establish sustainable systems to track all of our investigations through the better use of case managrement software and efficient inbound case processing.
Improve how readers and editors interact with content by applying research during product development
Working with Communications we are developing a comprehensive Audience Segmentation and conducting additional research on Contributing (editing) audiences. Primary designs and audiences crafted.
Summarize, process, and apply output of New Readers research project
Contribute to and collaborate on design research in additional regions / countries is complete and a public report was generated on the analysis. Synthesis of collected data is ongoing into remaining goals for the year.
Program 2: Maintain and improve content discoveryEdit
Completed community feedback and planning phase, currently working on prototype to convert into A/B tests on 4 wikipedias. This will enhance discovery into sister projects for the casual user when they arrive at the search results page.
Increase relevancy of search results
Continue to enhance language analyzers and detection of language of search terms. Language analyzers improve support for their respective language, thus significantly increasing the usefulness of search in multiple languages.
Maplink has been launched on all wiki projects, mapframe has been launched on only a select Wikipedias. At risk: launching mapframe on all remaining Wikipedias.
Connect Wikimedia projects
Graphs, tabular and map data namespaces have been launched on Commons. At risk: continued work to improve and fix bugs.
We are working towards putting this work on pause. Expectations around deployments, goal setting, and communications were set with the Interactive Team which were not met, which ultimately led to this pause. We are currently evaluating our options for deployment; we cannot deploy a new service to production unless we have a maintenance plan.
Program 3: Increase our global reach by increasing readershipEdit
A lot of work goes into maintaining and supporting our existing codebase. There are more than a dozen extensions that the web team is responsible for supporting, in addition to their work in evolving the experience. There is a fair amount of "interrupt" work caused by things breaking in a distant part of the codebase.
The apps have a smaller, more unified codebase, but given the high velocity of development a good portion of team focus goes into ensuring the continued stability of the platforms.
The objectives here are all complete or continuing in terms of output. We released 2 major features per platform and notifications on iOS. In terms of outcomes, we are on track on the apps and do not have the infrastructure to measure yet on the web.
Improve article rendering
This objective has seen tremendous progress but is also ongoing. Growing appreciation for the improved mobile site is evident. We do not have the infrastructure to measure "retention" on mobile web*, but we are seeing promising signs in other measures.
We are building this and other new KPI's during the remainder of the year.
The objectives here are mostly complete. We have implemented 2 of the 3 wikidata integrations on all platforms. The only achievement we are not expecting to hit here is using wikidata in the infoboxes--this is currently being worked on by others in the movement.
On track. We currently have 3 prototypes that have undergone their first round of evaluative research. Work on building at scale should commence in Q4. While we anticipate this to impact reach in target markets, we do not expect a global bump from this first feature.
We created a new wikitext editing mode with the visual editor, integrated into its workflow, and with all its tools, and released it as a beta feature with lots of support from community members to find issues so we can make it better during the rest of the fiscal year. We have incrementally improved the visual editor in dozens of small ways, improving stability and performance.
Modify and improve curation tools
We have designed, user-tested, & iterated a beta feature of improving Recent Changes, adding ORES-based and other filters. This beta feature will give reviewers insight into contributors’ intentions and enable identifying damaging but good-faith edits by new users. This will be released in the second half of the fiscal year.
We have explored some early concepts of how such a system could be implemented. We have not prioritised this objective.
Work on the back-end improvements to wikitext and modernisation of the pipeline is going well. The user-facing tools to help editors find and improve broken wikitext will be released soon.
Expand the core of MediaWiki to support new kinds of content
This work on Multi Content Revisions will now be picked up as part of the grant-funded work for Structured Commons headed up by the Reading department.
Streamline metadata editing
Work to migrate the Commons community gadget ImageAnnotations to be a full extension providing a better meta-data editing and using experience is mostly complete. It will be released for community comment soon, before it is deployed.
We've increased the functionality and stability of the content translation tool and are now working toward integration of the visual editor which represents the last significant change required before we being to move it out of beta. It is slightly behind the hoped-for schedule when the annual plan was written.
Program 5: Maintain and build contributor toolsEdit
Wish #5: Deployed numerical sorting in categories on English WP and other languages. Wish #9: Launched CopyPatrol, an interactive interface for the plagiarism detection bot. Wish #4: Defined a technical plan for the cross-wiki watchlist. Also shipped improvements to AbuseFilter, Programs Dashboard and a new PageAssessments API.
Launched Striker, a tool that connects developers' LDAP account with their Wikimedia user account, and allows them to easily create git repositories. Extended Striker to help new developers create an LDAP account, removing barriers in the new account creation process. After community discussion, agreed upon a new policy for developers taking over abandoned tools.
Conducted the 2016 Community Wishlist Survey in November and December, with 1,100 contributors proposing and voting on 265 proposals. (See results here.) In calendar year 2017, the team will investigate and address the top 10 wishes from the survey, as well as some other projects that support admins and stewards, program and campaign organizers, grantees and Wikisource users.
Program 6: Wikidata – Extend the reach of structured knowledgeEdit
Note: Program 6 was added after Annual Plan approved in June 2016
We have made significant progress and improved the processes. We have also started work on redesigning input widgets and switching more of our UI to OOUI. The topic of usability and design improvevments is a large one and work on it is continuing.
Improve Wikimedia Commons build out
The first test system has been published and well-received by the Commons and Wikidata community. We have taken the feedback we have received so far into account for development and will continue that. Usage tracking is ongoing and will be greatly helped by the Sloan grant and work by WMF
Add article placeholder feature
We have rolled out the ArticlePlaceholder on 9 projects now. Rollout on larger Wikipedias is requested (Russian) but on hold for performance reasons for now.
Connect Unit conversion & new data types
Thanks to work by the Discovery Team (that we supported) unit conversion for the first dimensions in the query service is available now. We are doing a staged roll-out of more dimensions based on community wishes and feedback.
The new Chief Technology Officer was hired in November 2016. The Technology Department experienced localized resource challenges during the reporting period which delayed progress in some areas, most notably security. A new plan has been put in place to bring in a senior security leader and resources are allocated to remedy the shortages.
Program 1: Improve tools that help us understand user needsEdit
Increased throughput capacity of pageview API by an order of magnitude while reducing latencies. We made available pageviews for all wikis that have requested them through said API. Currently working on reconstructing mediawiki edit history to easily be able to calculate and release edit metrics with WMF and community.
Real-time data processing
EventStreams has been launched, deprecating RCStream. Work on real-time data processing is postponed due to low prioritization, and the decision to wait until Kubernetes is useable. (Kubernetes will make it easier to experiment with this, as we don't have to provision new cluster hardware.)
Build compatibility between key databases
This rather a concrete milestone witin Objective 1, not an objective on its own.
Program 2: Expand research capabilities to equip product development to better meet users’ needsEdit
We completed in Q2 a major collaboration with Jigsaw aiming to characterize personal attacks and disrespect. We collected labeled data to design and train algorithms to help us understand the prevalence of attacks in Wikipedia talk pages and their impact on editor retention. We successfully submitted a paper to an academic conference (WWW '17); hosted public presentations of the results; publicly released data and code. We completed in Q2 a collaboration with Carnegie Mellon University to characterize supportive communication at the Teahouse, a new editor help forum. We analyzed the prevalence of supportive communication directed at new editors and the impact of non-supportive communication. Results will be submitted to a conference for publication in the second part of the year.
Evaluate AI as a service to help volunteers improve quality
Since being productized at the beginning of Q1, ORES has seen a significant increase in adoption: as of the end of Q2, over 18,000 editors have manually enabled the ORES beta tool on Wikipedia. External organizations (such as the Wiki Edu Foundation) have adopted the platform to power their dashboards. We grew language support to 24 projects/languages and expanded the set of models available across languages. As of Q2 ORES is being integrated into edit review tools designed by the Editing team. In Q2 we started research on the design and evaluation of stub expansion recommenders (in collaboration with editathon organizers) and reading recommender systems (in collaboration with the Reading team).
Create an open research infrastructure
In Q1 and Q2 the Labs and Research teams grew and promoted the adoption of PAWS, a notebooks infrastructure for open and replicable research. As of the end of Q2, the platform has been used by 800+ users to create over 3,000 notebooks. Computer science classes at 3 graduate schools (Colorado-Boulder, Carnegie Mellon, University of Maryland) are using PAWS for their students. A private instance of PAWS has been deployed internally to support the work of data analysts in Product teams. We met our availability goal as of the end of Q1. The addition of dedicated hardware for the platform, as well as a more prominent announcement, have been put on indefinite hold in favor of gradual outreach.
By the end of Q2, three papers have been successfully submitted and accepted for publication at a conference (WWW '17, CSCW '17). A proposal for our annual workshop has been accepted at WWW '17 and is currently being organized: http://wikiworkshop.org/2017/
Initiate formal research collaborations
As of the end of Q2, 5 new formal collaborations (subject to WMF's open access policy) have been established with research teams at: Carnegie Mellon University, TU Dresden, GESIS, EPFL, the Institute for Scientific Interchange.
6 total projects delivered to product teams: 3 for Reading, 1 for Editing, 2 for Discovery. Design research team lost one FTE resource, and responsibility for this work was moved into product teams.
Analyze and apply generative research for new readers
Facilitated cross team collaboration in order to focus the team in on 3 of 24 findings to address (Awareness, Offline, and Affordability). Designed and facilitated concept generation and evaluation for offline findings. Working with the Reading Mobile Web team, who produced prototypes of offline concepts, to test with end users. Provided some consulting support for Communications in designing surveys to inform awareness campaign designs for India, Nigeria and Mexico. Currently we are begining the process of concept generation for affordability findings.
Expand capacity for standardized iterative user testing
We engaged a contract with User Zoom. Mid year, there was a change in requirements for this as responsibility for the design research done within teams (and within their timelines), was moved to the product teams.
Create a Research minisite that would act as a landing page for all resources related to research at the organization
In Q1 and Q2 we ran a small-scale survey to collect user stories on needs of our research audience. We conducted a detailed audit of our research reports on wiki and an evaluation of the structure and information architecture of a landing page for Research and discussed it during our team retreat in December. We postponed the creation of the first version of this landing page to the second part of the year.
Strengthen research communications across teams and to the community
We hosted 11 presentations on research results at our monthly showcase and hosted a weekly cross-departmental research group meeting to share findings across the org. We delivered pragmatic personas to designers on the Reading team to evolve into resaerched personas. We delivered several reports and presentation on the New Readers research, both on wiki, off wiki.
Significant progress was made on reduction of operational technical debt, including the conclusion of our migration from Varnish 3 to Varnish 4, modernization of our Puppet configuration management infrastructure, a new Labs DB replicas platform and modernization of our metrics and incident monitoring infrastructure. Externally monitored availability across readership services for FQ1/FQ2 was 99.945%. Excluding the externally hosted wikimedia blog, availability was 99.966%. This is slightly below the availability target of 99.97% for the year.
Improve software security
The Security team continues to triage reported security issues and requested security reviews on a weekly basis, and works toward completion of these tasks on a rolling basis. Efforts to ensure uptime of automated static and dynamic scanning services are ongoing.
Improve the security capabilities of the Foundation engineering teams
Due to inadequate staff, the Security team has not expended efforts (beyond normal developer training) on the goal to increase intra-team capacity to address security concerns.
Improve privacy protection
The Security team has worked with the Legal team to implement portions of the Privacy Roadmap. The data mapping stage of the project is nearly complete. The teams anticipate reviewing the data and making recommendations for storage and retention changes during Q3 and Q4.
A stable Kubernetes cluster has been deployed within Tool Labs, and support for using Kubernetes as backend for Tools web services was made available, compatible with GridEngine. However, development of support for other forms of Tools has slowed down, as the Labs team needed to focus on other technical debt projects in the Labs IaaS product, such as the NFS storage infrastructure and the Labs DB replicas. In January 2017, 198 Tools were running on the Kubernetes backend.
Provide architectural leadership
Responsibility transferred over to Architecture Committee
Improve service platform
We deployed and documented rate limiting, coordinated timeouts and retry behaviors across important services. With operations, we secured all services with firejail. With operations and release engineering, we started research towards an integrated development, testing, production & third party deployment solution.
Build out Rest API
The REST API is used for several production features, and is currently serving close to 50% of overall API traffic. Several new entry points are in development. Version 1.0 is scheduled for official release in Q3.
Canceled. With the unexpected departure of the Performance team's manager, we reassessed the objectives. In addition to the reduced team capacity, this objective seemed like Mediawiki technical debt without a clear performance benefit.
Improve site availability
In Q3 and Q4, the process of data center switchovers is being streamlined with further automation and reduction of individual, manual steps which is expected to reduce the required read-only time from 30 minutes during a primary data center switch.
Deploy Asian caching POP
Scheduled for end of FY16-17, this work has just started with preparation work and vendor/site selection in Q3, contract execution, hardware procurement and start of buildout happening in Q4.
We have launched 4 new Wikipedia Zero partnerships, one of them in one of our priority countries (Nigeria), adding 138,000 new potential readers. However, most of these launches have been inbound requests, and not actively pursued by the team. Additionally, out of the 6 priority countries, 3 have public policy regulations that have either fully banned or challenged zero-rating offers, and/or where local community members have voiced concern over deploying the program (India, Brazil, and Mexico). The number of current Wikipedia Zero users was also heavily impacted during the first part of the FY, as major partners decided not to renew the 3 year agreement, resulting in a decrease of potential Zero users.
We need a technical solution to survey our users that is completely on wiki as they shouldn’t be asked to pay data charges to give feedback to a zero-rating program. This requirement created critical internal dependencies, particularly with the product team, who currently owns the on-wiki survey features. Currently, we are beginning to explore options with the fundraising tech team to build a targetable survey into the central notice system. The technical capability to do so seems straightforward, but we anticipate that additional engineering resources will be needed.
Establish at least one outreach or community support activity with a Wikipedia Zero partner
We have succesfully established a proper marketing campaign to support a new Zero launch in Iraq, with full support of the local community and the operator. The launch event of the Wikipedia Zero partnership in Iraq will take place at the Mobile World Congress, the biggest mobile industry conference in the world happening in Barcelona later in February 2017. This is the first time we host a launch event at the conference. Additionally to that, this is the first time we will create a baseline to measure the impact of outreach/marketing efforts in a country where we are launching Zero. We will survey Iraqi citizens before and after the partnership launch. The pre-launch survey is currently in progress.
Program 3: Increase reach through new types of partnershipsEdit
We successfully started 2 pilots in Latin America with 2 private sector companies (Free WP on Wifi in Mexico and free WP via USSD in Argentina). In addition, 3 partnerships with private companies (1 in each region: Brazil, Nigeria, and India) are close to being finalized and launched by the end of this FY.
Completed. Conducted media audit covering 2014-2016. Built understanding of Wikimedia press coverage from past two years including top narratives, geographies, outlets, and keywords. The audit was used to inform the Communications messaging strategy.
Continue media monitoring
Regularly produced daily (M-F) media reports of the Wikimedia movement and campaign/announcement media coverage reports to understand the impact of our work and inform the messaging strategy.
Completed. Developed a high-level messaging strategy with 5 key pillars: Reintroducing Wikimedia, Building an Inclusive Community, The Future of Open, Wikipedia for the World, and Where We're Going Together. Strategy tied to key organizational priorities and milestones, and informed by learnings from media audit.
Build strategy and tactics
Aligned messaging with pillars from messaging strategy. Used strategy to prioritize proactive media and storytelling pushes, including in the areas of anti-harassment work (Building an Inclusive Community), Wikipedia in a post-fact world (Reintroducing Wikimedia), and collaborations with open movement organizations (The Future of Open). Adopted Metrics Meetings as new workflow to align with messaging strategy.
Placed 2 op-eds advocating for public policy issues that impact the free and open internet (The Future of Open pillar), supported 10 ED interviews and 6 public speaking opportunities with messaging and training related to Inclusivity, Reintroducing Wikimedia, and The Future of Open.
Increase engagement with top narratives
Produced materials such as the annual report, fundraising campaign blog post, Facts Matter video, etc. that align with messaging strategy pillars. Share materials on community channels for community feedback and discussion to guide future work in each pillar. Working on more proactive rollout of the messaging strategy to community audiences and feedback in Q3 and Q4.
Program 2: Make our brand more consistent, relatable, and more easily understoodEdit
Completed. Developed “Facts matter” concept for annual report
Gathered donor and financial information Wrote copy for the annual report that supports organizational messages: how Wikimedia relates to 10 important facts from 2016.
Expand design capacities
Completed 33 unique design projects across 8 Foundation departments. These projects have resulted in 18 print documents, 8 digital designs, and more than 5 apparel/swag materials.
Continue to develop brand strategy for the Wikimedia movement
Updated our logo to include free fonts, built guidelines and templates for many assets from presentations to badges, and ran14 workshops to teach staff (and Wikimedia Deutschland) how to best use our new guidelines. Interviewed community members on Wikimedia identity and analyzed results for brand perceptions.
Program 3: Identify and build understanding of new audiencesEdit
Completed. We conducted a cross-team audience reearch project that led to the identification of 400 different audiences that the Foundation serves on en-wikipedia. These were developed into a framework and prioritized for future research: Communications/Audience research/Process
Canceled. This goal was set before an audience growth lead was hired and was changed after her arrival. The project described under "define target audiences" was completed. The next phase of this work will be to conduct research on the highest priority identified audiences.
We are taking a different approach to better understanding our users. In that process, we may develop feedback loops, but we need to assess the best methods given our objectives.
Program 4: Increase awareness and product adoption among specific audience segmentsEdit
Canceled. This goal was set before an audience growth lead was hired and was changed after her arrival. When that person was hired, we learned that the need for audience understanding was higher than product promotion at this time. The insights from the audience research project will inform future outreach to specific audiences.
We decided on a different approach once we hired the new audience development manager.
Program 5: Grow and improve audience and engagement on the Wikimedia digital channelsEdit
Facebook pic frame campaign explored new capabilities and ads; Twitter media views up 61% Q/Q, 121% Y/Y; newest accounts grew well with Instagram up 10%, Pinterest up 8%; Facebook group of Wikimedians growing and providing good feedback to Foundation.
Create social campaign prototype
Completed. Facebook pic frames were adopted by 14,000 and seen by 5 million; process shared with community and discussed with Facebook.
Completed. We put eight Wikipedians up as our profile pic, telling why they love Wikipedia, profiled 17 more in the blog. "Why I Edit" and the community digest have given us more ways to show Wikimedians' work.
Increase Wikimedia blog readership
Pageviews down 15% Q/Q, but still up 11% Y/Y. Blog visitors down 19% Q/Q, but up 21% Y/Y. "Wait, What?" feature launched to bring popular topics.
Facts Matter reached 95K views via social; Wiki Loves Monuments got 60K views as a GIF – a format never tried before.
More than 14,000 adopted Facebook pic frame, while 8,000 posted with the new hashtag #ilovewikipedia; Facebook engagements down 4% Q/Q, but still up 129% Y/Y; 63K media views on Twitter much higher than ever before.
Connect with growing communities
Completed. Questions posed to India community helped show how they see Wikipedia; added dozens of new Wikipedians to Facebook group from Asia and Africa.
Note: Program 6 added after Annual Plan published in June 2016.
Goal: Support movement strategic direction through communications
Support the ED and Core team in developing and running a process to determine the strategic priorities of the Wikimedia movement
Supported drafting and publishing of the outlines to share with community. Supported strategy conversations at Wikimedia events to gather feedback. Developed communications strategy to increase awareness for upcoming strategy process.
Helped draft the process and built presentation for November Board meeting. Worked with Community Engagement to speak with Wikimedia communities. Continuing to support the core team with messaging and updates.
We advised individual community members, user groups, chapters, and others on the application of our open Trademark Policy, with a goal of supporting the communities’ work in furtherance of their mission.
Granting mission-oriented licenses
We reviewed requests to use trademarks and drafted licenses when uses furthered or promoted the mission.
Enforcement against misuse of marks
We streamlined takedowns with Fiverr to reduce paid editing on the projects. We sent cease and desist letters to those misusing Wikimedia trademarks.
Maintain and defend project marks worldwide
Trademark registrations - We register, maintain, and defend project trademarks to protect against improper use.
Program 2: Support community members in unfair litigation while defending our missionEdit
We set up policy newsletter to better inform the community and allies about Wikimedia policy work. We hosted speaker event on digital policy topic to bring awareness to policy issues important to the movement. And we completed the public copyright strategy consultation to identify and prioritized copyright issues important to the movement.
Planning, drafting, and execution
We gave a keynote at the launch of the Nordic Centre for Internet & Society (Oslo) to introduce awareness of free knowledge and online collaboration to the research agenda. We gave input at launch of informal copyright reform alliance convened by Copia/Techdirt to make sure the needs of free knowledge are incorporated. We provided our views on copyright issues in TPP in a Reddit AMA organized by Fight For The Future to raise awareness of our stance. We updated content on public policy portal with community input to reflect work we have done in the past year and provide accurate information for volunteers. We participated in Internet & Jurisdiction Conference to voice WMF viewpoints and concerns, held workshop at Internet Governance Forum 2016 to build alliance for policy around Wikipedia as a public space, and presented panel at WikiConference North America to increase awareness and understanding of policy work. We also responded to the newly released EU Copyright reform package to explain its potential impact on Wikimedia projects and free knowledge.
Objective 3 (unplanned):
Delivery of censorship toolkit to aid community to avoid gov't and other censorship.
On track (confidential)
Program 5: Maintain transparency in our defense of content and user privacyEdit
Resist demands to remove content or provide nonpublic user information
We received 184 requests to remove or alter content and only granted 0; received 12 DMCA notices and only granted 4; received 12 requests for nonpublic user information and only partially granted 1; received 1 subpoenas and granted 1.
Maintain and defend Foundation trademarks worldwide
Trademark registrations - We register, maintain, and defend Foundation trademarks to protect against improper use.
Draft, enact, and ensure compliance with best practices for privacy
We drafted, enacted, and ensured compliance with best practices for privacy.
Evaluate and improve our data retention practices to ensure we are storing information safely
We continued to evaluate and improve our data retention practices to ensure that we are storing information safely.
Provide useful information about our privacy practices to community members and the public
We updated information on survey privacy practices, to provide community members conducting surveys with useful information.
Negotiate, review, and approve contracts the Foundation signs
We protected WMF and minimized risk through contractual agreements.
Train Foundation staff on legal policies and procedures
We continued to create and conduct staff training on legal policies and procedures.
Support the Board of Trustees
We supported scheduled September, November, and December Board meetings, updated the Board on results of CC 4.0 consultation, created and delivered to Board four potential candidate profiles to narrow focus and prioritize recruiting for open appointed seats, and created and delivered initial draft of board recruiting materials.
Update Wikipedia copyright license to Creative Commons 4.0
We held a successful community consultation on whether to upgrade our license to CC BY-SA 4.0 and communicated the community approval to the Board.
Develop and maintain procedures to comply with employment law
We continued to develop and maintain policies and procedures to comply with applicable employment laws.
Goal: Provide core Finance & Administrative support for the Foundation
The Annual Planning retrospective and development process has generated a lot of new insights and learnings within the Foundation. For example, the F&A team has learned:
More time is needed early in the Annual Plan process to collaboratively identify cross-functional programs and activities
Key definitions need further refining and clarification, ie. Program, Goal and Outcome
New tools are needed, particularly for travel and conference budgeting, Capital Expenditure budgeting, and program narratives
There was a lack of clarity on how certain budget items should be budgeted (ie. certain travel expenses, team development expenses, etc.)
Improve Annual Planning
F&A is highly focused on improving the Annual Planning process within WMF. F&A partnered with the Team Practices Group and Learning and Evaluation departments to conduct a retrospective following the completion of the 2016-17 Annual Plan. An outcome of the retrospective was the hiring of an outside consultant to support the improvement of Annual Planning within WMF.
The consulting firm RSM was hired to develop an Annual Plan with better planning tools, a greater emphasis on programmatic planning, continued use of the SMART framework, and continuing transparency. The consulting firm has engaged teams across the foundation and will provide recommendations and an implementation roadmap for an improved Annual Planning process in time for the beginning of the 2017-18 Annual Plan development (early Feb).
Improve Annual Plan reporting
The F&A group partnered with Community Engagement, Legal, and Product to develop the Mid-Year report. It is the first iteration of reporting on our Annual Plan. We expect that the work being done to improve the Annual Plan process will influence future Mid-Year reports.
A cross departmental group led by Legal and Communications developed a quarterly reporting process that will be published here: https://meta.wikimedia.org/wiki/Wikimedia_Foundation_Annual_Plan/Quarterly_goals
Improve Community Finance
In collaboration with Community Resources and FDC, identified high priority guideline and resource areas related to budgeting, defined requirements and created a project timeline to implement in 2017-18 APG round 1.
Consulted with two affiliates to resolve a small outstanding financial misappropriation issue and support implementation of new accounting and financial procedures.
Communicated financial requirements needed to create a formal Multi-Year agreement.
Improve Internal Financial Management and Reporting
Consulted with budget owners to provide actual to budget updates for their individual departments, and check in for any upcoming changes with financial impacts.
Worked with budget owners in need of additional fundings for new objectives / projects
Worked with Procurement to monitor/control new/recurring services indicated within our Annual Plan.
Increased financial stewardship by repurposed permanent under-spendings to objectives / projects that better serve the community (global reaching) and Foundation (team/process improvement).
The Independent Audit Report for FY 2015-16 was completed on schedule on Sept 27, 2016. WMF received a clean audit opinion.
Complete Form 990
The 990 tax filing is in progress and expected to be completed on schedule in April.
Secure an Office Lease
The WMF lease is set to expire on 9/30/17. The Wikimedia Foundation has been reviewing our office space options in consideration of our current lease expiring in September 2017. After identifying and reviewing our needs with people working in the office over the course of several months of analysis and discussions, several conversations with the people working in the office, and with input from our commercial real estate broker and design firm, we have decided to relocate to 1 Montgomery Street in San Francisco. This new location is a couple of blocks away from our current office, meets our anticipated office space needs, and will allow us to prevent increased spending on office space.