Wikimedia Foundation Annual Plan/2013-14
- Please see wmf:File:2013-2014 WMF Plan As Published.pdf for the official version of this document.
- Also see wmf:2013-2014 Annual Plan Questions and Answers for the official Q&A regarding that Annual Plan.
About this document:
- Amounts reflect management reporting, not generally accepted accounting principles (GAAP). GAAP amounts are noted where they appear.
- Management reporting reflects primarily cash-basis revenues and spending. As such it excludes non-cash items such as in-kind amounts and depreciation and includes total spending for capital items. Revenue projections and plan do not include ancillary revenue such as interest income, speaker fees, and misc. income.
- Restricted amounts do not appear in this plan. As per the Gift Policy, restricted gifts above $100K are approved on a case-by-case basis by the WMF Board.
- 1 EXECUTIVE SUMMARY
- 2 2012-13 Performance
- 3 2013-14 Plan
- 3.1 Overview
- 3.2 Revenue, Expenses, and Staffing
- 3.3 Activities, Goals, and Targets 2013-14
- 4 Board Resolution
- 5 Appendix
- 5.1 Editor decline could turn out to be an intractable problem.
- 5.2 Movement governance issues could detract from other priorities.
- 5.3 Grantmaking may not prove to be an effective way of achieving programmatic impact.
- 5.4 The Executive Director transition could hinder execution capability and negatively impact employee retention.
- 5.5 Proliferation of consumption-focused devices could further erode participation in the Wikimedia projects.
- 5.6 The shortage of Silicon Valley technical talent could hurt our ability to recruit and retain technical staff.
- 5.7 Wikimedia's ability to implement positive change could be constrained by actual or perceived lack of community approval
- 5.8 The international legal context could shift in ways that threaten the Wikimedia projects.
- 5.9 Readership could begin to flatten or decline, in part due to read-only aggregators of Wikipedia material, and/or the advent of other knowledge projects.
- 5.10 Revenue targets could not be met.
- 5.11 An unforeseen major expense or inability to raise revenues could use up the reserve and cripple the WMF financially.
- Background and Context to the 2013-14 Plan
In 2010-11, the Wikimedia Foundation embarked on Year One of our five-year strategic plan. 2013-14 is Year Four.
Fiscal year 2011-12 was spent capacity-building, conducting research and analysis, and experimenting. Last year we began to shift into an execution phase, which Year Four continues. We’ve learned a lot over the past three years, and we’ve been putting into practice what we’ve learned: focusing on critical activities, shutting down activities that don’t show promise in terms of impact, and investing in those that do.
We have known for several years that we will not achieve the 2015 plan targets, which were audacious guesswork, but our progress towards them is steady and real. All indicators (site uptime, site performance, time-to-rollout, number of readers, and number of articles) are moving in the right direction except number of editors, which has proven, unsurprisingly, our most difficult challenge. Today we have a far deeper understanding of the issue, have demonstrated an ability to achieve measurable impact on editing activity, have brought in new contributors through mechanisms such as mobile uploads, and have a good outlook for materially impacting the overall size and diversity of Wikimedia's community in the year ahead.
- Summary of 2012-13 Performance
Our top priorities in 2012-13 have been the Visual Editor project, which planned to roll out by July a new user-friendly editing system that wouldn’t require editors to learn wiki markup. The Editor Engagement Experimentation Team planned to find and productize ways to engage new editors. We planned to continue enhancing the user experience for mobile devices, including expanding access to readers in the Global South by creating partnerships with mobile firms enabling them to offer access to their subscribers to Wikipedia for free. We also planned to expand our grantmaking capacity by launching a volunteer-driven Funds Dissemination Committee (FDC) and revamping the Wikimedia Grants Program.
The majority of that work has been successful. The Visual Editor is on schedule to roll out by July. The Editor Engagement Experimentation Team has had some important small breakthroughs due to rigorous data-driven experimentation. The user experience for mobile readers improved considerably in 2012-13, and we launched new contribution-supporting functionality including a Wiki Loves Monuments upload app and other mobile uploads functionality. Mobile readership continues to show strong growth. Wikipedia Zero has proven a tougher slog than anticipated; our targets called for 600 million people to be given access to Wikipedia via the project by the end of the year, generating 200 million page views monthly. In reality, though, we will do less well; by the end of July we expect to have given 500 million people access, with page view results well short of expectations.
In 2012-13 the Wikimedia Foundation put a great deal of organizational energy into grantmaking, leading to the successful establishment of the Funds Dissemination Committee and revamp of the Wikimedia Grants Program. The Wikimedia Foundation also began, in 2012-13, to facilitate the support of legal defense and guidance for editors when appropriate, and we successfully brought Wikivoyage into the Wikimedia universe. We provided support to chapters in crisis.
And, importantly, in 2012-13 we conducted an internal exercise designed to narrow our focus, which resulted in a re-imagining of the organization as primarily focused on product development and engineering as well as grantmaking, with activities not supporting those areas being shut down or deprioritized beginning in October.
- Summary of 2013-14 Plan
In 2013-14, we will continue our primary focus on product development and engineering, with a secondary focus on grantmaking.
Product and Engineering will be split into two departments, to bring additional leadership strength to the organization and to rebalance the senior team so one person doesn’t need to represent both disciplines. We will invest further staffing resources in both areas, especially in historically under-resourced domains (data analysis and analytics support, design, operations support for projects, security engineering, and community support). This will enable our core engineering projects to launch more quickly and at a higher level of quality in the future than resources have allowed to date.
Grantmaking will continue iterating its processes, and we will make a small new investment in programmatic evaluation capability, with the goal of equipping grantmakers to make good decisions about where to invest in order to increase strategic impact, and to help people who run programs to share best practices with the goal of achieving better results. We will also make a small investment in additional governance expertise to help the WMF assess the organizational state of the chapters and other fundingeligible entities, and to enable support of chapters-in-crisis without pulling focus from other work.
In general 2012-13 has gone well. Financially the Wikimedia Foundation is in excellent shape. We had some important successes in Product and Engineering, and the FDC has gotten off to a very good start.
Revenue, Expenses, and Staffing
In 2012-13 our plan was to increase revenue to $46.1 million from a revenue projection in 2011-12 of $34.8 million, for an increase of 32%. We have exceeded this goal, with projected revenue for 2012-13 of $50.9 million. This projected revenue number includes $5M for the reserve, which was raised this fiscal year instead of next fiscal year.
The fundraiser was done in two parts with an end-of-2012 fundraiser focused on eight countries (Australia, Canada, France, Germany, New Zealand, Switzerland, the United Kingdom, and the United States) and a second fundraiser in March 2013 for the rest of the world. The end-of-2012 fundraiser was by far our shortest to date. We showed banners for nine days and then experimented with a new feature: showing banners only to users who hadn’t seen them before, and only for a limited number of views (usually one or two). By comparison, in 2011 the fundraiser showed banners to all readers in nearly every country in the world for 46 days.
The decision to split the fundraiser into two parts was deliberate. We wanted to use the end-of-year fundraiser to experiment with new types of messaging and banners, primarily in English, and then use several months afterwards to customize and localize the messaging for other audiences. This was very successful; it enabled us to reduce annoyance to readers by showing them fewer banners, while still growing revenues. In the future, therefore, we will aim to fundraise more consistently throughout the year, rather than focusing all our energies on a single end-of-year push.
In 2012-13, the Wikimedia Foundation planned to increase spending 33% from 2011-12 projections to $42.1 million. The goal was to fund the FDC grants process and build additional capacity in Engineering. Our actual spending is projected to be under budget due to lower than planned grantmaking via the FDC process, cost savings from Narrowing Focus, and lower than planned hiring in Product and Engineering. Wikimedia Foundation 2013-14 Annual Plan Page 4 of 29 We anticipate a 2012-13 year-end projected staff headcount of 167, compared with a planned end-of-year headcount of 174. The end-of-year headcount represents an increase over 2011-12 of 40%, and is 4% below plan.
Recapping 2012-13 Finances
|2012-13 Plan||2012-13 Projected||% Variance from Plan|
Activities, Goals, and Targets 2012-13
In 2012-13, the Wikimedia Foundation did its normal work of supporting the continued operations of the Wikimedia projects, including:
- Managing the hosting of the projects;
- Maintaining the MediaWiki software, supporting volunteer contributions to it, and developing new functionality to support existing and prospective editors;
- Managing/facilitating global media and public relations and issues management;
- Supporting volunteers in a variety of ways, such as helping in facilitating legal defense to community members in appropriate cases, serving as liaison with various community leaders and groups (such as ArbCom and Ombudsmen Commission), and facilitating community consultation on various WMF policies;
- Defending aggressively and successfully WMF and our projects against legal threats, including numerous outside efforts to censor content, handling DMCA takedown requests annually, stewarding the trademark and privacy portfolios, supporting WMF and community governance; and managing a variety of other legal and community processes;
- Giving out funding to other entities and individuals in the Wikimedia movement to enable work that helps fulfill the mission;
- Maintaining the Wikimedia Foundation's status as a 501(c)(3) public charity and complying with all applicable regulations;
- Fundraising globally to fund the above and a reserve ensuring the projects' continued existence.
Additionally, the 2012-13 plan called for us to continue our focus on combating editor decline via the Visual Editor and our editor engagement work and on increasing mobile readership, as well as calling for us to launch the Funds Dissemination Committee (FDC), a new volunteer-driven body aimed at supporting the Wikimedia Foundation in giving out money enabling qualified Wikimedia entities to do important programmatic work.
During the first quarter of 2012-13, the Wikimedia Foundation carried out an exercise in assessing the strategic alignment and impact of its activities; this resulted in the board-approved Narrowing Focus strategy. Narrowing Focus redefined the Wikimedia Foundation as an organization focused on engineering and grantmaking, and confirmed four major priorities for 2012-13, related to the Visual Editor, Editor Engagement, Mobile, and Grantmaking. Below is what we set out to do in those areas and the results we achieved.
2012-13 Goals: Launch a new default environment for Wikipedia projects that does not require markup. Limited English Wikipedia release for real-world editing in December 2012. Deployed to majority of Wikimedia wikis and ready for default usage by July 2013.
- December 2012 release: Done. Visual Editor was released in December 2012 as an alpha-release in production on the English Wikipedia (supporting basic content editing but not yet templates and citations) as an opt-in user preference and has since been made available in 14 additional Wikipedia languages.
- July 2013 release: On track. We are still planning to change the default for editing to Visual Editor by July, while retaining the old editor as a prominent “edit source” tab in the interface. Work has progressed to procure required hardware, resolve bugs, and add missing core editing functionality, especially for templates and citations.
Editor Engagement (Features)
2012-13 Goals: Implement shallow notifications system for user-relevant events by end of Q2. Launch new user-to-user messaging and scalable notifications system by end of Q4.
- Notifications system: Done. We released prototypes of Echo (the notifications system) to mediawiki.org by the end of Q2 and deployed a scalable implementation to English Wikipedia in May 2013, with other projects/languages to follow.
- Messaging system: Delayed. We’re still focused on ensuring the notifications system is stable and meets all user needs before shifting resources to messaging, but we have built an interactive design prototype of the new messaging system (Flow) and begun community outreach and research activities. We expect development on Flow will kick off before the end of the fiscal year, but we are unlikely to launch more than experimental production functionality before the next fiscal year.
Editor Engagement (Experimentation)
2012-13 Goals: Support editor engagement work with rapid experimentation and directly prioritize the most successful experiments. Conduct a minimum of 15 product and community experiments designed to directly increase new editor engagement and retention.
- Rapid experimentation: Done. The team has launched multiple features, including a revised Account Creation page, the first onboarding features for new account creators, and various tools to measure performance of these features.
- # of experiments: Done. We completed a total of 12 A/B tests and 4 cohort analyses across a total of six projects (Getting Started, Guided Tours, Account Creation User Experience, Donor Engagement, Community Portal Redesign, Post-Edit Feedback.) We have, however, shifted away from disconnected experiments to working within a user lifecycle framework. We measure the ongoing success of the team based on impact on key metrics such as % of attempted account creations that lead to accounts registered, % of account creators who make at least one edit, etc.
For more detail on Editor Engagement, see the Targets section on pages 7 – 9.
- Develop “mobile photographers” as the first cohort of mobile contributors. “Wiki Loves Monuments” (September 2012) will be the pilot; if successful, we will integrate uploading functionality more deeply into the mobile app/web experience.
- Experiment with micro-tasks and simple text manipulation. We can’t assume that full-page editing is the right answer for the mobile format. By the end of Q3 we will start piloting mobile editing/micro-contribution functionality and begin prioritizing successful contribution features.
- Wikipedia Zero: Mobile partnerships with operators in the Global South to offer Wikipedia for no data charges to their users and generate interest in Wikipedia. (See Targets below.)
- Mobile photo uploading: Done. We launched the Wiki Loves Monuments app for Android as planned. We then implemented photo uploads for the mobile web (articles missing images have a call-to-action) and launched dedicated photo upload apps for Android and iOS.
- Micro-tasks and simple text manipulation: Partially done. We’ve launched mobile editing in beta and have also added categorization features to the mobile upload apps. We’ve not yet focused on other micro-contribution features, but have prioritized work on uploads and basic improvements to the mobile experience instead.
- Wikipedia Zero: Done. We have signed up five major mobile carriers (Orange, Telenor, Saudi Telecom, Vimpelcom, and Axiata) for our Wikipedia Zero program, giving free access to Wikipedia to 410 million mobile subscribers in 37 Global South countries. The program has launched in 13 of those countries, with free access available to 125 million subscribers.
For more detail on Mobile, including Wikipedia Zero, see the Targets section below.
2012-13 Goals: Establish the FDC, improving grantmaking so it is strategic, impactful, and accountable.
- Done. In 2012-13, we established the FDC, a board-selected committee of seven volunteers (to be expanded to nine in 2013), charged with the responsibility for making recommendations that allocate general funds to support the annual plans of movement entities. In its first year of operations, the FDC proved itself to be responsible, prudent, and fair, and has disseminated $9.17M to the Wikimedia movement (including an award of $4.5M to the Wikimedia Foundation itself). Community response to the FDC has been positive despite several fund-seeking entities receiving substantially lower funding than requested; to date, only one formal complaint has been filed with the board representatives to the FDC regarding a recommendation.
- Stabilize number of active editors (all projects except Commons) to 86,000 by July 1, 2013 from 85,000 in March 2012.
During March 2013, there were a total of 88,284 active editors across all projects, excluding Commons. However, due to seasonality, we do not expect the July 1, 2013 target of 86,000 to be met. Through improved onboarding, we have achieved statistically significant results in helping a greater number of new users to make their first edit after registration (from 23.5% to 25.6% of new accounts on English Wikipedia, May 2013 split test). Our priority is to move more users across the threshold to 5 edits, thereby increasing the population of active editors.
- Increase the number of contributors who make at least one upload to Wikimedia Commons from 18,600 in March 2012 to 25,000 in June 2013, including 1,000 mobile uploaders per month (from 0).
During March 2013, a total of 20,176 contributors made a least one upload to Commons. It is possible that the target of 25,000 contributors by June 2013 will be reached. Seasonality and the continued growth rate we've observed throughout the year will determine whether this goal is achieved. Hiring of the multimedia engineering team was deferred to focus on core priorities, so most of this growth is organic or due to incremental usability improvements.
The mobile uploaders target of 1,000 has been surpassed earlier than anticipated. As of May 20, a total of 1,782 users uploaded at least one photo via their mobile device in the previous 30 days. Wikimedia Foundation 2013-14 Annual Plan Page 7 of 29
- Continue to expand participation in the Global Education Program from 79 to 150 classes with at least 50% female participation, leading to an increase in quality content added by students from 19 million characters in 2011-12 to at least 25 million characters in 2012-13.
Participation in the program was expanded to 128 classes with an average of 74% female participation. Collectively, students added just over 24 million characters (24,131,766) in 2012-13.
- Reach 4 billion mobile page views per month by June 2013 with at least 15% from the Global South.
During March 2013, there were 3.3 billion mobile page views. Given the year-to-date growth rate of mobile, it is unlikely that the target of 4 billion mobile page views by June 2013 will be met.
- Serve 200 million page views/month at no charge via Wikipedia Zero partners in the Global South.
12 million page views/month were served at no charge via Wikipedia Zero partners in the Global South.
In addition, in 2012-13 we achieved the following:
- In response to a community request, we supported the move of Wikivoyage to become a Wikimedia project. This included defending two Wikimedians against lawsuits and then supporting a complex migration of existing heterogeneously configured wikis onto Wikimedia servers.
- In partnership with Wikimedia Deutschland, we supported the development of a new Wikimedia project called Wikidata, for which we provided code review and architectural feedback, and operationally supported the launch. As of March, there are more than 4.5 million items in Wikidata, and it is now used to provide interlanguage links and structured data to all Wikipedias.
- We developed the Universal Language Selector, which integrates user interface language selection, font delivery, and input method selection, and deployed it into production use to several of our multilingual wikis, with a larger scale rollout planned before the end of the fiscal year. We’ve enhanced translation tools with translation memory support and a fully redesigned user interface. Input methods, language selection, web fonts, and client-side internationalization code have been componentized as jQuery plugins to broaden community development. 150+ input methods are supported.
- We developed and deployed support for Lua scripting in templates, radically improving rendering performance of pages with a large number of complex templates. For example, the rendering time of citation templates on a page with >300 citations in English Wikipedia has been reduced from 18 seconds to 3 seconds.
- We developed and launched the international legal fees assistance program designed to help protect administrators and functionaries when appropriate - a program which complements our already existing policy to assist contributors in special circumstances where their free speech on wiki is threatened by private or government authority. We have provided a wide variety of support, including legal fees in certain cases, for individual community members threatened with legal action. We won or favorably resolved six cases protecting user content and free licenses outside the United States.
- We launched an API for geo-coordinates associated with Wikipedia articles and other content.
- We launched “New Pages Feed,” a feature designed to improve the interaction between new page patrollers and new contributors, and to make new page patrolling more efficient.
- We supported the Wiki Loves Monuments launch with improvements to Upload Wizard and integrated Flickr import functionality into Upload Wizard.
- We launched a new HTML5 video player with WebM support and transcoding of uploaded originals to derivatives in multiple resolutions.
- We expanded volunteer access to MediaWiki core (from one non-WMF core maintainer to 11 between March 2012 and March 2013); developed a new mentorship program for volunteer developers; organized participation in Google Summer of Code and the Outreach Program for Women; organized hackathons; and presented at various tech conferences.
- We identified and developed closer relationships with dozens of active, highly-credible non-English Wikimedians around the world: people who have been involved with the projects for a long time, who have not been traditionally recognized, but who regularly contribute large amounts of quality content.
- We pushed out more information about the projects to more people than ever before, increasing our number of blog posts by roughly 10% (328 posts in the first 9 months of 2012-2013, up from 291 in the same period in 2011-2012), increasing our Facebook following by roughly 15% and increasing our Twitter following by more than 50%.
- We launched the Program Development Group, a small team designed to educate program leaders about how to assess programmatic impact/effectiveness and to equip Wikimedia grantmakers to make good decisions about where to invest resources.
- We supported the investigation of the UK conflict-of-interest issue. We helped the Wikimedia UK Board select an independent investigator, advised them on media and PR issues, and participated in the review itself. We then developed conflict-of-interest guidelines for the movement overall. We also advised and supported three other chapters-in-crisis.
- We created and implemented new legal and financial safeguards for our grantmaking, including developing new agreements with a stronger focus on good governance and accountability.
- We carried out an internal Wikimedia Foundation audit of compliance with Florida non-profit law and procedure, which resulted in amendments to the Wikimedia Foundation bylaws and to various practices of the organization.
- We made significant advances towards building a global Wikimedia trademark portfolio, aggressively retrieved our domain names from cyber-squatters, and handled about 200 trademark usage requests.
- Existing grantmaking practices were refined to be more responsive and effective. The backlog of requests to the Grants Program was eliminated, with review of proposals and reports now up to date.
Product Development and Engineering have achieved major milestones in the past year, and we will build on those by investing further in this area. We will split Product and Engineering into two departments to bring additional leadership strength and to rebalance the senior team so it better reflects the core work of the organization. We will invest further staffing resources in both areas, particularly product development. This will enable faster launch of new features.
In 2012-13 we successfully launched the Funds Dissemination Committee, and in 2013-14 we will continue to develop our grantmaking practices. Currently we have a good framework for distributing funding throughout the movement. In 2013-14 we will invest a small amount of new resources into grantmaking, including a modest new investment in programmatic evaluation, designed to help equip our grantmakers to make good decisions about where investments will achieve strategic impact, and to equip people who run programs to achieve better results. We will also make an additional investment in additional governance expertise to help the Wikimedia Foundation assess the organizational state of the chapters and other funding-eligible entities, and to enable support of chapters-in-crisis without pulling focus from other work.
Revenue, Expenses, and Staffing
In 2013-14 our plan is to decrease revenue to $50.1 million from a 2012-13 projection of $50.9 million, a decrease of 1.6%. Our revenue plan was calculated to fund investments in the Wikimedia Foundation (primarily product development and engineering) as well as increasing the amount of money given out by the Wikimedia Foundation to other movement entities in the form of grants. We believe that if we chose to, we could increase revenues more than is reflected in this plan, but we believe this target reflects an appropriate balance between funding growth while minimizing annoyance to the readers of the projects. This plan adds no money to the reserve, as an additional $5 million in funding to the reserve was added in the current fiscal year. To achieve the revenue target, we will continue to innovate to increase the effectiveness of our messaging, and we will improve translations and localizations of the banners, enable mobile donations, and use video better.
In 2013-14, we plan to increase spending 30% from 2012-13 projections of $38.5 million to $50.1 million, three percent lower than last year’s increase of 33%. We are planning for a 2013-14 year-end headcount of 192, an increase over 2012-13 headcount of 15%, representing significantly lower staffing growth than in previous years.
2013-14 Plan Finances and Staffing
|2012-13 Projections||2013-14 Plan||% Increase |
|Expenses: WMF||$32.3M (ii)||$42.1M||30% (iii)|
|Expenses: Grants||$ 6.2M (iv)||$ 8.0M||29%|
|Contribution to reserve||$12.4M||$ 0.0M||(100%) (v)|
|Reserve at end of year||$37.8M||$37.8M||0%|
|Staffing at the end of year||167||192||15% (vi)|
Total Spending by Functional Area
In 2013-14 spending is planned to increase throughout the organization except in Management & Governance, where it is flat. Each department’s share of the overall expenditures remains steady in 2013- 14 compared with 2012-13, except that the overall proportion of spending by Product & Engineering increases two percent compared with the previous year, with the overall proportion of spending by Legal and Community Advocacy, Human Resources, and Finance & Administration decreasing by two percent.
In 2012-13, the year the Wikimedia Foundation launched the Funds Dissemination Committee, we set aside $6.7M for it to disseminate to eligible entities (excluding a request for $4.5M for the Wikimedia Foundation). This represented an increase over the previous year’s funding for eligible entities of 7%. However, based on the FDC’s assessment of proposals received from fund-seeking entities, it recommended dissemination of $4.67M (excluding the WMF request of $4.5M). Given the large variance between planned and projected spending in grantmaking, in 2013-14 we have increased the budget for all Wikimedia Foundation grantmaking activities by 29% against projected spending, rather than against plan. The increase for project grants (rather than FDC grants) is largest, because this is where we believe we will and should see the most growth in the coming year, in supporting emerging and developing entities.
Grants Spending Compared to Wikimedia Foundation
A note on our grantmaking: As we build out the WMF grantmaking strategy, 2013-14 is intended to be a year for setting baselines for our internal performance as grantmakers, including our administrative costs in distributing grants. It is currently difficult to measure ourselves against other grantmakers, since we are unusual in the grantmaking industry for a number of reasons: most particularly the extent to which our grants process is participatory, transparent and community-led, as well as the fact that we give grants internationally and to individuals. All of these key factors make our grantmaking significantly more cost and labor intensive than conventional domestic grantmaking in the U.S. For 2007-2009 (the most recent years for which there is currently any data), the median expense ratio for staffed US-based independent public and private foundations was 8%. Foundations that engaged regularly in international grantmaking had expense-to-qualifying distribution ratios that were on average 16%, i.e. twice as high as those that did not.
For the Wikimedia Foundation, we are estimating that our expense ratio for the grantmaking team will be 28% in 2013-14. Since there is no disaggregated data for international grantmakers currently available, we will reach out to international foundations with similarly sized grantmaking budgets over the next year, to set some informal benchmarks for ourselves that will help us execute our grantmaking goals as effectively as possible.
2012-13 Spending (Projected)
Compared with 2012-13 and 2013-14 Plans (see original image)
|VP of Product||269||266||(3)||-1%||1%||236||(30)||-11%||0%|
|VP of Engineering||514||509||(5)||-1%||1%||452||(57)||-11%||1%|
|Management & Governance||782||769||(13)||-2%||2%||769||-||0%||2%|
|Legal & CA, HR, Fin & Admin||8,224||9,138||914||11%||d||24%||11,001||1,863||20%||dd||22%|
The 2013-14 plan calls for adding 25 positions to the WMF headcount, an increase of 15%. Staffing investments follow the “Narrowing Focus” emphasis on engineering and grantmaking, and are comprised of 19 additional positions in Product Development & Engineering (mostly data analysis and analytics support, design, operations support for priority projects, security engineering, and community liaison support), five additional positions in Grantmaking and Program Development (supporting program evaluation, grantmaking, movement governance, and Wikipedia Zero), and one additional position in Communications (a role designed to support community translation work).
Product and Engineering have built into their plan a stage gate; five of their 19 positions are locked to hiring until January 1, 2014. If at that point they are achieving appropriate progress against targets and have filled all scheduled medium-high priority requisitions to date, they will “unlock” the last five positions for hiring.
Staffing by Functional Area
Activities, Goals, and Targets 2013-14
In 2013-14, the Wikimedia Foundation will do its normal work of supporting the continued operations of the Wikimedia projects, as described earlier in this document.
Activity planned for the four major priority areas for 2013-14, the Visual Editor, Editor Engagement, Mobile, and Grantmaking, is detailed below.
Please note that the Wikimedia Foundation needs to preserve the ability to be responsive to a fast-changing product environment and to a large and diverse community, therefore, the plans below may change. In order to create internal and external accountability in that context, the Wikimedia Foundation maintains a process of quarterly reviews of strategic high-priority initiatives. These reviews are used to assess progress, resolve blockers, and agree on changes to the plan. Minutes are publicly captured.
Also, in addition to the priorities below, the Wikimedia Foundation undertakes other critical work. This includes keeping the Wikimedia sites and services running and ensuring continued reliability and uptime; additional features development in areas such as multimedia and language engineering; driving critical improvements to the MediaWiki technology platform and site architecture; supporting and nurturing the Wikimedia technical community; providing technical infrastructure for volunteer developers; developing analytics infrastructure and data products enabling data-driven decision-making; providing legal defense for the projects including anti-censorship work and the handling of DMCA takedown notices; stewarding the privacy and trademark portfolios; supporting a variety of community processes; serving as liaison to community leaders and groups; facilitating community consultation on various issues; supporting the community of volunteer translators; managing global public and media relations; running a small number of programmatic activities such as the Global Education program; maintaining the Wikimedia Foundation's status as a 501(c)(3) public charity; and organizing and improving Wikimedia's fundraising drives.
As per the 2012-13 plan, we will deploy the Visual Editor as the default editing environment (invoked when clicking the “Edit” button) to all/most Wikipedias (some language-related blockers may remain) by July 2013. Provided we achieve that goal, by July 2013 we’ll have an editing environment that will be of sufficient quality for new contributors and for at least routine edits by experienced contributors.
To increase usage and to support rollouts to remaining wikis, we expect that significant time in 2013-14 will need to be spent on the following:
- implementing required features;
- fixing bugs and improving language compatibility;
- improving performance.
As Visual Editor becomes a robust, stable, and performing default editing environment, we will shift our attention to a new frontier: real-time collaboration and chat. Collaborative editing environments (Google Docs, Etherpad, etc.) have proven that this combination is very powerful.
In the Wikimedia context this will help address edit conflicts and inefficient collaboration on articles that receive attention from multiple users at the same time. It will also create wholly new opportunities for engagement and mentoring.
Our efforts to increase recruitment and retention of new contributors can be divided into two main workstreams:
- system-level improvements to the Wikimedia user experience which are complex, take significant time and effort to develop, may not produce immediate impact on editor engagement targets but are expected to do so in the long run, and which are needed for the long-term sustainability of our communities. This work is done by the Editor Engagement (“E2”) team.
- focused interventions exploring engagement opportunities that exist today and that can be leveraged with relatively low technological effort, with the aim to achieve near-term impact. This work is done by the Editor Engagement Experiments (“E3”) team.
System-level improvement goals (“E2”):
In 2013-14, we will modernize Wikimedia’s discussion system. The existing discussion systems and ad-hoc page/template-based workflows represent a critical barrier to contribution for new contributors.
Improvements in this area are also a key dependency for future work, as almost all improvements to the new user experience depend on successful communication among users.
This project is known as “Flow,” and design/prototyping work is already in progress. The complexity of this effort is very high. Talk pages are fundamentally unstructured, different in no significant respect from other pages in a wiki. At the same time, talk pages and similar page-based workflows are at the heart of how Wikimedia projects function today.
The roll-out strategy for “Flow” will emphasize multiple iterations and consistent user feedback in realworld scenarios. We intend to release early and often with small, limited prototypes to obtain end-user feedback. This type of frequent release and feedback will help surface likely challenges our editors will face using the new discussion system and feed/subscription framework.
Focused intervention goals (“E3”):
The team will continue to work in a highly iterative fashion, enabling it to flexibly shift gears among the following areas as the data warrant it:
- Acquisition: How do we draw new users into the signup process in a targeted way?
- Activation: How do we get new users to contribute to the site after joining?
- Retention: How do we get users who have started to contribute to continue to do so?
- Reactivation: How do we get users who were once contributing productively to return to the site?
Additional system-level changes:
There are additional areas of the user experience that need to be transformed, but progress in these areas will be difficult with the current levels of resourcing. The launch of a new team, however, always carries significant risk, especially if such an effort is undertaken before our primary objectives are met. For this reason, we are proposing to make a set of additional hires conditional upon a set of objectives being met.
This set of additional hires would be primarily focused on two areas that are known to be strongly linked to user engagement: identity and affiliation. The first area describes changes to how a user is visible in our projects, extending unstructured user pages with globally shared profile information that can be surfaced throughout the user experience.
“Affiliation” describes functionality that captures and accelerates the growth of the interest graph of our user community: As a user, I am interested in certain topics, and I possess certain skills. This information, currently at best captured project-locally in categories and templates, can be used to surface tasks the user may be interested in (“article A in topic B needs work”, “skill A is required to complete task B”). It can also be used to drive engagement by notifying the user of relevant events (“the collaboration of the week in topic A is B”, “there is a request for comments about issue X”).
Even if we begin hiring for this team in 2013-14, the full team is unlikely to be in place until close to the end of the fiscal year. Early engineering efforts will likely focus on user profile improvements but may not go past the prototyping stage.
We expect to go into the year continuing to focus on uploads as the primary work activity of the team. We will also continue experimentation and assessment regarding non-uploading contributions and microtasks. We will shift away from uploads as the primary contribution target if/when we’re reaching the point of diminishing returns, or we’ve discovered through targeted sprints, spikes, and research that a change in focus could lead to higher net benefit.
In addition to these efforts, we want to begin to at least dedicate some effort to improving the user experience for tablet users. As of February 2013, we receive about 850 million page requests per month from (recognized) tablets of which >90% are directed to the desktop sites. The desktop user experience, both in terms of support for touch input and the general information density of the UI, has a lot of potential for optimization. Such optimization could drive editing (which is more feasible on tablets than on smartphones) and increase usage.
- Grantmaking design
We will improve the design and implementation of all grants programs in order to effectively disburse funds to the movement, ensuring timeliness, transparency, accessibility, accountability, and impact. We intend to rationalize the grantmaking team’s structures, processes, and systems to support different constituent elements and levels of the movement appropriately (individuals, emerging groups and entities, thematic groups, user groups, established chapters, and movement partners). We will also create guiding principles, parameters, and communication materials for our team and grantmaking committees that will help both the team and the community make effective choices among the various grants programs and ensure our grants are impactful.
- Indicators and best practices for governance and growth
We will facilitate the creation and sharing of a strong set of indicators and best practices around movement governance, effective leadership, and organizational effectiveness. We intend to work in partnership with Legal and Finance and partners across the movement to define the parameters and practices for good governance. In particular, we need to build appropriate parameters for different levels of organizational lifecycles and growth in the movement, and share research and learning about appropriate and innovative models of organizational growth, including different ways to move from volunteer-based groups to those with staffing or other support.
- Global South and Gender Gap strategies
We intend to develop and execute an innovative grantmaking and community growth strategy for supporting under-resourced and emerging regions, languages, and communities in our movement, particularly in the Global South, that will build upon the learnings from the catalyst projects. While continuing to be global, and supporting other Global South and emerging communities, we will have a specific focus for the next two years on 8-10 geographies and languages with high potential, and build community and content through grants as well as pro-active community and leadership development. We intend to focus on the following countries and the main language communities associated with them: Argentina and Mexico (Spanish), Brazil (Portuguese), Egypt (Arabic), India (Indian languages and English), Indonesia (Bahasa Indonesia and English), Philippines (English and Tagalog), Turkey (Turkish), and Vietnam (Vietnamese). We arrived at this list through a combination of factors that we believe indicate strategic potential for our movement, including the Wikipedia active and very active editor count in those countries, the extent of internet penetration, language/country divisions, and our ability to support the existing community relatively easily, particularly through grants. We will also facilitate the creation of a strategy for narrowing the gender gap that will encourage the increased participation of female contributors, including through grants.
- Research, learning, and evaluation
In partnership with the Program Evaluation and Design team, we will develop and facilitate a strong and effective learning and evaluation framework and strategy for grants and grantee partners with both quantitative and qualitative elements of understanding and measuring good outcomes and impact. We intend to develop baselines based on the size and nature of grant, as well context (region, language, community, project, etc.); these baselines can then be used year-over-year to assess the effectiveness of our grantmaking overall and in different portfolios or programs. We will also design indicators and offer research on organizational effectiveness that will complement the program evaluation work. We will develop and facilitate platforms, toolkits, and trainings for self-evaluation and learning across different grantee partners and the movement.
- Program Evaluation and Design
In the last quarter of 2012–13, the Wikimedia Foundation started establishing a small Program Evaluation and Design team, with two purposes: to help program leaders evaluate the impact of their work and share best practices in order to design and run effective programs, and to support Wikimedia grantmakers in making good decisions about where to invest resources in order to achieve programmatic impact. In 2013– 14, the new Program Evaluation and Design team will support program leaders in learning to use evaluation to improve their program design, and will, in collaboration with program leaders, develop and share a high-level understanding of which activities have the most impact at scale on our projects. The team will also work with program leaders throughout the movement to develop “program toolkits”: blueprints for program components and processes that have proven in the past to support the achievement of impact. The focus of the Program Evaluation and Design team will be peer learning, practices sharing, and the adoption of shared evaluation methodology across the movement, in order to support the creation of programs that will achieve real impact.
- Wikipedia Education Program
The top priority for the Wikipedia Education Program will be to increase the overall impact of educational programs worldwide. In order to achieve this, the team will partner with a number of chapter-driven and non-chapter-driven, high-potential education programs in different countries. The team will also work closely with volunteers and educators in Algeria, Egypt, Jordan, and Saudi Arabia on driving impact in this high-potential region. In order to improve the sharing of learnings among program leaders from different parts of the world, the Education Program team will lay the foundation for a global education “cooperative” model that allows educational program leaders worldwide to share best practices, harness local programs' efforts to grow educational opportunities, and drive student contributions on local Wikipedias worldwide.
- Wikipedia Zero
The Wikipedia Zero team will focus on improved operations with faster launches and better scalability. Program partners will have access to marketing assets and analytics information through a new portal that will also support self-serve set-up by smaller telecommunication companies. This will result in improved marketing, education, deployments, testing, features, and product enhancement that should result in increased customer usage. The team will also explore new delivery channels to reach a larger base of users via USSD/SMS, J2ME, and other apps. The Wikipedia Zero team will also sign up at least two more major partners and launch Wikipedia Zero in at least 20 more countries in order to give more people access to Wikipedia on their mobile devices free of charge.
- By end of December 2013, after additional stability improvements and bug fixes, Visual Editor will be deployed as the default editor for most Wikimedia wikis, beyond the July Wikipedia rollout. By end of June 2014, opt-in experimental real-time collaboration and chat will be deployed to production, leading to full build-out in the default mode in 2014-15.
- By end of December 2013, user-to-user messaging will be fully deployed in production on one or more Wikipedia projects. This release would be available to all user groups, and deployments in additional languages and projects would follow soon after the initial release. By end of July 2014, visible progress will have been made towards Flow-enabled article talk discussions, a workflow language and feed/subscriptions framework, with precise milestones to be determined. However, we don't expect that we will have “Flow” on all talk-based workflows by year-end.
- By end of June 2014, we expect to achieve a sustainable increase in the Total Active Editors core metric (registered users across all projects who make >= 5 contributions in content namespaces, de-duplicated) by 2.4K per month, adjusting for seasonality and length-of-month. This target is deliberately conservative about the estimated impact of larger scale feature changes (Visual Editor, Echo, Flow, etc.) because we cannot make realistic assessments about their likely impact at this time. Is is therefore primarily driven by current data on our continuing editor engagement interventions.
- By end of June 2014, the number of unique registered users per month who contribute content to the projects from a mobile device (i.e. at least one mainspace edit or upload) will increase to 6K.
- By the end of June 2014, Wikipedia Zero will serve 35 million page views/month at no charge via Wikipedia Zero partners in the Global South.
Grantmaking and Program Development:
- By end of June 2014, we will have distributed up to $8 million to the Wikimedia movement. Funding to individuals will increase to 7% or more of total grant spending from 4% in 2012-13. Funding to emerging communities and the Global South will increase to 10% or more of total grant spending from 6% in 2012-13. Funding for grants aimed at redressing the gender gap will increase to 1% or more of total grant spending from 0.6% in 2012-13. We will develop and share indicators for organizational effectiveness and growth based on research from within the movement and other volunteer-based movements and sectors.
- By end of October 2013, we will deliver a basic, high-level impact assessment of Wikimedia programmatic activities to Wikimedia grantmakers and people who run programs. This will improve the FDC's ability to make funding decisions and also give chapters and other movement entities indicators about which programmatic initiatives to pursue in the future. By end of June 2014 we will have delivered a first iteration of in-depth reports on two to three particularly promising programs that are most likely to impact our projects at scale. We will increase Arab world studenteditor contributions on the Arabic Wikipedia by 5%. We will lay the foundation for a global education “cooperative” with at least five programs actively participating and sharing best practices in a local cooperative by the end of 2013.
RESOLVED, that the Board of Trustees hereby approves management's proposed 2013-14 annual plan, which includes $50.1 million of revenues, $50.1 million of spending, with no increase to financial reserves. If, during the year, management anticipates the reserve at each quarter-end will differ materially from the plan, the Board directs management to consult the Chair of the Audit Committee promptly. Reference: Management's currently anticipated quarterly breakdown of this approved annual plan.
- Quarterly Breakdown of the Annual Plan
|2013-2014||Q1 (Jul.-Sept.)||Q2 (Oct.-Dec.)||Q3 (Jan.-Mar.)||Q4 (Apr.-Jun.)||Total|
All amounts USD, in millions.
- Risks Considered in Developing 2013-14 Plan
Editor decline could turn out to be an intractable problem.
The number of Total Active Editors in Wikimedia projects continues to be flat. The number of Active Editors in the English Wikipedia continues to decline slightly year-over-year. Significant Wikimedia Foundation effort is tied up in addressing and ultimately reversing this trend. Should these efforts prove intractable, this represents a threat to the future health, diversity, and viability of Wikimedia’s projects.
Response: We do not believe editor decline is intractable or irreversible. It’s a complex problem, and solving it is requiring multiple simultaneous interventions of different types, but some of our activities (e.g, improvements to the user onboarding experience, the Global Education Program, uploads via mobile) are yielding results today, and we believe that others (e.g., Echo, Visual Editor) will begin to yield results in the coming year.
In the Fiscal Year 2012-13, the Wikimedia Foundation has significantly improved its ability to understand the impact of our work at a granular level so that we can accurately measure our progress. We now have the ability to measure:
- conversion rates from account creation to first edit including ability to A/B test various onboarding strategies;
- retention, productivity, and various other characteristics of any cohort of editors, allowing us to measure and compare the impact of various programs, including non-technical initiatives like the Global Education Program; and
- clicks, other interactions, and response rates for specific features and calls-to-action.
In 2012-13, we focused on the following efforts to directly address the stagnation/decline tendencies in Wikimedia’s communities:
- the Visual Editor, which is scheduled to launch in June/July 2013 and not expected to impact editor behavior until then;
- the Echo notifications system, which provides a foundation for surfacing relevant events to users on-wiki and via email;
- features to contribute photo uploads via the mobile web and mobile apps; and
- improvements to the account creation and onboarding experience, including making it easy for new editors to discover articles that are suitable for improvement.
We expected, and were able to measure, the most immediate positive impact resulting from the last two initiatives. We are successfully engaging new users to upload photos via mobile, and we have managed to increase the success rate of new account creations and the rate by which new users are completing their first edit. Our challenge here is to move more users to cross the five-edit threshold. We expect a larger and more sustaining impact of system-level changes like Echo and the Visual Editor, which will play out over coming months.
We also continued to scale the Global Education Program, Wikimedia’s world-wide effort to engage students in improving or creating Wikipedia articles as course assignments. The GEP team has successfully increased the number of participating students and classes while simultaneously reducing staff effort and the cost-per-student. In the Arabic Wikipedia alone, during the most active months of the terms, Education Program students now make up around 10% of the active and very active editors on the project.
Overall, we now have a much greater ability to understand where interventions are achieving impact, to course-correct as needed, and to work in more clearly defined increments towards growing our community. It’s a hard, multi-faceted problem, but the evidence indicates clearly that our work is making a difference, and we expect to see more progress in the coming year.
Movement governance issues could detract from other priorities.
As anticipated, in 2012-13 movement governance issues cost the Wikimedia movement significant time and resources. There were a number of governance issues in chapter organizations through the year, most notably the Gibraltarpedia issue, which engendered significant negative international press and required substantial Wikimedia Foundation managerial, legal, governance, and media relations resources to handle.
Movement governance issues also had implications for chapter payment-processing in the annual fundraising campaigns. Negotiating and managing agreements is labor-intensive ordinarily, and even more so when movement governance issues require renegotiation or non-renewal of agreements. That, combined with a dramatically more efficient and effective 2012-13 Wikimedia Foundation fundraising campaign, means that we consider it an open question whether chapter payment processing “adds sufficient value to the movement to offset the additional significant time and money required from both the chapter and the WMF.”
Our recent emphasis on good governance of incorporated entities is non-optional but it does also have a cost; we worry that we may, collectively, be taking the focus off mission work and (in effect) turning valuable community members into bureaucratic corporate officers.
Response: None of these issues is new, and the Wikimedia Foundation is discussing with other movement actors our collective best path forward. We have been having conversations about payment processing with the chapters for several years. In February 2013, the jointly-commissioned WMF-Wikimedia UK was published; it documented 50 potential areas for improvement for Wikimedia UK and, we hope, will serve as a model for other movement entities for improving their governance. In April, the Wikimedia Foundation adopted conflict of interest guidelines requiring disclosures of personal and financial interests in requests for movement resources; we hope it will also serve as a model for other movement entities. Also in April, the Wikimedia Foundation asked the Funds Dissemination Committee “to ask grant applicants to reconsider institutionalizing, or to decide to grow more slowly, thoughtfully and deliberately, with time built in for assessment and reflection, in their early stages of institutionalization.”
However, we do believe that movement governance issues will continue to require a substantial amount of Wikimedia Foundation time and energy. Therefore, we intend in 2013-14 to take a more active role in promoting proper governance in the movement. We will update and improve our governance standards for grant recipients and continue to support the Funds Dissemination Committee in evaluating, monitoring, and assessing compliance with them. And, with the help of external governance experts and the community, we will be putting forth an online wiki guide on minimum organizational governance. We also will be hiring legal governance counsel and an administrator to handle the increased workflows and provide support in this effort. In summary, in 2013-14 the Wikimedia Foundation will increase resources dedicated to movement governance oversight, in order to support good governance throughout the movement and mitigate risk to the movement’s reputation and ability to do its job, as well as the Wikimedia Foundation’s ability to do its other work.
Grantmaking may not prove to be an effective way of achieving programmatic impact.
Launching the Funds Dissemination Committee process in 2012-13 has required considerable time, energy and focus from the Wikimedia Foundation, other movement entities, and the FDC members. Thus far the work has gone very well. The framework for grantmaking seems to have been well-designed and is successfully giving out funding, and it seems to have created some significant initial value for the movement: for example, grant recipients have vastly improved their publishing of activity reports, stable governance is increasingly being recognized as an important precondition for programmatic impact, and a shared conversation has begun about what impact actually can look like in our movement. However, at this point, we’re not in a position where we can claim our grantmaking consistently funds programmatic work that’s having a strong demonstrated impact helping to achieve the mission.
We see two major challenges: 1) As a movement we don't currently have a solid understanding of what programmatic impact looks like and how it can be achieved, and 2) The movement seems to be on a path of rapidly creating incorporated, professionally-staffed entities, and it is not clear whether that is the best path towards achieving programmatic impact.
It is not within the scope of grantmaking to determine what types of entities make up the Wikimedia movement; the role of grantmaking is to support the funding of competent, effective organizations and individuals. That said, it is not clear that incorporated professionalized entities are the best way to achieve impact, and yet that is where most of our spending and attention is being directed. Currently, 96% of Wikimedia Foundation grantmaking dollars go to incorporated entities, and in 2012-13, those entities requested very large increases in funding over previous years, primarily to professionalize and institutionalize their organizations. As described above, the Wikimedia Foundation is investing in movement governance oversight in order to reduce the risks of mismanagement or poor governance, but there is still an open question about whether there are other imaginative and innovative ways of building organizational support for our global community of volunteers other than professionalization and institutionalization of movement entities.
Response: In order to help the movement achieve a solid understanding of what programmatic impact looks like and how it can be achieved, in 2013-14 the Wikimedia Foundation will launch a small new team, Program Evaluation and Design, intended to support evaluation of the impact of programmatic work in the Wikimedia movement. Importantly, this team is premised on the assumption that the Wikimedia movement is still developing and learning and is operating in an emergent space. Solid answers aren't yet known, and it will take considerable time and experimentation to develop them. Therefore, the Program Evaluation and Design team won't pretend to have all the answers; instead, it will facilitate conversations among program leaders, intended to identify and document best practices and expertise where it already exists. The team will have two goals: to equip program leaders with the information they need to direct their energies towards activities that are proving to be effective, and to equip grantmakers with the information they need to make good decisions about where funding should be directed. It's important we begin it now so that we can increase the ability of our grantmaking activities to consistently fund programmatic work that's having a strong demonstrated impact in helping to achieve the mission.
Meanwhile, in 2013-14 Grantmaking will increase funding to individuals from 4% of total grant money to 7%. This is not just appropriate because individuals are the heart of our movement; we hope it will deepen our support of on-wiki engagement and activities, and our understanding of different solutions for the concern around editor retention.
The Executive Director transition could hinder execution capability and negatively impact employee retention.
Turnover at the CEO level could harm organizational performance and retention of employees, particularly senior staff. Additionally, although the Wikimedia Foundation has experienced changes at the senior leadership level over the past half-dozen years, it has only experienced one prior ED transition, in 2007. It is possible the Wikimedia Foundation will have difficulty hiring an ED in a timely fashion, and it is possible the onboarding of the new ED will take longer, or be more difficult, than anticipated. A new ED will mean change regarding how the Wikimedia Foundation does its work, which may result in a temporary hit to the organization’s productivity.
Response: The major mitigating factor here is that the groundwork for a successful transition process has been laid. Namely, the warm and mutually supportive relationship between the current ED and the board is significant, as is the experience level of the Transition Team in navigating leadership change, and the commitment of the current ED to participate in the search for a capable, credible successor while continuing with her full workload, and the commitment of the C-level team to design, develop, and support a robust transition and onboarding program for the incoming ED. The initial announcement to the organization has not impacted current retention, we believe, due in large part to the effectiveness of the announcement plan and the strength of the overall office team.
Proliferation of consumption-focused devices could further erode participation in the Wikimedia projects.
Response: As we pointed out in the 2012-13 risk analysis, some of the technological changes and associated shifts in behavior are out of our control, but we can develop functionality that targets the specific contribution potential that new devices offer. In 2012-13, as detailed elsewhere in this plan, we developed comprehensive offerings to upload photos via mobile devices, including a dedicated app to upload photos to the Wiki Loves Monuments competition, and we launched the first mobile editing functionality. In 2013-14, as detailed elsewhere in this plan, we will expand these offerings further and grow the number of contributors on mobile devices, as well as improving the user experience on tablets.
The shortage of Silicon Valley technical talent could hurt our ability to recruit and retain technical staff.
The job market for engineers began heating up in 2010, and is now extremely competitive, particularly in the Silicon Valley area. The Wikimedia Foundation, as a non-profit, does not offer equity or large cash bonuses that are often awarded by multi-billion-dollar tech companies. The organization therefore faces a risk of not being able to hire the key engineers, designers, product managers, and other staff needed to deliver on this plan.
Response: In recognition of the challenging hiring environment, we've put increased focus on developing internal recruiting expertise for technical staff. We've optimized Wikimedia's value proposition (a combination of a world-changing mission, open source values, reasonable compensation near the mid-point of the tech sector, strong health and wellness benefits, and the opportunity to achieve huge impact as an individual) and the ways in which we communicate that proposition to the world. We've created increased flexibility to hire opportunistically, as opposed to following a strict hiring calendar.
We've also improved our ability to work as a distributed organization through team practices improvement and investments in videoconferencing and other remote staff support. Nearly half of the Wikimedia Foundation engineering/product team is not based in the San Francisco Bay Area, so we're operating with the flexibility to hire people where they are, supporting them in making the shift to work at the central location if they're willing to do it and it makes sense for a given position. This gives us increased flexibility in recruiting and hiring.
Significant effort has also gone into leadership development and career pathing, which will help Wikimedia Foundation staff see the organization as a long-term employer as well as a gateway to future opportunities.
That said, new leadership, an evolving org structure including the split into engineering/product, as well as the ED transition, may all impact retention of engineering talent. Given the factor of uncertainty in these dynamics, some planned-for hires in 2013-14 are explicitly stage-gated on the number of vacancies not exceeding a manageable amount, accounting for attrition, by January 1, 2014.
Wikimedia's ability to implement positive change could be constrained by actual or perceived lack of community approval
Our objective is to help the community become stronger, more cohesive, more loving, and more fun, and our ability to achieve that is very much affected by community willingness to support our work and to partner with us.
The 2012 Editor Survey found the community believes the WMF is doing a good job; when asked about seven specific activity areas of the Foundation, its general performance and the direction in which it is going, the overwhelming majority of opinions given by our users were positive. Reactions to the release of prototype versions of Visual Editor have been positive, suggesting that many community members appreciate such editor retention initiatives.
That said, to successfully make change of any kind requires the Foundation to maintain goodwill within the community. That goodwill could be lost quickly if the Foundation supports positions or projects contrary to the needs or desires of the community.
During 2012-13, the Wikimedia Foundation earned some additional trust and goodwill:
- The board passed the resolution approving the Legal Fees Assistance Program, which demonstrated important support to the community, both users and functionaries. We also reaffirmed our internal policy to help find support for community members who are threatened with actions to force censorship.
- We help facilitate the defense of users and functionaries and win matters that are properly perceived as protecting the community and its projects, including the Internet Brands victory, the French censorship matter, and multiple wins in foreign courts.
- We’ve increased frequency of communications through the Wikimedia blog (328 posts in the first nine months of 2012-13, up roughly 10% from the 291 published in the same period in 2011-12), including an increase in multilingual blog posts and a rise in the number of Wikimedian profiles over last fiscal year.
- Community translators now have much-improved tools to work with, including the Translate extension on Meta. The Meta translation functions now allow us to post all major reports (including the WMF Monthly report) for translation to over a dozen languages.
- Through 2012-13 the Wikimedia shop processed over 2000 online orders, added product variety including editor and multi-lingual products, carried out editor-focused sales and promotions, and added public visibility through sidebar and short central notice pushes. The shop’s volunteer recognition program launched on English Wikipedia with more than 100 nominations in the first month. The shop also supported movement events with merchandise packages and Wikimania specifically, with a conference t-shirt and on-site store for the first time.
- In addition to continuing to dedicate WMF engineering resources to code review, we’ve shifted more effort towards mentorship of promising volunteer developers in the community and enabled a larger number of them to help with code review and to merge code into our production repository. From March 2012 to March 2013, the number of volunteers with core merge access has increased from 1 to 11. In December 2012, 155 MediaWiki developers were active, a 90% year-over-year increase.
- We hosted key community stakeholders such as the Ombudsman Commission, members of the Arbitration Committee, and a representative from the stewards (and FDC), in an effort to better understand and respond to issues they're facing.
- The fundraiser was much shorter in 2013 than 2012, and was the shortest ever by far at less than nine days of solid banner plus additional days of light banner impressions only to those who had not seen any before. This responded directly to long-standing community-expressed desires that the fundraiser not be overly aggressive. In a “Thank you” campaign that invited readers to become editors, we featured a wide array of contributors from multiple Wikipedia language versions and other Wikimedia projects.
Still, investing in positive community relationships does not necessarily translate to smooth sailing on WMF initiatives, and we continue to need to carefully negotiate changes that we make to Wikimedia’s sites. As an example, the ArticleFeedbackV5 rollout on English Wikipedia (a tool to solicit and curate reader suggestions for article improvement) was discontinued after an RFC on the English Wikipedia showed strong community objections to the tool.
Response: Part of the answer is moving faster and failing more quickly (structuring projects to deliver minimally viable products as quickly as possible and to iteratively improve from there). We also intend to invest in more thoroughly understanding the non-en-WP communities, creating a center for change management, and growing goodwill. To that end, in 2012-13 we began to build out our team of two community advocates, in addition to the two previously-existing headcount, inside the Legal and Community Advocacy department. In 2013-14, the Community Advocacy team will take on a role as a center for knowledge about change management, embedding with Engineering or other parts of the organization to support them in deploying high-value strategic initiatives. We will also build a small network of contracts with community liaisons who can assist on a more transactional basis with product rollouts and related communications.
The international legal context could shift in ways that threaten the Wikimedia projects.
The Wikimedia projects’ ability to achieve the Wikimedia mission depends on people everywhere having access to a free and open Internet. Increasingly, this is under attack. We are disturbed to observe around the world governments and other parties taking steps that could inhibit people’s ability to read and/or contribute to sites like Wikipedia. This includes countries such as Bahrain, Belarus, China, Iran, Saudi Arabia, Russia, and Vietnam, which routinely filter or remove content, imprison people who create or share certain types of material online, and shut down or slow people’s Internet access. These types of actions aren’t limited to countries generally understood to be authoritarian; in recent years we have begun to see, including in democratic countries, the development of legislation which violates user privacy by surveilling their activity online, which punishes people who post certain types of material on the Internet, and which makes intermediaries legally liable for content posted through their services. Wikipedia is not immune to this general trend; in spring 2013, for example, a French Wikipedian was detained and threatened by authorities in an incident related to a Wikipedia article he had never read or edited, and the Wikimedia Foundation and Wikipedia editors are threatened with meritless lawsuits by politicians and others. Badly drafted legislation and attempts to block content are a major problem for Wikipedia; for example, in 2013 Russian authorities added the Russian Wikipedia article on “cannabis smoking” to a blacklist of sites required to be filtered by Russian ISPs, and the U.S. House of Representatives passed a bill titled “Cyber Intelligence Sharing and Protection Act” (“CISPA”) that was criticized for being potentially harmful to user privacy on the Internet. Also, increasingly, actions taken by governments, Internet service providers, and others are resulting in an online experience that is fragmented and distorted due to the user’s geographic location.
The Wikimedia Foundation does not participate in, or condone, efforts to inhibit people's access to knowledge online; we provide the same, uncensored service for everyone in the world.
Response: In 2011-12, we considered whether to significantly invest in efforts aimed at protecting the free and open Internet through a direct legislative and lobbying program. After due consideration, we conclude that neither lobbying nor public advocacy is a core area of expertise for the Wikimedia Foundation. We will continue, however, to speak up for freedom and openness online wherever we see our ability to do our job is threatened, and we will support like-minded organizations who specialize in legislative issues affecting our editors and community. In 2012-13, we began to dedicate a small amount of organizational energy towards cultivating a community of Wikimedia volunteers interested in advocacy issues, and we will continue that work in 2013-14.
Readership could begin to flatten or decline, in part due to read-only aggregators of Wikipedia material, and/or the advent of other knowledge projects.
The trend we pointed out in the 2012-13 risk analysis continues: Google, Apple and other major players strive for deep integration of content into their products to increase usage and advertising revenue. In the case of Google, a recent example is the (now disabled) Google QuickView feature on mobile devices, which makes a cached version of Wikipedia articles available with a single tap.
In addition to reducing traffic directly to our sites (which is not as important to us as it is for ad-driven sites, but which could reduce our perceived reach if we do not get analytics), this type of feature also removes any invitations to contribute.
Despite this, we are not seeing major impact on readership growth yet. From March 2012 to March 2013, total page views have increased from 16.5 billion to 20.6 billion, a 25% increase (compared with a 12% increase from March 2011 to March 2012). Mobile page views have increased from 1.8 billion to 3.2 billion in the same time period, a 77% increase (compared with a 200% increase in the prior year, but much of this was due to massive initial improvements to our mobile offerings and device detection). Non-mobile page views have increased by 19% (compared with only 4.1% in the prior year). comScore’s panels only register an increase of 489 million to 517 million unique visitors during that time period (5.7%), but the panel data has significant limitations and ought not to be taken at face value.
Response: We will continue to work with companies that integrate Wikipedia content to assess impact and, as we build new contribution features, may reach out to them and request small calls-to-action to be integrated as well.
One concern we pointed out in the 2012-13 risks analysis is being addressed: that structured data from our projects is hard to re-use and that third parties are therefore motivated to work with proprietary data providers or third party re-processing efforts of Wikipedia data. In 2012-13, we partnered with Wikimedia Germany to get the Wikidata project off the ground, which is shaping up to be a success story in community collaboration. As of April 2013, Wikidata already contains more than 10 million items, 2 million of which have associated structured data claims. With already more than 300 possible properties per item, the Wikidata project promises an ever-increasing breadth in available structured data than many alternatives.
Revenue targets could not be met.
The 2013-14 revenue targets for WMF are a modest decrease from our prior year projections. It is possible we cannot meet that target. Risk factors include the possible decline in readership in high-income countries (where most of our revenue is raised), possible loss of effectiveness of last year’s banners, and possible failure to discover a new kind of banner that works as well. A common misconception of the WMF revenue mechanism is that we can simply run banners longer to raise more money. In reality, the marginal gain from each additional day of fundraising drops rapidly after just the first week of fundraising, as the maximum pool of donors is used up. Additional fundraising days might only raise a few hundred thousand dollars each, at first. Prolonged fundraising can bring the daily total down as far as one hundred thousand dollars.
Response: We do have a track record of conservative planning; every year we have exceeded revenue targets, and in most years we underspend against the plan. We consider this prudent and responsible. It has enabled us to build a reserve of approximately $38 million by the end of the fiscal year, which we do not plan on increasing this year due to adding $5 million to the reserve in the current fiscal year. It is also true that fundraising more consistently, at very low levels, throughout the year, will reduce the risk of a critical failure during a few crucial weeks of a short annual fundraiser. Overall, we are confident we can meet this target. In the event we have difficulty, the Board Treasurer will notify the Chair of the Audit Committee and the Executive Director should reserves vary materially from the plan at the close of any quarter.
An unforeseen major expense or inability to raise revenues could use up the reserve and cripple the WMF financially.
Response: The likelihood of this happening is extremely small. The Wikimedia Foundation's costs are predictable and we have a track record of conservative planning. We have never in our major history faced a major unplanned expense. Our operating reserve is not expected to fall below six months of total expenses. In addition, we have made significant improvements to our insurance to protect against a variety of financial risk.
- Renz, Loren, Benchmarking Foundation Administrative Expenses: How Operating Characteristics Affect Spending. Foundation Center, 2011. Page ix.
- Renz. Loren. Benchmarking Foundation Administrative Expenses: Update on How Operating Characteristics Affect Spending. Foundation Center, 2012. Page 2.
- Calculated as the ratio of estimated operating expenses to estimated grants distribution of the Grantmaking team for 2013-14.
- https://meta.wikimedia.org/wiki/Metrics_and_act ivities_meetings/Quarterly_reviews
- See https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor for metric definitions.
- See http://reportcard.wmflabs.org/ for detailed numbers.
- See http://reportcard.wmflabs.org/ for detailed numbers.
- See http://techcrunch.com/2013/04/16/google-makes-mobile-search-faster-adds-expandable-sitelinks-and-experimental-quick-view-cards-for-wikipediaarticles/