March 2011 Update
What Wikimedia has accomplished in the last decade is amazing, and it’s thanks to you. Thank you for being awesome.
My name is Sue Gardner, and I run the Wikimedia Foundation – a San Francisco-based non-profit organization with a staff of sixty. We pay the bandwidth bills and buy the servers and provide legal defense for the Wikimedia projects. We also help develop technical and social solutions that support your work. The purpose of this note is to tell you about some of the work we've got planned, and also to share some new data about trends in the editing community.
Let me start with some context. By practically every imaginable measure, the Wikimedia projects are flourishing. Readership continues to increase – the Wikimedia projects are now serving more than 400 million people every month. The total number of articles, and the quality of those articles, continues to climb. Each year, more people donate than ever before. And the material in the projects continues to freely propagate: it's visible in maps and social networks; it's available on mobile phones, tablets and e-readers; it's copied by countless other sites and used in many different ways. That's all good news.
But part of my job is to be on the lookout for problems. And so, a few months ago, the Wikimedia Foundation commissioned some research to better understand the internal dynamics of our communities. The result is the Editor Trends Study (Results), which investigates five large Wikipedia languages: English, German, Russian, French and Japanese.
You may notice the study results are a little rough: not every detail is perfectly nailed down. That’s not an accident. We’re sharing these new findings early, so you can help us interpret and understand the results we’ve got so far, and start talking with others in your community about how to respond to them. We’ve also open sourced the code used to run the study, so you can check the methodology, look at the data from different angles, or examine other projects and languages.
What We've Learned
Here’s what we think the Editor Trends Study tells us: Between 2005 and 2007, newbies started having real trouble successfully joining the Wikimedia community. Before 2005 in the English Wikipedia, nearly 40% of new editors would still be active a year after their first edit. After 2007, only about 12-15% of new editors were still active a year after their first edit. Post-2007, lots of people were still trying to become Wikipedia editors. What had changed, though, is that they were increasingly failing to integrate into the Wikipedia community, and failing increasingly quickly. The Wikimedia community had become too hard to penetrate.
These general patterns also emerged in the other languages we studied.
The below chart from the study shows this quite clearly for the English Wikipedia. What it shows is the number of active editors (blue) plotted against the percentage of editors who joined in that month who are still active one year later (red). Please note that these trends hold true even when looking at new users who have completed at least 50 edits – it’s not just an increase in experimentation and vandalism.
Our new study shows that our communities are aging, probably as a direct result of these trends. I don't mean that the average age of editors is increasing: I'm talking about tenure. Newbies are making up a smaller percentage of editors overall than ever before, and the absolute number of newbies is dropping as well. That's a problem for everyone, because it means that experienced editors are needing to shoulder an ever-increasing workload, and bureaucrat and administrator positions are growing ever-harder to fill. Experienced Wikipedians have observed those changes for years: this is the first time there’s been data supporting what they’ve said.
Based on this and other research (links below), here’s what we think is happening: As successful communities get really big, they naturally suffer growing pains. New people flood in, creating an Eternal September effect, in which the existing community struggles to integrate the newbies while at the same time striving to preserve the ability to do its work. It does that by developing self-repair and defense mechanisms – which in our case, turned out to be things like bot- and script-supported reverts, deletions, user warnings, and complex policies. All those mechanisms are obviously helpful – after all, they were developed for a reason, in response to real problems. And they do their job: they do successfully help experienced editors preserve and maintain quality. But they’ve also made it harder and harder for new people to join us, which in turn seems to have made experienced editors' work harder as well. People tell me that editing back in 2001 or 2003 or 2005 was more rewarding —and more fun!— than it is today. I believe that some of that is ordinary nostalgia. But I think some of it is true.
Openness Begets Participation
I believe we need to make editing fun again for everybody: both new editors and experienced editors. Some of you will question whether we can do that without compromising quality. My personal opinion is that that’s a false choice, and a trap. Quality and openness go hand in hand: if that weren't true, Wikipedia wouldn't —it couldn’t!— exist. Wikipedia is the largest and best and most-used informational resource ever compiled in human history. Openness works.
So. Where quality assurance mechanisms hurt good-faith newbies, the answer to that is better quality assurance mechanisms, which will support quality while doing less unintended damage. Bringing in more new contributors will lighten the work load, and make the whole endeavor more pleasurable, for everyone. I also believe that we need to turn up the volume on activities that help acculturate new users and that make everyone feel appreciated and welcome.
Here are some questions we’re thinking about:
Should new editors be encouraged to share more about themselves on their userpages, so that good-faith people can be identified more easily and given support and encouragement? Should we build more automated mechanisms for editors to express appreciation for each other? Should we build automated tools for connecting new editors with experienced mentors? Do we need better tutorials? Should there be improved semi-public draft spaces, like on the Russian Wikipedia, to give new articles a chance to incubate rather than being deleted? What else should we do?
This is important work. It’ll succeed if you —the heart and soul of the projects, the people who are most active and most knowledgeable— work together, with the Wikimedia Foundation and with each other, to make it happen. The Wikimedia Foundation wants your ideas, your expertise and your support. And we hope you’ll be talking with each other too. This is about starting a new chapter in our history, opening our communities up further, while ensuring we create an ever higher quality resource for the world.
The Year Ahead
In February, we released the Wikimedia Foundation’s strategic plan, which sets our long-term priorities. Based on the strategic plan, we’ve completed a comprehensive analysis of our product priorities, the “Product Whitepaper.”
The editor trends data informs where we’re focusing our attention in the year ahead. Overall, our top priorities are focused on growing the community -- creating an environment that’s diverse and welcoming to everyone who wants to help. Some of these are projects that will increase the inflow and diversity of new editors, but we need to simultaneously increase retention. Here’s where we’re putting most of our energy this year:
- Create a visual editor: We’re creating a new editing environment for Wikimedia projects that’s simpler to use and doesn’t require users to learn any special wiki syntax. This work is just beginning, and new Lead Architect, Brion Vibber, will play a key role in helping get our software platform ready for this big change. This is a long range project – it’s a massive change, but we have to start now, to nurture a more open and diverse community together.
- Improve the newbie experience: We’re running a series of community and engineering experiments to improve the experience of new good-faith contributors. This will range from tools for expressing appreciation to community experiments around mentoring (learning from experiences like the German Wikipedia mentoring program and the Russian article incubator) to thinking about how to improve the interface for new users. Our goal is to do lots of parallel weekly experiments, and to continually feed lessons learned back into our product development process and to the community. We’ll create lots of pages where you can help with this, but feel free to comment directly on the discussion page here.
- Support community growth in developing countries: We believe lots of future growth in our projects will come from places like India, Brazil, the Middle East and Africa -- and we want to be there now, helping communities to grow around Wikimedia’s free knowledge mission. We’re helping to catalyze the communities, through both online and in-person activities. We’re also helping to remove technical barriers around text input and display in many languages. The India Programs Meta page, India mailing list, Brazil Catalyst Project Meta page and Brazil mailing lists are where much action is being discussed and coordinated right now. WMF has also expanded our grantmaking programs in support of global community and chapter initiatives.
- Serve audiences on all devices: Whether you use a smartphone, a low-end feature phone, a tablet, or you want access when offline, you should get the best possible experience. That’s essential for reaching billions of new readers, and for enabling people to edit who will never touch a PC. We’re just configuring this work now, but you can read more about our mobile strategy.
- Create a delightful experience for contributing and reviewing multimedia: Images, sounds and videos make our projects richer and better, and they are areas where we’re seeing strong growth in contributors, quality and quantity. They can also be lightweight entry opportunities: on-ramps for deeper involvement. We’ve started this work with the multimedia usability project (see report), but we’re not done yet. To avoid creating tension between quality and inflow, we’ll take a two-pronged approach that also builds out new review/moderation tools which are designed to be both socially aware and highly effective.
We’re continuing to invest in other areas, including improved discussion systems, quality review and labeling tools, and content packaging for offline use. There are also some important general site infrastructure improvements we’re working on. But the above five are our priority projects that we’re pushing with maximum effort this calendar year.
We’ll be partnering with Wikimedia’s world-wide chapter organizations in doing this work, and ultimately our success depends on partnering with you, as well.
Please get involved: look at the data and research, join our active projects, help us make our technology and our processes better. Be bold and do the unexpected - everyone is a leader.
Thanks for reading - we look forward to hearing more from you.
- Sue Gardner
- Executive Director, Wikimedia Foundation
- The Editor Trends Study was conducted by Diederik van Liere and Howie Fung and supervised by Erik Moeller. The Editor Trends Study was completed on March 10, 2011.
- See the study for more detailed breakdowns of cohorts of individual editors over time, and for data from other languages.
- Diederik van Liere, Howie Fung. Editor Trends Study. March 2011.
- Resolution:Openness. Passed 8 April 2011 by the Wikimedia Foundation Board.
- Bongwon Suh, Gregorio Convertino, Ed H. Chi, Peter Pirolli. The Singularity is Not Near: Slowing Growth of Wikipedia. In Proc. of WikiSym 2009. October 2009. Florida, USA
- Definitions of research terms used.
- Attrition pipeline (PDF) — also available as a publicly-viewable (but not editable) Google spreadsheet. This is a spreadsheet that attempts to define impediments to editing at various stages of the editor life-cycle. The editor life-cycle framework used in this document is an earlier version of the one used on this page.
- Decline Theories and Support (PDF) — also available as a publicly-viewable (but not editable) Google spreadsheet. This is a spreadsheet that documents various theories purporting to explain Wikimedia's decline in editor retention and tests them against the available research.
- In November 2009, User:WereSpielChequers ran the Wikipedia:Newbie treatment at Criteria for speedy deletion study on the English Wikipedia, in which experienced Wikipedians posed as newbies to experience the CSD process through new people’s eyes. The experiment itself was controversial, with one critic saying “I don’t believe anyone is seriously denying there is a problem with the way newbies are treated.”
- In December 2009, Mark Graham, a Research Fellow at the Oxford Internet Institute, published an analysis of the English Wikipedia’s geotagged coverage that found coverage of Africa and China to be extremely weak. Graham concluded that “it is clear we are far from running out of topics to write about.”
- The Former Contributors Survey, published in February 2010, found that about half of 1,200 lapsed editors said they stopped editing due to personal reasons. About a quarter said they stopped contributing because of issues with the community, including interactions with other editors they found stubborn or biased or bullying.
- Shiju Alex has created a statistical report on the Indian language Wikipedias, covering 2010. In it, he observes that article creation in many Indian language Wikipedias has slowed, and “more language wiki communities have started focusing on the quality than the quantity.”
- In February 2011, Sue Gardner published a blog post that collected together comments from women talking about their experiences editing Wikipedia, culled from dozens of online discussions that sprouted up in the wake of the New York Times gender gap story.
- Wikimedia fellow Lennart Guldbrandsson and Head of Public Outreach Frank Schulenburg have launched the Account Creation Improvement Project aiming at increasing the number of people who create a user account and actually start editing.
- User:Kaldari has made a script called Wikilove — with the goal of encouraging WikiLove within Wikipedia, by making it easy to add awards and gifts to people’s talk pages. You should install it! Use it! :-)