Archives for category: Wikimedia Foundation

About a decade ago I migrated into community work [1] from a non-community background. This is the guide I wish I had read back then.

[1] When I say community work, I am talking about stuff like Wikipedia: large distributed groups of people doing something together, usually online, often unpaid. Usually international, often nerdy, often (but not always) FLOSS or FLOSS-adjacent.

You are going to be doing a lot of writing. Do it well. Phone calls and Hangouts don’t scale. Face-to-face is expensive for distributed groups and therefore rare. Real-time tools like IRC and Slack disadvantage people in minority timezones. And so, in an online community, your main way to communicate is likely going to be email. Which means you need to be good at it. Take the time to write carefully, fully and precisely. And since text is going to do so much of your communicative heavy lifting, consider being a little more explicit about emotional signal-sending than you might be otherwise. (“I am happy…”, “it is sad…”, “I am grateful.”)

In all your communications, be conscious of your group’s diversity. The group you’re speaking with is likely more diverse than what you’re used to. There may be children and teenagers as well as adults. For many or most, English won’t be their first language. They probably live in multiple countries, have had a broad diversity of experiences, and and hold a wide array of beliefs. This creates a bunch of potential pitfalls. Your jokes may fall flat or offend people. Cultural references (sports, movies, history) may be meaningless. Even for those of us who aren’t American, it’s easy to come across as U.S.-centric. Metaphors, allusions and convoluted sentence structures may not be worth the time they’d take readers to untangle, and make translations much more difficult. High diversity argues for a style that’s literal, straightforward, and well-structured.

Be cautious about creating an insider culture. This is a tough one, because inside jokes and shared history and assumptions foster a sense of belonging. But every in-group requires an out-group, and having a lot of shared lore is unavoidably exclusionary: it makes it harder for new people to join. A strong culture will also inevitably move you towards demographic/attitudinal narrowing, rather than the reverse.

Publish early, publish often. If you are building a plan or a proposal, don’t wait until it’s flawless and polished to publish: release it while it’s still raw and half-baked. For folks from a non-community background this will feel dangerous, like you’re leaving yourself vulnerable to criticism. But in a community context it builds trust and empathy, and will be understood as an invitation to collaborate. Do tons of signposting. Explain what you’re trying to do, and why. Sketch out how you imagine it may work.

Be aware that volunteer time is different from paid time. Staff need to begin their public work as soon as they possibly can (sooner!), and to build in lots of elapsed time for community discussion. Community members have other priorities: school, jobs, families. You can’t expect them to make your work their top priority, so you need to give them the biggest-possible window in which to contribute.

Write (and publish) a greater volume of stuff than you think you should. This feels counter-intuitive for people who’ve been execs in traditional contexts, because in an ordinary executive context the scarcest resource is time, and so we get used to providing summaries, bullet points, upshots, and key takeaways. Succinct=good. In a community context though, comprehensive beats succinct. This is only logical: if you’re writing for a wide diversity of stakeholders, they’re going to want to know a wide variety of stuff. Manually asking and answering questions is slow and laborious and splinters the information so you can’t get it all in one place: it’s faster and better, as much as possible, to anticipate questions and answer them in your original communication.

Assume good faith. This is so easier said than done ;/ But for real, assume good faith. When someone asks a question and you think they are trolling, it’s entirely possible they are not. (Maybe they are 15 years old, or their English is imperfect, or they have an impairment of some kind.) Even if they are trolling: there will always be onlookers who don’t know it, and who, whatever the provocation, will recoil if you are curt or unkind. Trolling also gives you an opportunity to equip onlookers with reasonable arguments that they can go on to use themselves.

Bias towards transparency. Way, way, way more than you think you should. I remember being taught change management back in the early 2000s. Our instructor beat into us that wherever there is a communications vacuum, it will be filled by gossip and fear. That is a million percent true, and even more so in online communities. Gossip and fear grow out of power imbalances and information asymmetry which are to some degree unavoidable in distributed and voluntary groups, and you need to compensate for that. Publishing everything also scales well, because it equips everybody, not just you and your inner circle, to help explain what’s going on.

Note that if you’re the boss it’s insufficient to ask your staff to be transparent, because as long as there is any risk of penalty for over-publishing, they will do the opposite. You need to make it clear that nobody will ever be punished or shamed for being transparent, then you need somebody to publish something they shouldn’t have, and then you need to not punish them. Only then will people begin to take you seriously about transparency.

When you change your mind, say it publicly, and explain why. This is another one that’s tough for execs from a non-community context, where we got trained to express more confidence than we felt. But for real: in a community context, changing your mind and explaining why will not erode your credibility; it will earn you more.

Pay attention to people you disagree with. In an ordinary work environment executives get insulated and protected from honest disagreement. This is bad for them and for their company. Community work is different: there is no shortage of people who will disagree with you, loudly and repeatedly, in public. It’s natural to avoid or ignore those people, but if you do, you’re wasting an opportunity. Consider instead that they may, occasionally, be right.

This post requires a number of caveats and acknowledgements. They’re at the bottom.

In 2008 I was interviewing a candidate for an engineering position at the Wikimedia Foundation and as we talked I found myself imagining what a terrific impression he would make on donors. He’s so shiny and cheerful and mission-oriented, I found myself thinking — donors will love him!

As soon as I thought it, I had the grace to be embarrassed. And although we ended up hiring the guy, we did it because he seemed like a talented engineer, not because he was charming. I was horrified at myself for a while afterwards anyway, and the whole thing ended up being a bit of a turning point for me, as well as a cautionary story I sometimes tell. Because that was the moment that crystalized for me what’s *actually* wrong with nonprofits.

Preface! I’ve always been irritated by people who assume nonprofitland is self-evidently suckier than forprofitland. I’m particularly irritated by people who say that nonprofits “should be more businesslike,” with businesslike as a kind of confused stand-in for “better.” That just seems dumb to me — I feel like it’s obvious that nonprofits function in a specific context including challenges unique to the sector, and that solutions aimed at increasing our effectiveness needed to be designed to respond specifically to those actual, real circumstances. That’s what this post is about: my goal is to describe a serious problem, and point to where I believe we’re beginning to see solutions emerge.

Here it is.

Every nonprofit has two main jobs: you need to do your core work, and you need to make the money to pay for it. In the for-profit sector when you make better products, you make more money — if you make awesome socks, you sell lots of socks. Paying attention to revenue makes sense in part because revenue functions as a signal for the overall effectiveness of the org: if sales drop, that’s a signal your product may be starting to suck, or that something else is wrong.

Nonprofits also prioritize revenue. But for most it doesn’t actually serve as much of an indicator of overall effectiveness. That’s because donors rarely experience the core mission work first-hand — most people who donate to Médecins Sans Frontières, for example, have never lived in a war zone. That means that most, or often all, the actual experiences a donor has with a nonprofit are related to fundraising, which means that over time many nonprofits have learned that the donating process needs –in and of itself– to provide a satisfying experience for the donor. All sorts of energy is therefore dedicated towards making it exactly that: donors get glossy newsletters of thanks, there are gala dinners, they are elaborately consulted on a variety of issues, and so forth.

By contrast, when I buy socks I do not get a gala dinner. In fact it’s the opposite: the more that sockmakers focus relentlessly and obsessively on sock-making awesomeness, the likelier I am to buy their socks in future. This means that inside most of nonprofitland –and unique to nonprofitland– there’s a structural problem of needing to provide positive experiences for donors that is disconnected from the core work of the organization. This has a variety of unintended effects, all of which undermine effectiveness.

It starts with the ED.

EDs prioritize revenue because a fundamental job of any CEO is to ensure their organization has the money it needs to achieve its goals. That means fundraising is necessarily the top priority for a nonprofit ED. That’s why the head of fundraising normally reports to the ED, and it’s why, I’d say from my observation and reading, the average ED probably dedicates about 70% of his or her energy to fundraising.

Optimizing for fundraising distorts how the ED behaves. To the extent EDs optimize themselves for fundraising, they tend to spend time outside their organization — being interviewed, attending conferences, publicly demonstrating wisdom and thought leadership. An ED must hone his or her self-presentation and diplomatic abilities, even at the expense of other attributes such decisiveness or single-mindedness, because that’s what donors see and respond to. There’s an obvious opportunity cost as well: spending 70% of your time on fundraising leaves only 30% for everything else. (That’s why, in a different context, Paul Graham argues that start-ups should have only one person designated to handle fundraising: to preserve the bulk of organizational resources for other stuff.)

The second effect: Optimizing for donor experience promotes a general emphasis on appearances rather than realities. Appearing effective rises in importance relative to being effective.

Here’s how the mature nonprofits I know self-present. Everyone is very polite and the offices are quiet. Their reception areas display racks of carefully-designed marketing materials. One I know has gorgeous brushed stainless steel signs attached to its conference room doors, engraved with an exhortation to be silent in the hallways. Typically the staff dress like academics — the women wear interesting jewelry, with the men in shabby suit jackets and corduroys.

By contrast I noticed in my early days running the WMF, we were quite different. Our staff were young and messy and wore hoodies. They were smart and blunt, sometimes obnoxiously so. The office was often half-deserted because everybody worked all the time, often while travelling or from bed. I’m pretty sure at one point we had a foosball table in the middle of the room, and later there was a karaoke set-up and a Galaga game. What if donors think we’re erratic, undisciplined slobs, I found myself worrying. What if they’ve never met programmers before?

Most nonprofits, it seemed to me, optimized to self-present as competent, sober, and diligent. I think if they optimized to get stuff done, they might look different.

The third effect. Nonprofits are generally conservative in their approach to regulatory compliance, administration, finance and governance practices. (Why? Partly it’s because the core work is complicated: hard to do and hard to measure, so people drift towards stuff that’s simpler. Also, the nonprofit sector is too small to support a diverse array of service providers, and so the services provided by consultants tend to be extremely generic. Boilerplate recommendations on term limits and that kind of thing.) Optimizing for donor experience makes that worse.

Why? It’s easy to describe for donors the core problem a nonprofit is trying to solve, but explaining the work of solving it –and how impact can best be measured– is hard. Far easier to show that the 990 was filed on time, that the org got a clean audit letter, and that the ED’s compensation was determined according to a highly responsible process. And donors seem relatively willing to accept the proposition that administrative effectiveness is a good proxy for overall organizational impact, even though such a proposition is actually pretty weak. A whole industry has developed around this: supporting good compliance and measuring it, as a service for potential donors.

This effect is amplified by the presence of major donors, who are typically wealthy retired business executives.

That’s because major donors like to feel their advice is as useful as their money, and they have decades of experience of people taking their opinions seriously. But they can’t necessarily say much that’s useful about the specifics of helping victims of domestic violence or rehabilitating criminals or protecting endangered gorillas in the Congo. So many nonprofits create opportunities where they can help. They are put on the investment committee, they are asked to help with the audit firm selection process, their advice is sought about when to launch an endowment campaign. This has the effect of focusing the ED’s attention in those areas — because the ED, of course, wants to make sure the major donor’s experience with the org is a positive one. More unintended consequences: “providing a good donor experience” becomes an unstated job requirement for the head of finance. A great head of nonprofit finance needs to not just be a person who’s financially and administratively competent: he or she also needs to be credible, composed, tactful and likable.

So. A major structural flaw of many nonprofits is that their revenue is decoupled from mission work, which pushes them to focus on providing a positive donor experience often at the expense of doing their core work. That’s bad.

What can we do about it?

I believe the problem is to some degree newly now solvable. I know that, because we solved it at the Wikimedia Foundation.

Here’s what we did.

From 2008 until late 2009, the WMF played around with various fundraising models. We applied for and got restricted grants, we cultivated major donors, we made business deals that brought in what’s called in nonprofitland “earned income,” and we fundraised online using what we grew to call the many-small-donors model. After two years we determined we’d be able to be successful using any of those methods, and an important study from Bridgespan had persuaded us to pick one. And so we picked many-small-donors, because we felt like it was the revenue model that best aligned with our core mission work.

Today, the WMF makes about 95% of its money from the many-small-donors model — ordinary people from all over the world, giving an average of $25 each.

It’s awesome.

We don’t give board seats in exchange for cash. Foundations’ priorities don’t override our own. We don’t stage fancy donor parties (well, we do stage one a year, but it’s not very fancy), and people who donated lots of money have no more influence than people who donate small amounts — and, importantly, no more influence than Wikipedia editors. Donors very rarely visit the office, and when they do, they don’t get a special dog-and-pony show. I spend practically zero time fundraising. We at the WMF get to focus on our core work of supporting and developing Wikipedia, and when donors talk with us we want to hear what they say, because they are Wikipedia readers. (That matters. I remember in the early days spending time with major donor prospects who didn’t actually use Wikipedia, and their opinions were, unsurprisingly, not very helpful.)

The many-small-donors models wouldn’t work for everyone, mainly because for it to succeed your core work needs to be a product or service that large numbers of people are aware of, understand, and want to support. About a half-a-billion people read Wikipedia, and we get on average 11 cents a year from each one, which is not much. I know a couple of nonprofits that’ve backed away from the many-small-donors model after doing that math. But I think the usefulness of the many-small-donors model, ultimately, will extend far beyond the small number of nonprofits currently funded by it.

Why? People are slowly getting used to the idea of voluntarily giving smallish amounts of money online to support stuff they like — look at Kickstarter and Donors Choose and Indiegogo. These are not self-interested transactions made after a careful evaluation of ‘what’s in it for me’: they’re people funding stuff because they think it’s great. Meanwhile, the online payment processing market is maturing, with an increasing number of providers supporting an increasing number of currencies and countries, and fees are starting to drop. And, note that donations to the WMF have risen steadily every single year (we’ve been named the nonprofit with the fastest growing revenues in the United States, which probably actually means in the world) — even though the WMF’s fundraising is deliberately restrained. Eleven cents per user per year is nowhere near a ceiling, for Wikipedia or for anyone.

The advent of the internet has given ordinary people access to the means of production, and now they (we) can easily share information with each other on sites like Wikipedia. That’s been playing out for more than a decade, and its effects have included the disintermediation of gatekeepers and middlemen of all types. I think we’re now seeing the same thing happen, more slowly, with the funding of mission-driven work. I think that among other things, we’re going to see the role of foundations and major donors change in surprising ways. And I think the implications of these changes go beyond fundraising itself. For organizations that can cover their costs with the many-small-donors model I believe there’s the potential to heal the disconnect between fundraising and core mission work, in a way that supports nonprofits being, overall, much more effective.

Notes: This post is written from the vantage point of somebody who thinks many nonprofits do good work in difficult circumstances: please read it from that perspective. Lots of people think nonprofits are lazy and inefficient and woolly-minded. That’s sometimes true, but no more so in my experience than at for-profit orgs. The world has no shortage of suck.

I also want to thank some of the people who’ve influenced my thoughts in this area. Although the views expressed here are my own, Erik Moeller and I have talked a ton about this stuff over the past half-dozen years. He was the first person to point out to me the absurdity of overheard ratios, and has written about them extensively and publicly, starting back in 2009. Afterwards, he and I discovered the good work of Dan Pallotta and also the Urban Institute, investigating overhead ratios and explaining why they’re bunk. I’ve also benefited from reading Jim Collins’s monograph Good to Great and the Social Sectors, as well as two books from Michael Edwards: Just Another Emperor? The Myths and Realities of Philanthrocapitalism, and Small Change: Why Business Won’t Save the World. I was helped by a conversation about difficulties facing new nonprofits a few years back at the Aspen Institute, as well as by dozens of less structured conversations with fundraisers including particularly Zack Exley, as well as with my fellow EDs, including ones on whose boards I serve. David Schoonover has done some analysis of U.S. non-profit funding models that has influenced me, and he and I have talked extensively about challenges facing the nonprofit sector, including this one. The folks at Omidyar have also been helpful, including pointing me towards the very useful Bridgespan study linked above.

Because I’ve been working lately on issues related to grantmaking and Wikimedia movement entities, it might be tempting to assume my arguments here are somehow aimed at informing or influencing those conversations. They’re not. To the extent anything here is useful to those conversations that’s great, but that’s not why I wrote this.

Wikipedia Anti-SOPA Blackout Design

Below is the text of a talk I delivered Monday at the 2013 MIT-Knight Civic Media Conference in Boston. Republican Congressman Darrell Issa, Icelandic member of Parliament Birgitta Jónsdóttir and I spoke on the theme of “Insiders/Outsiders: what is the right approach to change.”

Unlike many of the people in this room, I’m not an academic or a public policy expert and so I won’t be bringing you statistics or analysis or theories today. I run a big website. I’m also a journalist. If we consider ourselves to be in a war for the free and open internet, I am here to tell you some stories from the trenches.

Wikipedia is pretty much the consummate insider-outsider: the #5 most-popular site in the entire world, read by a half a billion people every month, yet written by utterly ordinary people with no special power or authority at all. If they have credentials, they park them at the door.

Wikipedia is a tremendous success story. It launched in 2001 and took off very quickly: by 2006 it had surpassed all the other news and information sites in terms of popularity. Today it’s a behemoth. And people love to point to it as an example of everything great about the internet. There’s only one problem with that. Wikipedia is pretty much alone. It’s NOT the general rule: it’s the exception that proves the rule.

Wikipedia is operated by a non-profit. In Silicon Valley, people often find that astonishing – they ask me why Jimmy Wales “left so much money on the table,” and whether he regrets doing it. (Answer: no.) Sometimes people ask me why we don’t just put ads on the site, and whether we are against advertising, against for-profit companies, against capitalism.

We say no. Our view is that the internet should have balance, much like the offline world. A city has restaurants and shops and banks as well as schools and libraries and parks. Wikipedia is like a park. It’s a public space, accessible and used by everybody.

But where are the other parks?

  • Wikipedia is the only donor-supported site in the top 50
  • Wikipedia and Mozilla are the only two nonprofits in the top 25(*)
  • The average person spends practically all their time online on the sites of for-profit companies, the vast majority of them American. (Caveat: mainland China.)

This worries me. The internet is evolving into a private-sector space that is primarily accountable to corporate shareholders rather than citizens. It’s constantly trying to sell you stuff. It does whatever it wants with your personal information. And as it begins to be regulated or to regulate itself, it often happens in a clumsy and harmful way, hurting the internet’s ability to function for the benefit of the public. That for example was the story of SOPA.

My first war story happened soon after I joined the Wikimedia Foundation. It’s about censorship in the United Kingdom.

The internet industry is, of course, generally hoping to remain unregulated. In the UK a coalition of ISPs have formed an association called the Internet Watch Foundation, which is essentially a group of retired police officers, paid by the ISPs to investigate complaints of child pornography online. In 2008, that group got a complaint about an image on Wikipedia of an album cover from 1976(**) – an album called “Virgin Killer”, by a German heavy metal band called the Scorpions. The album cover image is a young girl, nude, which has been treated with an effect that makes it look like she’s looking at you through a pane of glass that has been shattered by a bullet. It’s deliberately provocative – it’s heavy metal.

The Internet Watch Foundation decided this was child porn, and attempted to block it from the view of UK internet users. In doing that, they accidentally made it impossible for anybody to edit Wikipedia from inside the UK.

People went nuts. There was a lot of press coverage, both inside the UK and internationally. The Wikimedia Foundation spoke to the press, and individual Wikipedia editors in the UK spoke to the press and blogged and tweeted and so on. And after a few days the IWF reversed its decision.

Two interesting things:

  1. When they reversed their decision, they explicitly said that they still believed the image was child porn, but that the public outcry was too much for them. They backed down because they couldn’t win a PR war against fans of the number five website in the world. If we had been Joe’s Album Art History Wiki, it’s clear the decision would not have been reversed.
  2. Importantly and invisibly, while this story was playing out, and was being written about by journalists internationally, at the Wikimedia Foundation we noticed Amazon had quietly pulled the Virgin Killer album from its site. It still sold a version of the album that had a different cover, but it no longer sold the version with the image that was being challenged. Amazon didn’t call us to ask what was going on, or to offer us help. They didn’t even silently watch and wait. They pulled the album off their shelves — not just in the UK but worldwide.

Don’t get me wrong, I love Amazon. I spend a significant portion of my disposable income at Amazon every year, and I am grateful that it’s made my life easier and given me choices I didn’t have before it existed. Amazon is fantastic. But it’s also true that Amazon’s job is not to protect the public interest – it’s to advance the interests of Amazon.

Another story.

In 2011, there was a Wikipedia conference in Mumbai at which Jimmy Wales spoke as well as our Board member Bishakha Datta, and a few of our staff. To our considerable surprise, a popular Indian political party picketed outside our conference and demanded that the police arrest us. They were doing that because the map of India displayed on Wikipedia shows the country’s borders as per the United Nations – with the borders with China and Pakistan “disputed” – and not as per the map defined by the government of India. It is only legal, in India, to publish a map showing India’s borders as they are defined and understood by the government of India.

As you can imagine, the protest made us exceedingly anxious. We knew that although India is a democracy with a commitment to free speech, that commitment is variable and laws and community standards inside India are somewhat volatile. And so we retained a bunch of lawyers. We spent weeks researching the legal and PR issues. Where we could, we took a variety of small non-controversial steps to protect ourselves. And ultimately we got lucky, and the issue seemed to fizzle out.

What we did not do was change the map of India displayed on Wikipedia. Partly because we can’t – that’s a Wikipedia community decision – but also because we shouldn’t. It’s perfectly reasonable to publish a map of India with the UN borders.

What was interesting here, as we researched our position, was what everybody else does. It seems that inside India, every major player except Wikipedia displays the map of India with the borders as defined by the Indian government. If you’re in India, that’s what Google shows you. When the Economist magazine prints a map of India, I was told by our lawyer, the version of the magazine they sell inside India shows a map different from the version in the magazine they sell elsewhere.(***)

It’s also worth noting that the Wikimedia Foundation has a legal team and a PR team, and Wikipedia is a popular site, much-loved by its readers. Not everyone has those resources. Of those that do, most are private and for-profit. Again, some of those players are doing great things. But on the whole, over time, they will put profits before public service. That’s their job and their obligation.

Governments, in my experience, aren’t helping. Mostly they’re just befuddled, but even if they knew what to do, there’s no reason to believe they’d do it. Too often they’re corporate captives. We saw it with SOPA. Today they listen too much to the entertainment industry – the copyright owners. Tomorrow, maybe they’ll be listening too much to giant technology companies. Either way, the voices of ordinary people will only rarely be heard, and I have difficulty believing that more or better civic engagement will fix that anytime soon. I agree with Larry Lessig: structural problems – fundraising, gerrymandering – have made for a powerful incumbency with skewed incentives.

And so, as a soldier in the trenches, my message to this conference is caution and concern.

Aside from Wikipedia, there is no large, popular space being carved out for the public good. There are a billion tiny experiments, some of them great. But we should be honest: we are not gaining ground. Our schools, our libraries, our parks – they are very, very small and they may or may not sustain. We certainly have no information-sharing participatory Garden of Eden, the promise of the internet that we all originally believed in. Though we are not lost, we are losing.

I say this because it’s easy to come together for a conference like this and get excited about awesome experiments and interesting breakthroughs. It’s worth doing! We want to celebrate success! But if you’ve read Tim Wu‘s Master Switch, if you’re reading Robert McChesney‘s Digital Disconnect, you know that the insiders are winning. We are not.

The internet needs serious help if it is to remain free and open, a powerful contributor to the public good. That’s what I’m hoping you’ll discuss over the course of this conference. How to create an ecosystem of parks and libraries and schools online … that supports participation, dialogue, sharing.

Thank you.

(*Turns out I was wrong about this. Mozilla is #60 globally according to comScore Media Metrix, the industry standard for web audience measurement. Therefore, I should actually have said Wikipedia, at #5, is the *only* non-profit in the top 25.)

(**When I delivered the talk I said 2009 and 1979. I’d been misremembering: it was 2008 and 1976.)

(***Since delivering this talk, Tilman Bayer at the Wikimedia Foundation pointed me towards this BBC article, in which the Economist accuses the Indian government of hostile censorship after it forced the magazine to place a blank white sticker over a map of Kashmir in the 30,000 copies of the May 2011 Economist that were distributed in India.)

On 5 July 2013 I updated the blog post image to be the design that actually appeared on Wikipedia during the anti-SOPA blackout, as recommended to me by Brandon Harris, the guy who designed it :-)

For the past sixteen months, the Wikimedia Foundation has been having uncomfortable conversations about how we handle controversial imagery in our projects — including, a few weeks ago, the staging of a referendum on an image hiding feature requested by our Board. The purpose of this post is not to talk specifically about the referendum results or the image hiding feature: for that, I’ll be talking in more official venues. The purpose of this post is to step back and assess where we’re at, and to call for a change in tone and emphasis in our discussions.

Please note also that due to the nature of the topic, you may find yourself offended by this post, and/or the materials linked from it.

In March 2010, editors on the German Wikipedia ran a poll asking their colleagues whether they would support a rule restricting the types of material that could appear on the German home page. Thirteen voted in favour of restrictions, and 233 voted against. A few weeks later, the German Wikipedia featured the article about the vulva on its home page, which included a close-up photograph of an open vagina. Twenty-three minutes after the article went up, a reader in Berlin wrote “you can’t be serious?!,” and called for the image to be taken down. This initiated an on-wiki discussion that eventually reached 73,000 words – the length of a shortish novel. It included a straw poll in which 29 people voted to remove the image and 30 voted to keep it. The image was kept, and the article remained on the front page for its full 24 hours.

A few months later, in June, the Wikimedia Foundation Board of Trustees began to discuss how the Wikimedia community was handling controversial imagery. Why? Because some people seemed to be using Commons to stockpile commercial porn; because the German community had put a close-up photo of a vagina on its homepage; and because upskirt photos and controversial editorial cartoons seemed to be being categorized in ways that seemed designed to be provocative, and the people who complained about them were being shot down.

The Wikimedia Foundation was concerned that a kind of market failure might be happening — that the Wikimedia community, which is generally so successful at achieving good decision quality through a consensus process, was for some reason failing to handle the issue of controversial material well. It set out to explore what was going on, and whether we needed to handle controversial imagery differently.

That triggered community members’ fears of censorship and editorial interference. And so we find ourselves today, sixteen months later, locked in angry debate. At a meeting in Nuremberg a few weeks ago, German Wikipedian User:Carbidfischer furiously denounced our Board Chair Ting Chen. The other day –as far as I know for the first time ever– somebody called someone else an asshole on one of our mailing lists. User:Niabot created this parody image. It’s unpleasant and unconstructive, and if you’re familiar with transactional analysis, or with the work done by the Arbinger Institute, you’ll recognize the bad patterns here.

The purpose of this post is to figure out why we aren’t handling this problem well, and how we can get back on track.

So: backing up.

Is there a problem with how the Wikimedia projects handle potentially-objectionable material? I say yes. The problems that led the Board to want to address this issue still exist: they have not been solved.

So what’s the solution? I have read pages upon pages of community discussion about the issue, and I sympathize and agree with much of what’s been said. Wikipedia is not, and should never be, censored. It should not be editorially interfered with.

But refusing censorship doesn’t mean we have no standards. Editors make editorial judgments every day, when we assess notability of topics, reliability of sources, and so forth. The German Wikipedia particularly is known to have extremely rigorous standards.

So why do we refrain from the expression of editorial judgment on this one issue?

I think there are two major reasons.

First, we have a fairly narrow range of views represented in our discussions.

We know that our core community represents just a sliver of society: mainly well-educated young men in wealthy countries, clustered in Europe and North America. It shouldn’t surprise us, therefore, when we skew liberal/libertarian/permissive, especially on issues related to sexuality and religion. Our demographic and attitudinal narrowness is a shame because at the heart of the projects is the belief that many eyes make all bugs shallow and yet, we’re not practicing what we preach. Instead, we’ve become an echo chamber: we hear only voices like our own, expressing points of view we already agree with. People who believe other things fall silent or abandon the conversation or are reduced to impotent rage. Or, and even likelier, they never made it to the table in the first place.

Second, we are confusing editorial judgment with censorship.

Censorship is imposed from outside. Editorial judgment is something we do every day in the projects. Applying editorial judgment to potentially-objectionable material is something that honourable journalists and educators do every day: it is not the same as censorship, nor does it constitute self-censorship.

In newsrooms, editors don’t vote on whether they personally are offended by material they know their readers will find objectionable, and they don’t make decisions based on whether the angry letters outnumber the supportive ones. They exercise empathy, and at their best they are taking a kind of ‘balance of harm’ approach — aiming to maximize benefit and minimize cost. The job is to provide useful information to as many people as possible, and they know that if people flee in disgust, they won’t benefit from anything the newsroom is offering. That doesn’t mean newsrooms publish only material that’s comfortable for their readers: it means they aim to exercise good judgment, and discomfit readers only when –on balance– discomfort is warranted.

How does that apply to us? It’s true that when people go to the article about the penis, they probably expect to see an image of a penis, just like they do when they look it up in a reference book in their public library. It’s also true that they probably wouldn’t benefit much from a gallery of cellphone camera shots of penises, and that’s why we don’t have those galleries on our articles. In lots of areas, we are currently doing a good job.

But not always.

When an editor asks if the image cleavage_(breasts).jpg really belongs in the article about clothing necklines, she shouldn’t get shouted down about prudishness: we should try to find better images that don’t overly sexualize a non-sexual topic. When an editor writes “you can’t be serious?!” after vagina,anus,perineum_(detail).jpg is posted on the front page, the response shouldn’t be WP:NOTCENSORED: we should have a discussion about who visits the homepage, and we should try to understand, and be sensitive to, their expectations and circumstances and needs. When we get thousands of angry e-mails about our decision to republish the Jyllands-Posten Muhammad cartoons, we should acknowledge the offence the cartoons cause, and explain why, on balance, we think they warrant publication anyway. None of that is censorship. It’s just good judgment. It demonstrates transparency, a willingness to be accountable, and a desire to help and serve our readers — and it would earn us trust.

I believe that in our discussions to date, we’ve gotten ourselves derailed by the censorship issue. I know that some people believe that the Wikimedia Foundation is intending to coercively intervene into the projects, in effect overruling the judgment of the editorial community. I don’t see it that way, I regret that others do, and I dislike the ‘authoritarian parent / rebellious adolescent’ dynamic we seem to be having trouble resisting.

Wikipedia is not censored. It should never be censored. That doesn’t relieve us of the obligation to be thoughtful and responsible.

So: what needs to happen?

We need to have a discussion about how to responsibly handle objectionable imagery. That discussion doesn’t need to happen with the Wikimedia Foundation (or at least, not solely with the Wikimedia Foundation). The projects should be talking internally about how to avoid unnecessarily surprising and offending readers, without compromising any of our core values.

Those community members who are acting like provocateurs and agitators need to stop. Demonizing and stereotyping people we disagree with pushes everyone into extremist positions and makes a good outcome much less likely. We need to look for common ground and talk calmly and thoughtfully with each other, staying rooted in our shared purpose. Some editors have been doing that throughout our discussions: I am seriously grateful to those people, and I wish others would follow their example.

“Wikipedia is not censored” is true. And, we need to stop using it as a conversation killer. It’s the beginning of the conversation, not the end of it.

We need to set aside anxieties about who’s in charge, and quit fighting with each other. We need to be aware of who’s not at the table. We need to bring in new voices and new perspectives that are currently lacking, and really listen to them. Those community members who’ve been afraid to talk need to speak up, and those who’ve been driven away need to come back.

The purpose of this post is to call for that responsible engagement.

Like I said at the top of this post, my purpose in writing this is not to talk about the referendum results or the image hiding feature: for that, I’ll be talking in more official venues.

I haven’t posted here for a while because I was on a long and slightly gruelling trip, timed to coincide with Wikimedia’s 2010 fundraiser. (Which has just successfully concluded with more than a half million people contributing a total of USD 16 million, exceeding our goal in our shortest annual campaign to date: you can read more about it on the Wikimedia Foundation blog. My thanks to everyone who donated / helped with the campaign, and congratulations to Zack and the fundraising team!)

The gist of the trip: Between November 20 and December 6, I was in Stockholm, London, Dubai and Delhi. I spoke at three conferences, was interviewed by 16 journalists, and met about 50 Wikimedians as well as a few supporters and friends at cultural institutions, and I interviewed candidates for the Wikimedia Program Director for India. Afterwards I went on three weeks of holiday in India, including a 10-day silent Vipassana meditation retreat in Karnal, north of Delhi. The trip had multiple overlapping goals — to advance awareness of Wikipedia (particularly during our fundraising season) and encourage people to try editing, as well as to support chapters and find out a little more about the challenges and opportunities faced by editors in the Middle East. On the whole, the trip was successful and we’ll probably aim to do similar ones in future, building on lessons learned this time. I’ll blog more about it here as I get my notes organized.

Now, I’m back in San Francisco, and catching up on the 650-ish e-mails that accumulated (and didn’t get responded to) while I was away. Happy 2011 :-)

The Wikimedia Foundation Board of Trustees met in San Francisco a few weeks ago, and had a long and serious discussion about controversial content in the Wikimedia projects. (Why? Because we’re the only major site that doesn’t treat controversial material –e.g., sexually-explicit imagery, violent imagery, culturally offensive imagery– differently from everything else. The Board wanted –in effect– to probe into whether that was helping or hurting our effectiveness at fulfilling our mission.)

Out of that agenda item, we found ourselves talking about what it looks like when change is handled well at Wikimedia, what good leadership looks like in our context, and what patterns we can see in work that’s been done to date.

I found that fascinating, so I’ve done some further thinking since the meeting. The purpose of this post is to document some good patterns of leadership and change-making that I’ve observed at Wikimedia.

Couple of quick caveats: For this post, I’ve picked three little case studies of successful change at Wikimedia. I’m defining successful change here as ‘change that stuck’ – not as ‘change that led to a desirable outcome.’ (I think all these three outcomes were good, but that’s moot for the purposes of this. What I’m aiming to do here is extract patterns of effective process.) Please note also that I picked these examples quickly without a criteria set – my goal was just to pick a few examples I’m familiar with, and could therefore easily analyze. It’s the patterns that matter, not so much the examples.

That said: here are three case studies of successful change at Wikimedia.

  • The Board’s statement on biographies of living people. Policies regarding biographies had been a topic of concern among experienced Wikipedians for years, mainly because there is real potential for people to be damaged when the Wikipedia article about them is biased, vandalized or inaccurate, and because our experience shows us that articles about non-famous people are particularly vulnerable to skew or error, because they aren’t read and edited by enough people. And, that potential for damage –particularly to the non-famous– grows along with Wikipedia’s popularity. In April 2009, the Board of Trustees held a discussion about BLPs, and then issued a statement which essentially reflected best practices that had been developed by the Wikipedia community, and recommended their consistent adoption.  The Board statement was taken seriously: it’s been translated into 18 languages, discussed internally throughout the editing community, and has been cited and used as policies and practices evolve.

  • The strategy project of 2009-10. Almost 10 years after Wikipedia was founded, the Board and I felt like it was time to stop down and assess: what are we doing well, and where do we want to focus our efforts going forward. So in spring 2009, the Wikimedia Board of Trustees asked me to launch a collaborative, transparent, participatory strategy development project, designed to create a five-year plan for the Wikimedia movement. Over the next year, more than 1,000 people participated in the project, in more than 50 languages. The resultant plan is housed on the strategy wiki here, and a summary version will be published this winter. You can never really tell the quality of strategy until it’s implemented (and sometimes not even then), but the project itself has accomplished what it set out to do.

  • The license migration of May 2009. When I joined Wikimedia this process was already underway, so I only observed first-hand the last half of it. But it was lovely to watch. Essentially: some very smart and experienced people in leadership positions at Wikimedia decided it made sense to switch from the GFDL to CC-BY-SA. But, they didn’t themselves have the moral or legal right to make the switch – it needed to be made by the writers of the Wikimedia projects, who had originally released their work under the GFDL. So, the people who wanted the switch launched a long campaign to 1) negotiate a license migration process that Richard Stallman (creator of the GFDL and a hero of the free software movement) would be able to support, and 2) explain to the Wikimedia community why they thought the license migration made sense. Then, the Wikimedia board endorsed the migration, and held a referendum. It passed with very little opposition, and the switch was made.

Here are nine patterns I think we can extract from those examples:

  1. The person/people leading the change didn’t wait for it to happen naturally – they stepped up and took responsibility for making it happen. The strategy project grew out of a conversation between then-board Chair Michael Snow and me, because we felt that Wikimedia needed a coherent plan. The BLP statement was started by me and the Board, because we were worried that as Wikipedia grew more popular, consistent policy in this area was essential. The license migration was started by Jimmy Wales, Erik Moeller and others because they wanted it to be much easier for people to reuse Wikimedia content. In all these instances, someone identified a change they thought should be made, and designed and executed a process aimed at creating that change.
  2. A single person didn’t make the change themselves. A group of people worked together to make it happen. More than a thousand people worked on the strategy project. Probably hundreds have contributed (over several years) to tightening up BLP policies and practices. I’m guessing dozens of people contributed to the license migration. The lesson here is that in our context, lasting change can’t be produced by a single person.
  3. Early in the process, somebody put serious energy towards achieving a global/meta understanding of the issue, from many different perspectives. It might be worth pointing out that this is not something we normally do: in order to do amazing work, Random Editor X doesn’t have any need to understand the global whole; he or she can work quietly, excellently, pretty much alone. But in order to make change that involves multiple constituencies, the person doing it needs to understand the perspectives of everyone implicated by that change.
  4. The process was carefully designed to ask the right people the right questions at the right time. The license migration was an exemplar here: The people designing the process quite rightly understood that there was no point in asking editors’ opinions about something many of them probably didn’t understand. On the other hand, the change couldn’t be made without the approval of editors. So, an education campaign was designed that gave editors access to information about the proposed migration from multiple sources and perspectives, prior to the vote.
  5. A person or a group of people dedicated lots of hours towards figuring out what should happen, and making it happen. In each case here, lots of people did lots of real work: researching, synthesizing, analyzing, facilitating, imagining, anticipating, planning, communicating.
  6. The work was done mostly in public and was made as visible as possible, in an attempt to bolster trust and understanding among non-participants. This is fundamental. We knew for example that the strategy project couldn’t succeed if it happened behind closed doors. Again and again throughout the process, Eugene Eric Kim resisted people’s attempts to move the work to private spaces, because he knew it was critical for acceptance that the work be observable.
  7. Some discussion happened in private, inside a small group of people who trust each other and can work easily together. That’s uncomfortable to say, because transparency and openness are core values for us and anything that contradicts them feels wrong. But it’s true: people need safe spaces to kick around notions and test their own assumptions. I know for example that at the beginning of the Board’s BLP conversations, I had all kinds of ideas about ‘the problem of BLPs’ that turned out to be flat-out wrong. I needed to feel free to air my bad ideas, and get them poked at and refuted by people I could trust, before I could start to make any progress thinking about the issue. Similarly, the Board exchanged more than 300 e-mails about controversial content inside its private mailing list, before it felt comfortable enough to frame the issue up in a resolution that would be published. That private kicking around needs to happen so that people can test and accelerate and evolve their own thinking.
  8. People put their own credibility on the line, endorsing the change and trying to persuade others to believe in it. In a decentralized movement, there’s a strong gravitational pull towards the status quo, and whenever anyone tries to make change, they’re in effect saying to hundreds or thousands of people “Hey! Look over here! Something needs to happen, and I know what it is.” That’s a risky thing to do, because they might be perceived in a bunch of negative ways – as naiive or overreacting, as wrong or stupid or presumptuous, or even as insincere – pretending to want to help, but really motivated by inappropriate personal self-interest. Putting yourself on the line for something you believe in, in the face of suspicion or apathy, is brave. And it’s critical.
  9. Most people involved –either as participants or observers– wanted more than anything else to advance the Wikimedia mission, and they trusted that the others involved wanted the same thing. This is critical too. I have sometimes despaired at the strength of our default to the status quo: it is very, very hard to get things done in our context. But I am always reassured by the intelligence of Wikimedia community members, and by their dedication to our shared mission. I believe that if everyone’s aligned in wanting to achieve the mission, that’s our essential foundation for making good decisions.

Like I said earlier — these are just examples I’ve seen or been involved in personally. I’d be very interested to hear other examples of successful change at Wikimedia, plus observations & thinking about patterns we can extract from them.

I spent this past weekend with Wikimedia trustee Phoebe Ayers at the Quaker Center in Ben Lomond, California, attending a workshop called Business Among Friends: Clerking as a Spiritual Discipline.

Neither Phoebe nor I are Quakers, but we’re curious about them. I first read about the similarities between Quaker and Wikimedian decision-making practices in Joseph Reagle‘s excellent new analysis of the Wikipedia community Good Faith Collaboration – and since then, I’ve read a dozen or so Quaker books and pamphlets. I’ve been especially interested in the practice of “clerking.”

The job of the Quaker clerk is to facilitate Quaker meetings – to create the agenda, set the tone, traffic-cop the discussion, listen, help resolve conflict, and understand and document agreement. It’s a role that reminds me a lot of leadership (both formal & informal) at the Wikimedia Foundation, and so I’ve been curious to learn more about it. The purpose of this post is to share some impressions and identify a few Quaker practices that I think Wikimedia could usefully adopt.

Disclaimer! Experienced Quakers will probably find that my grasp of some practices is shaky, and I may have mischaracterized things. People who attended the workshop may remember things differently from me. And, I am going to use words like “consensus” in the non-Quaker sense, so that they’ll make sense for readers who aren’t familiar with Quakerism. My apologies for errors and misunderstandings.

Based on what I’d read, I expected the Quakers to be mostly middle-aged or older, mostly white, and really, really friendly. They were exactly that.

But I was surprised to discover also some unexpected commonalities with Wikimedians. Both speak in acronyms (WP:NPOV, meet M&O, FCL and FAP). Both are really proud of their work, and yet tend towards self-criticism rather than self-promotion. Both talk a lot, and are precise and articulate in the way they use language (the Quakers I met spoke in complex sentences, studded with caveats and parentheticals). Both resist speaking on behalf of their group. And both have a strong individualistic streak, and describe themselves as skeptical about leadership and authority.

(To that last point: On Saturday night, Quaker adults and teenagers played a game called Big Wind Blows, which is kind of like musical chairs. Everyone’s in a circle and the person in the centre, who doesn’t have a seat, calls out “Big wind blows for everybody who [has X characteristic].” Everybody without the characteristic stays in their chair; everybody with it runs around looking for a new seat. On Saturday night, the first person in the centre said something like “Big wind blows for everyone with brown hair.” Second was “everyone who’s wearing blue jeans.” Third was “everyone who’s gone to jail for a matter of conscience.” Four Quakers in the group had chosen jail rather than, say, serving in the military or paying taxes. And doing that was considered ‘normal’ enough to be fodder for a game.)

Tomorrow, once I’ve cleaned them up, I’ll post some detailed notes I took. For now though, I’ll elaborate on a few Quaker practices that I think we Wikimedians could learn from. Most of this will be applicable for face-to-face meetings (i.e., our board meetings, Wikimania, meet-ups), but there may be relevance here for on-wiki work too.

Everybody who’s part of the movement shares responsibility for helping it succeed. Nobody gets to sit on the sidelines and watch things fail.

The Quakers talk a lot about “clerking yourself,” which basically means taking personal responsibility for the group’s collective success. People are expected to behave in a disciplined fashion, including managing themselves emotionally. They’re expected to be open-minded, open to learning and changing their minds. They’re expected to pay attention and listen carefully to each other. They’re expected to avoid the temptation to get mad or show off, and to instead speak “with love rather than judgment.” They’re expected to restrain themselves from talking too much, from interrupting other people, and from repeating the same arguments again and again. Quakers are expected to be willing and able to calmly, thoughtfully, explore areas of disagreement. If they’re feeling shy or reticent or silenced, they’re expected to say that, so that other people can find ways to support them and ensure they’re heard. And if other people are behaving badly, everyone is expected to try to help them behave better.

All this, obviously, is aspirational. As someone at the workshop said, Quakers aren’t paragons, and they’re just as likely as anyone else to be childish and whiny and egotistical. But they’re expected to try really hard not to be.

Setting the right tone is critical for success.

All weekend, I was struck by the Quakers’ skill at establishing and maintaining a rich, healthy emotional tone.

The most obvious example of this is the Quakers’ use of silence. Quakers really value silence: it’s built into all of their religious meetings and their discussions, and during the weekend, we probably spent a combined total of two or three hours together in silence — sometimes for long stretches, and sometimes just for a few minutes. That does something really interesting: it makes everybody more judicious. You have time to reflect, to organize your thoughts, to calm down. You get to listen to other people, rather than using their speaking time to plan what you’ll say next. What you say is smarter and more thoughtful than it would’ve been otherwise.

That’s just one technique the Quakers use: there were lots of others. Elizabeth and Eric, who facilitated the workshop, modeled warmth and patience and respect. They thanked people, a lot. They acknowledged and welcomed the new people. They opened the meetings in a circle, with everyone holding hands.

It reminded me of something Sal Giambanco of Omidyar Network once told me – that he recommends non-profit boards kick off their meetings with a recitation of their mission statement. It’s the same kind of thing – rituals and practices designed to remind us that what we’re doing together is meaningful, so that we can approach it in a spirit of love and respect.

Sometimes you have to kick out difficult people. Maybe.

The people attending the workshop were all experienced Quakers. And it was clear from the stories they told and the questions they asked, that Quaker meetings suffer from difficult people.

This reminded me of Wikimedia. Because it didn’t seem like difficult people were necessarily over-represented inside Quakerism. Rather, it seemed like a normal number of difficult people created stress and anxiety disproportionate to their actual numbers. Elizabeth says that many clerks have shared with her stories about a single problematic member in their meeting, who wants attention or influence and takes advantage of the consensus process to grandstand and delay or block action for months or even years. Quakers call these people ‘dissenting spirits’ or ‘chronic objectors,’ and characterize them as “needing to hold themselves out of alignment with the group.” Elizabeth describes them as people who, no matter how much trust is extended to them, are unable to develop trust in others. Their disruptive presence can drive away others, and sometimes even threaten the survival of the group.

Which sounded sadly familiar to me.

Here’s what I think happens. Where other groups might unhesitatingly excommunicate a person who repeatedly broke their rules, it seems to me that the Wikimedia projects and the Quakers both tend to agonize instead, presumably because both groups pride themselves on being highly inclusive and tolerant. (Remember I said the Quakers are strongly individualistic? I suspect that, like some Wikimedians, some Quakers have a history of getting kicked out of various groups, and so they have a lot of empathy for people having that kind of difficulty.)

But even the Quakers, it seems, have their limits. As Elizabeth wrote in her book on clerking, “A healthy meeting will provide spiritual nurture for the ‘difficult’ Friend, but will understand that protecting the safety of the meeting has priority. It will not confuse ‘being loving’ or ‘being Quakerly’ with tolerating the destructive behavior of an individual, but will understand that setting firm limits is loving.”

This was probably the most uncomfortable topic that got addressed during the workshop, and it was the only time I remember when Elizabeth and Eric seemed to disagree. It’s a tough topic, both for the Quakers and Wikimedians.

I want to thank Jacob Stone and Gretta Stone, directors of the Quaker Center in Ben Lomond, as well as Elizabeth Boardman (Davis Meeting) and Eric Moon (Berkeley Meeting), facilitators of the workshop. Everybody at the workshop was enormously welcoming to Phoebe and me: we are really grateful. Seriously: it was lovely.

I’ll publish more notes –rougher, longer– probably tomorrow.