A video http://bambuser.com/channel/pigsonthewing/broadcast/2140682 of Sue talking about the difficult issue of editor retention. I haven't followed everything she's said in public, but she seems much more open here and less reserved than I thought.
She talks about what WMF call the the 'holy shit' slide, showing that editing peaked at 2005. She says she is not going to use the words 'death spiral' in public (but then goes on to use the term not a few times). It is essentially "endless September".
Below is what I summarised from listening to the first hour, which was enough. I have a lot of time for what she is saying, in particular her point that it is the drive to quality that seems to be pulling the numbers down. I don't agree that it is possible to draw a compromise between those two, as she seems to think. Sue is essentially a tree hugger, witness her visit to Occupy Wall Street a few weeks ago. Her idea that you can resolve problems by sending pictures of kittens and just being generally huggy and pink and glittery is generally misconceived.
Interesting that there seems to be general acknowledgment at the Foundation that there is a rogue admin problem. Do they have the guts to do anything about it though?
New editors aren't making it to their first year anniversary. People are coming in large numbers as they always have, but they are getting rebuffed. Why are they failing to enter community. Warnings have gone up, criticism is 'way up', and praise and thanks have been in decline. People join and then it's warning template after template. The templates are well-intended but they make new editors feel like 'Wikipedia hates you'. Pre-2006 there were no automated warning. Established editors would talk to new editors and help them through. Now, 4 out of 5 messages are bot-delivered.
Someone asked about the 'goldrush' hypothesis - early gains are over. The low hanging fruit has all been picked. Sue said that "There is very little evidence to support that. On the contrary, we see the same pattern on other Wikipedias.Even the Hindi Wikipedia, which has only 5,000 articles.After a couple of years a Wiki gets filled up with rules and templates. Deletions and reversions stop people hanging around."
The indicator of good Wikipedian is to make lots of edits in first few days. But these are the most disincentivised. "People think the website is yelling at them. "
In 2005-6 the wall started to go up. She discusses FA and how the bar has got so high that certain articles will never go to FA. "I honestly think we have lost our way, I really do right. I sometimes think we have become Nupedia, and we need another Wikipedia to feed into it. I just think we are really really rigid ... We have lost sight of what makes the project special, which is that everybody really does have something to contribute, it's not just a priesthood".
Someone comments about a user who started an article on their user page, and told that if they did not stop their disruptive editing they will be blocked. Sue thinks this came out of the Seigenthaler thing, when Jimmy insisted there must be a focus on quality, also the same time as Essjay. "There was a moral panic created around quality". This gave a whole lot of people license to be jerks. People are playing Wikipedia like a video game, shooting down vandals, and every now and then a nun or a tourist wanders in front of the AK47 and just gets murdered. "What we think now is that it's all nuns and tourists. There's a massacre and there's one vandal in the background running away. And meanwhile everyone else is dead."
"I deliberately kept out the phrase 'death spiral', because I didn't want to put it in writing. It's too depressing".
So what shall we do. The WMF rarely speaks directly to editors. It used to be Jimmy who would proclaim something. As an movement develops matures it needs institutions, not individuals. Jimmy still plays a role as an individual, but the board, the institution plays much more of a role today. The first time was the BLP resolution. The board said on March 2011 that editor retention was its top priority. (At this point they read something which is not spoken aloud).
Are admins the problem, particularly rogue admins? (32:00) Sue mentions someone on Foundation-l (or internal l?) who was being 'a real jerk', and a number of people even called him on it. He told her "I don't think I was being uncivil, I was being icily sarcastic". Is icily sarcastic what we aim for? Is that success? "I would interpret our requirement for civility as a flaw. You cannot get worse than that". "As though I can be barely civil and that's what we are supposed to do here".
She then goes on about Occupy Wall St, where she was (at Manhattan) a few weeks ago. The think she liked most was that everyone there took responsibility for maintaining a good tone. She does think that Arbcom should play a role. Not banning, but maybe 'taking a role'.
She notes that the smaller projects have a big problem, e.g. Wikiversity.
What is WMF doing? The analysis comes first, on Wikimedia. She thinks there is a false dichotomy between quality and retention, as though there is a conflict between 'good' and 'open'. She looked at the abuse filter. She thinks this is great for participation, and great for quality.
40:48 - have they tried to identify who the top newbie killers are? They can't, but they have been thinking about mechanisms. "There are people we call 'moth people'. They are drawn to the flame. And some people are drawn to the flame in a good way (they negotiate, they mediate), others in a bad way, they like fighting. "Its really hard for data to tell you things like that". They had thought about a flag to identify the bad moth people.
The volunteers tend to increase complexity. The job of the WMF software engineers is to reduce complexity. [And 55 minutes was as far as I got - mostly questions and answers by this stage].