Site   Web

August 17, 2007

Interview with Wikipedia Administrator Durova

“I’m so glad I don’t manage Wiki, it’s like having squabbling kids high on sugar in the back of a mini-van for 10 hours!” EvilGreenMonkey – Forums Editor, SearchEngineWatch

I have just spent two and a half hours on Google Chat with someone who does. Controversial, quick witted and damned by those she has banned from Wikipedia, Durova is a much maligned but multi-faceted character. She is also one of the most recognized names surrounding Wikipedia. As one of the more outspoken of 1302 or so Administrators, Durova has attracted a lot of heat for her ardent defense of the Encyclopedic nature of Wikipedia. readers will recall Durova’s name from Ross Dunn’s article, “Is Wikipedia Corrupt?” which ran on Thursday August 9. In the piece, Durova’s name came up several times in relation to a dispute between herself and Gregory Kohs, an SEW Forums member who goes by the name of “thekohser”.  It was this dispute that caught the eye of several writers in the search marketing sector. Durova contacted me hoping that more open communication would help clear the air between the Administrators at Wikipedia and the search community.

The Chat-interview (transcript here) lasted far longer than either of us expected but, after speaking with Durova for such a long period of time, the conversation left me with a deeper respect for the work of the Wikipedians along with a deepened cynicism when reading the rantings of some of its detractors.

The interview covered over 13 pages (Arial, 10pt font) in a Word document. I’ve decided to reprint excerpts from it in this article along with a link to the full transcript.

To start the interview, I wanted to get some background on how the Wikipedia operates. I wanted a sense of structure, an understanding of how its hierarchy works to make and enforce decisions.

me: ok.  Is there a hierarchy of editors/volunteers at Wikipedia?

durova: Mostly it’s a meritocracy.

me: ok… a hierarchy based on merit

durova: In theory (I could show you a quote) an administrator’s opinion in a discussion carries no more weight than any other editor’s.

But certain tools have to be restricted for practical reasons.

me: k.

durova: There are tools I don’t have.

Checkuser, for instance, gives information about an account’s underlying IP address.  That’s very closely controlled for privacy reasons.

Likewise there’s a power called Oversight.  That’s done to erase personal disclosures.  Such as a twelve-year-old who posts her home address.

me: what are the various titles (ie: admin, editor, etc…) and how are they earned/granted

Sent at 11:10 AM on Wednesday

durova: Well, anyone’s an editor who edits the sight.  You could click the “edit” tab on an article and become an editor right now, if you haven’t already.

Here’s a page on administratorship.

Administrators and bureaucrats are elected by community consensus.

Arbitrators are a combination of election and appointment.  The other positions are appointed either by Jimbo Wales or the arbitration committee.

me: TY:  my goal here is to help users understand the environment.

durova: Sure.

me: how does one get noticed by the community / stand for election?

how long are the terms?

durova: Arbitration is a three year term.  The others have no specific duration.

me: as an estimate, how long does the average editor/admin/arbitrator serve?

durova: Most arbitrators resign before the full three years.  It’s a very demanding position.

me: no doubt

durova: For the rest, people are as active as they want to be.  Wikipedia had 1302 administrators when I checked this morning.  About 2/3 of those are active.

Sent at 11:15 AM on Wednesday

me: so you are one of about 860?

durova: There’s a voluntary program called “Administrators open to recall.”  These are sysops who’ve pledged to go through a reconfirmation vote if enough other editors ask for it.  About 100 administrators participate in that program.  I’m one of them.

I could check that number if you like for accuracy, but about that.  Yes.

Here’s a link to the current votes.

me: ok. this is like a democratic method of deposing admins for the editors?

durova: There’s been talk of that.  Nothing formal has come out of it.  The arbitration committee handles instances of misconduct.  Occasionally they desysop someone.

me: I’d like to get back to that in a moment or two.

durova: Sure.

me: why would a candidate’s nomination be voted down or fail to reach consensus?

how do the voters (editors) know about each nominee?

durova: Well, there could be concerns about an editor’s overall experience or about their conduct.

me: is there a lot of thought put into voting by the editors?  Is there debate?

durova: These are discussions rather than raw votes.  Reasoning counts for a lot. If you’d like to take a moment and look at the current discussions you’ll see a cross section of how this plays out.

Next, I wanted to get a sense of how much work was involved for any given editor or administrator. As Wikipedia is an entirely volunteer run organization (there are only 7 paid staff at the WikiFoundation, none of which are directly involved in the administration or editing of Wikipedia), the amount of work varies from volunteer to volunteer.

me: what does the job of admin entail?  how much time commitment?

durova: Totally flexible.

I could spend ten hours a day or do nothing for a month.

me: yes but the work must pile up if Admins don’t dedicate time

durova: There’s usually enough of us to keep on top of things.  We could always use more help.

I asked for help collecting statistics on the workload.  One sec while I dig up an example.

me: k

durova: Here we go.  In August 2005 the site deleted about 1000 articles a day.

12 months ago the average was 3000 deletions a day.

By 6 months ago that average was 5000.

Another area that eats up a lot of administrative time is Recent Changes patrol.

me: so, on average, there could be up to 150,000 deletions a month

durova: In that neighborhood, yes.

One hundred and fifty thousand deletions a month!  No wonder there are some awfully angry people out there. Many of the complaints against Durova and other editors or administrators come from disgruntled contributors whose postings were deleted. As the popularity of Wikipedia continues to grow, and as search and Internet marketers find ways to use Wikipedia to get solid backlinks or a guaranteed placement for a commercially focused posting, the number of editorial deletions is increasing. Rapidly…

durova: Here’s a humorous look at vandalism patrol.

me: 150K deletions in a month indicates a lot of (what?)

durova: Well, one classic example is garage bands.

A group of friends start a band, try to get gigs, aren’t signed yet.

So one of them writes a Wikipedia article to try to get attention.

Unfortunately there aren’t any reliable sources to prove whether the article is something real or a hoax.  Just the band’s homepage, which can be spoofed.

So we delete that.  Happens all the time.

me: what is the test for reliability, credibility?

durova: For your readership, I should note that a certain percentage of those deletions are about non-notable people and businesses.

Let me give you a link to the notability guideline.

The links lead to detailed discussions for specific areas.

me: ty

durova: Here’s another important link, our articles for deletion page.

Right now there are 96 open discussions for businesses and organizations.

Deletions are not done lightly. There is an enormous amount of discussion surrounding whether to keep or drop a posting. Some specific debates can go on for several pages with tens of editors and administrators writing their opinion of each post. When a super-majority consensus of approximately 75% is reached, a decision is made.

me: what constitutes “consensus”?

durova: Usually, two-thirds.

But closure is complex.

It’s pretty obvious when someone joined up just to comment at a discussion.  Those opinions usually get discounted.

And a single editor’s post may get extra weight for raising one or more important points that the opposing side doesn’t answer.

I was still curious how an article or entry was considered by the editors. One of my personal concerns about Wikipedia is the growing work-load for a staff of volunteers. Wikipedia has become an important social institution with an extraordinary power of placement at Google. A similar thing happened to the Open Directory Project. It grew much faster than its editorial staff could manage.

me: What is the process an entry goes through when made?

(from editor/admin perspective)

durova: Well, each new edit gets listed at Recent Changes.  There’s a link to that at the far left of your screen (depending on which skin you use).

Here it is for simplicity.

Some of the volunteers hang out here to screen for vandalism or new articles that need deletion.

They also add categories if the article doesn’t have any categories.

me: I see over 50 new entries in the past minute or two

durova: Yep.

me: are the rate of entries per day growing?

durova: This isn’t my quote. “At a rate of four hundred words a minute, twenty four hours a day, a person could read nearly twenty million words in a month. In the month of July 2006, Wikipedia grew by over thirty million words. In other words, a sleepless fast reader could never catch up with Wikipedia’s new content. Reading the current incarnation at that rate would take over two years, and by the time they were done, so much would have changed with the parts they had already read that they would have to start over.”

me: ty

durova: np

me: so if the readers can’t keep up, how can the editors?

durova: Heh.

We do our best.

Some bots crawl the site to sniff out vandalism and copyright violation.

me: is there a tremendous backlog of unedited or unfiltered posts?

durova: Mostly Wikipedia keeps on top of it.

Watchlists are important.  I wrote about them in my last article for SEL.

Even with a large number of volunteers, running the Wikipedia can be a frustrating, thankless task. Apparently, the Wikipedians have a very strong and supportive internal community. Though Durova’s name tends to draw heat outside the Wikipedia, reading posts by other admins and editors shows she has quite a bit of internal support.

durova: I’ve been hoping that better communication could reduce the frustration on all sides.

me: what frustration?

durova: 5000 deleted articles a day represents a lot of wasted time.

We all know that Wikipedia is a top Google return.  And of course the professionals in SEO and related fields pay attention to it.

me: oh yeah

durova: There are things you can do to achieve your goals that work within site standards without taking too much of your time.

Most of the professionals don’t know much about what these methods are.

And a lot of the things that get published in the mainstream press are written by people who don’t know Wikipedia very well.

me: What exactly do you mean?  Are you referring to the goals of a marketer or those of an editor?

durova: They may be well-intentioned, but they often misrepresent one or more aspects of site policy.  Their readers read those mistakes and follow them, then our volunteers have to go to work. Now in SEO your basic goal is to send readers to websites.  Would you say that’s a fair summary?

me: Absolutely. Increased traffic is the goal. That was generally done via search engines. Now, Wikipedia is sort of a search engine that has at least one guaranteed placement in Google SERPs

In this and a few other ways, Wikipedia is starting to remind me of the Open Directory project.  SEOs are going to pay a HUGE amount of attention to Wikipedia in the coming year

durova: Well the way to get durable outgoing links from Wikipedia to a client’s site is to treat Wikipedia as an encyclopedia first.  Its social media aspects exist to support the encyclopedia.

I wrote about one way to do that in my first SEL column.

My next column is going to address another way to create durable links.

me: excellent

durova: Wikipedians don’t like linkspam, but they don’t begrudge links that serve a clear encyclopedic purpose.

The interview meandered over several other topics ranging from her personal experience at Wikipedia to the relationship between Wikipedia and Google:

me: Does Wikipedia have a relationship with Google?

durova: Yes, we run away for intimate weekends in the mountains and cuddle up on the bearskin rug by the fire.

me: do you think they respect you in the morning?

durova: We’ll see how Wikia does in the marketplace.

I have no connection whatsoever to Wikia, by the way.

me: I figured that. Wikia is still very small

durova: I think Google is very smart and is watching that very closely, and probably has a business plan to address it if it takes off.

me: Oh yeah. Google watches the space like nobody else’s business.  seriously though. Does Wikipedia have any agreements w/Google regarding content?

durova: Not that I know of.  Wikipedia is all copyleft licensure, so anybody can use the material for free if they credit the source.

But a lot of people wonder why Wikipedia consistently ranks so high.

me: does the fact that entries will almost certainly get high placements under relevant keywords have any bearing on editorial decisions?

lots of links.

durova: Among Wikipedia regulars, no.  Among occasional visitors and newcomers, very much so.

So to the extent that the hardcores like me deal with it, we’re taking out links and explaining policies to those people.

I know this isn’t the ivory tower.  Wikipedia content and links have real world impact.

That’s a reality and I deal with it.

Occasionally I siteban people because of it.

That sitebanning is what gets folks so angry with Durova. There is little she can do about that except try to do her job with consistency and relative compassion. More on the argument between Durova and certain members of the search marketing community next week.

Author:  Jim Hedger is Executive Editor of Jim is also a writer, broadcaster, commentator and analyst in the search marketing sector. He is an avid Wikipedia user and temperate Wikipedia supporter.