Archive for November, 1999

Peter Cutler, @studio525, is obsessed with the wrong metrics

November 29th, 1999

Peter Cutler has 17,000 followers on Twitter. He gets them largely through gaming the system, looking for people who will follow anyone or who autofollow. He’s pretty indiscriminate in his follows and he doesn’t weight things like follower to followee ratio, picks up people using keyword search like SEO and marketing.  He’s not particularly interested in personalizing his tweets to people he follows who may be curious as to why he follows them. He’s busy pushing some links, bragging about his Twitter Grade, etc.

In other words, he is a social media spammer.  If you have 17,000 followers, there is no reason to be following new people.  You can’t possibly follow all 17,000 people.  You can follow some of them if you add them to lists.  (And in my case, he didn’t add me to any.)  Total followers is a meaningless metric that Peter Cutler is obsessed with, as he’s actively trying to grow his count.    The people clicking through links for a few thousand followers, where the followers were gained through autofollow attempts, is pretty low.  (I can break it down for you using Fan History if you care for data.)  The ROI is pretty awful in that regards.  Number of followers might be a nice ego boost.  (Kind of like people who write Bella/Edward Twilight stories to get a lot of readers.  Quality is often irrelevant there.)  He can’t articulate why follows individuals.  He offers little value for those following him in return.   He’s hard to reach, not responsive to people who follow him while he deliberately seeks out those individuals.  He’s gaming the system and a twitter follow spammer.

He might be a fine Creative Director, but he’s not some one I would ever point you at for social media he’s obsessed with the wrong metrics and his methodologies could cause backlash as he doesn’t think through the consequences of, well, you know, actions like gaming the system. 

Mahalo aggressively goes after Wikia

November 29th, 1999

Yes, Jason is trying to get people to leave wikia. (He has never approached us!) Given the licenses of most wikis on Wikia, he doesn’t really need to approach anyone: He can just copy the whole thing on to Mahalo, find some one willing to curate and be done with it. If he is only offering half the money to one person, large wikis aren’t going to move. If he doesn’t have the domain, there isn’t going to be a lot of money unless he can squeeze Wikia and beat them on SEO for the same content.

It will be interesting to see how this plays out.

Edited to add: Linked to Jason’s twitter account and the leavewikia account on #wikia on irc.freenode.net to mostly complain about Jason’s lack of understanding of wiki communities and fan communities. The result: [14:42] -ChanServ- You have been devoiced on #wikia by Dantman (Nadir_Seen_Fire) .

What country do Twilight fans live in?

November 29th, 1999

Where do Twilight fans live in?  The following chart should help give you an idea.

Where do American Twilight fans live?

November 29th, 1999

This is a continuation of my look at the Twilight fandom demographics that I started in lion_lamb: A sneak peak into the composition of the Twilight fandom. The idea is to get a better grasp of the demographics of the fandom.

One question I had is where did American Twilight fans live?  Were their states that just appeared to have a bigger fanbase than other states?  To answer this question, I gathered two sets of data one for LiveJournal and one for Facebook. 

I got data off LiveJournal following the methodology described in the lion_lamb post.  The difference was that I used a bigger pool of communities.  Communities represented by this sample include: jasperalice, lion_lamb, loveattwilight, french_twilight, i_hate_twilight, just_cullen, obs_twilighter, ed_bell_lovers, lionlambinlove, forks_high, jacobxnessie, early_evening, forksradio, all_pattinson, beautiful_ks, twispammers, twilighttees, twilightsongfic, twilight_book, twilight_addict, twilight_adult, sortofbeautiful, ru_pattison, ontd_twatlight, pattinson_daily, tm_switzerland, sv_twilight, thosetopaz_eyes, thosetopazeyes_, twsecret, twilightsagafan, twilightmanips, twilightfriends, twilight_xover, twilight_web, twilight_stamp, twilight_slash, twilight_series, twilight_quotes, twilight_ger, twilight_film, twilight_fics, twilight_aus, twilight100, twilicons, twiicons, twiconic, wildlove_jblack, weepdiscreetly, vavri, to_our_meadow, the_temptation, the_gazebo, robert_kristen, and realtw_secret.    All this data was collected on November 15, 2009.   There were 54,739 entries. No duplicate rows for individuals were removed as the assumption was made that duplicates would represent unreported data.  After compiling the list, all people who did not list their location were removed.  This brought the total down to 29,956 people.  After that, all people who did not list the United States as their country of residence.  That brought the total down to 16,728.  Next, all people who did not list a a specific location in the US were removed.  That brought the total down to 15,906.  Then people where the state could not be ascertained were removed.  This left 14,806 people for this sample.  It is important to note: This sample is too small to pick up the real size and scope of Twilight fandom on LiveJournal.  We would need to probably add another 200 communities to get to that stage where we can get a feel for its true size.   We do believe it to be representative of the community and these numbers speak for trends.   We base this on the fact that there was less than .6 change for the average year of birth from lion_lamb to the whole of this sample.

The data from Facebook was gathered using  Facebook’s Advertise on Facebook.  First, I made sure the fields were set to allow the maximum people included: No segmenting based on age, relationship status, etc.   I set the keyword interest as Twilight.  Then, I went down the line with segmenting by state and recorded the number of people from that state with that interest.  This was done on November 18, 2009.  2,393,460 people listed Twilight as an interest.  I believe the distribution using this method is accurate regarding the Twilight community on Facebook but it too has issues similar.  Most people update their interests when they join Facebook and not often after that.  People may be interested in Twilight and join fan communities but not list Twilight as an interest.

All of that out of the way, this data was compiled and the following two charts were created:

The Twilight fandom isn’t that big in the Mountain West, which makes sense as there is a smaller population pool to draw from though this is more the case on LiveJournal than Facebook. Twilight is big in the South, the Northeast and the Great Lakes. Proportionally LiveJournal fans are a bit more active in the Northeast than Facebook fans. For a look using hard numbers, the following chart gives a much clearer picture.

Much of this stands on its own for you to draw your own conclusions.  What interests me is when this data is then compared to US population rankings for 2003.  Where Washington D.C. ranks 50th for population by state, it ranks 42 on LiveJournal and 28th on Facebook.  That seems to be a huge difference, with DC over representing in terms of Twilight fans.  Missouri is another interesting case.  It ranks as the 17th most populace state.  On Facebook, it ranks as the 10th most populace state for Twilight fans.  LiveJournal is a lot different, where it is the 27th most popular state for Twilight fans.  One has a large over representation and the other has a large over representation.  New Mexico ranks as the 36th most popular state but on Facebook, it ranks as the 46th most populace state for Twilight fans.   Utah is the 34th most populace state but on Facebook, it ranks as 23rd.  That is an 11 state difference in rank.  This is not all that surprising given that the Twilight books were authored by a Mormon who hails from the state.  The last big difference is Washington, where the state ranks 15th in terms of population but on LiveJournal ranks 5th in terms of Twilight population.  

lion_lamb: A sneak peak into the composition of the Twilight fandom

November 29th, 1999

lion_lamb is one of the biggest Twilight specific LiveJournal communities. It has, as over November 15, 2009, about 16,600 members, 23,128 Journal Entries, 113 Tags, 1 Memory, 44 Virtual Gifts, and 9 Userpics. It is big and as influential as a number of other fansites in terms of sharing information.

I’ve always been interested in demographic surveys of fandom. There has been a fair amount of market research done that would be of interest to people involved with fan studies but a lot of that remains locked, controled by the people who produce it. There are people who produce great big data sets involving social media sites but a lot of that work is not easily minable by fan studies folks either, or the scope excludes them. Getting a picture is just hard.

So when one of the founders of YourWiki offered to help me get some demographic data off LiveJournal, I leaped at the chance. (And then nagged, harassed and busted his bandwidth limits. We’re special like that.) I wanted some data regarding lion_lamb because it is so huge and I figured it could give a nice snapshot of the Twilight fandom on LiveJournal. They obviously aren’t representative of the whole of the Twilight fandom. (No one has ever done a study that has linked the characteristics of LiveJournal fandom to the characteristics of broader fandom. It would interesting to do. If you’ve got some data scraping skills, have analytic data from your own site, run a convention, let me know and we can see if we can do that.) Still, the represent a fairly large group of fans who have influence on many other social networks including Facebook, Bebo, Quizilla, orkut, Twitter, Twilight Moms, etc. They are worth looking at.

First, we checked LiveJournal’s bot policies. Then we created a bot that went to the profile page of a LiveJournal community, grabbed the list of 11,000 or so people who both watched and are members of the community, and accessed the user page of every member of that community. The bot then collected data on location, birth year, birth date, year that the person created their LiveJournal, date that the person created their LiveJournal, year the person last updated their LiveJournal, date the person last updated their Livejournal, account type (Basic, Plus, Paid, Permanent, Sponsored, Early adopter), total number of journal entries, total tags, memories, virtual gifts, user pics and friends. This information was then dumped into a csv file. (More on methodology)

We now have a lot of data. It has a lot of holes in it because users do not need to complete all the fields. For the sake of this sample, we are going to assume that the non-blanks are representative of the whole set. People also lie about their ages and location. (People are not living in Antarctica or the Romulan Neutral Zone for instance. They are not likely to be born in 1900 or 2007.) Still, in terms of accuracy of the sample, I think it is better than survey research that may be done on community and it is less self selecting.

So what do we know about the watchers of lion_lamb?

lion_lamb watchers don’t pay for LiveJournal. In order to get extra features, the vast majority are willing to have extra ads displayed on their LiveJournals.

Things change a bit when you look at this based on nationality.

There are 3,378 people that list themselves as from the United States. Americans are a bit more willing (or able) to pay for their accounts and less willing to have as many ads.

There are around 300 Australians, 364 Canadians, 159 Brazilians, 197 Germans, 107 Mexicans, 23 New Zealanders, 28 Irish, 332 people from the United Kingdom, 116 French, 16 Japanese, 107 Italians.

Amongst those countries, Australians are the ones most likely to have permanent accounts with two percent of Australians having this account type. Permanent accounts go on sale around once a year for $150 to $250. Canada, Germany, Mexico, the United States and the United Kingdom all have one percent of total members with that account type.

Plus accounts are most popular in Brazil (87%), France (80%), Italy (85%), and Mexico (86%). Plus accounts are least popular in Germany (69%), and Japan (63%).

Paid accounts are most popular in Ireland (18%), Germany (16%) and the United Kingdom (14%). They are least popular in Brazil (4%), France (8%) and Mexico (5%).

Basic accounts are most popular in Japan at 25%. No other country in this sample is in the twenty percent range. Basic accounts are least popular in Brazil (9%), Ireland (4%), Italy (5%) and Mexico. (8%)

These patterns are interesting. Why are Japanese Twilight fans willing to do with less extra features with the reward of fewer ads? Why are Brazilians, Italians and Mexicans willing to have more ads? Why are the Irish, Germans and Brits most willing to pay for extras?

I’ve been curious about Twilight fandom on LiveJournal: How old are people in the fandom and did the fandom give LiveJournal a spike in new users? The fandom is characterized in many places as being filled with teenage girls. On LiveJournal, based on lion_lamb, that doesn’t appear that true. The average year of birth for people who listed theirs is July 1985. That’s about 24 years old, past high school and past college.

This average differs a bit my country. There are 1,193 people from the US, 135 from Canada, 48 from Brazil, 33 from Mexico, 139 from the United Kingdom, 86 from Germany, 122 from Australia, 50 from France, 9 from Ireland, 51 from Italy, 7 from Japan, and 7 from New Zealand.

The above chart doesn’t include the total people born by year before 1960 or after 1995 but people outside the range were included in the average. Brazilian, German, Japanese, Australian, and American members are the oldest at about 25 years old having been born in 1984. French and British fans are the youngest averaging the age of 22 and being born in 1987. The mode for the whole sample is 1989 and the median age is 1987. These are different than the average, implying that the community is a bit younger than the average implies.

Twilight may have given LiveJournal a registration bump in terms of this community, with the median and mode of 2008 for registering. The average registration year is 2007.07 in comparison. Members of this community are updating, with a last update year average of 2008.66, mode of 2009 and median of 2009.

lion_lamb members have posted an average of 132.25 times, a median of 11 times and a mode of 1 time. This indicates that several members are extremely prolific on their own journals. Two members have nine thousand plus times, one posted eight thousand plus times, two posted six thousand plus times. 320 people more than one thousand posts. This small group of prolific posts, representing two percentage of the total members, skew the average rather high. Most members are posting between one and eleven times.

For tags, the average number per person is 36.48, the mode is two, the median is four. For memories, the average is 37.77 per person, mode is zero and median is zero. Virtual gifts per user has an average of .23, median and mode of zero. The average number of userpics per user is 13.94, with a mode of one and median of fix. The average member of
lion_lamb has 34.08 friends, median of eleven friends and mode of one friend. Most of these you would only expect to have higher numbers if the community was full of active contributors who were utilizing all the features on LiveJournal.

lion_lamb has a number of users who are a bit older than the screaming teenage girl stereotype, though certain countries represented on this community are closer to that than others. Most users are not using plus accounts, choosing to view additional ads to get extras on the service. Members of the community are not very active on their own journals and many are relatively new to the site, having joined in the past twenty two months.

As a snapshot of the Twilight fandom on LiveJournal, it is an interesting one. It will be interesting to see if this sample mirrors other LiveJournal communities or as part of a larger sample of members of Twilight communities on LiveJournal.


Why didn’t I chose Twilight as an interest and then pull profiles based on that to get an idea about the shape of the Twilight fandom on LiveJournal? Because that’s largely a passive activity. First, Twilight doesn’t always mean series of books by Stephanie Meyer; it could mean that time when the sun sets. I also felt that joining a group was more active indicator of interest in the books and movies: You want to read what other people are saying so you join to do so.

Fan History organizational tree: Harry Potter

November 29th, 1999

During the past few days, we’ve been working on trying to visualize our organizational patterns on Fan History in order to understand our own patterns, how we conceptualize fandom, to check for organizational consistency, create tools to help users understand our organizational patterns, to identify areas where we lack stub content. This is our second post in this series. This one is about Harry Potter.

Click on the image to view it full size.

There are things about the top level category that drive me nuts. Namely, the category seems to have way too many categories for specific sites: LiveJournal, GreatestJournal, EZBoard, InvisionFree, JournalFen, Yahoo!Groups. I’m not certain how this can be fixed with out a major category restructure. The major hesitation about doing that is it would disturb our existing patterns, would make the tree a lot deeper and harder to navigate.

Once I start looking at the structure regarding communities in depth, it appears that we have started to move that content into its own separate structure. It just was never fully implemented. Or quite possibly, it was implemented and then people went in and changed it not realizing that we had attempted a restructure. This is a wiki and anything is possible.

Our character and pairing structure was taken apart at one point, with articles removed from them. There is the whole question of if we should restore that and put the articles back into the subcategories and restore that. Some like Draco/Hermione have a lot of content that could go in them including stories around the ship, etc. Having that structure existing just makes it easier to navigate and conceptualize in my opinion. If you specifically find that Snape/Harry LiveJournal community, than it helps to have that sorted if we don’t have an article list of Snape/Harry LiveJournal communities.  

The Harry Potter fandom also seems to be structured around pairings, characters and houses.  The community doesn’t necessarily have the emphasis on a coherent community identity.  Thus, it seems logical to me to structure the fandom more around these articles and rebuild them that way as a history of the fandom would be much more accurate if told from that perspective.

There is also the issue of how to organize fan fiction. We have fan fiction archives, fan fiction, stories, slash, saffic… and all of these deal with fan fiction. Should those sub categories be moved into Category:Harry Potter fan fiction? And looking at that category now, it really looks like everything in it should be moved to Category:Harry Potter stories. Oye.  If we combined communities and organized that way, it feels like it would be logical to combine this one too.  Things just seem really scattered like this; would it make it easier for article contributors to link various parts of the history they tell if the category structure made those connections more obvious?

If you have any feedback on this tree, any questions about how it developed, we would love your feedback. Do the organic patterns we’ve developed make sense? Is this construction too artificial? Is it not logical? Does it make sense in relation to the fandom this structure represents? And if you’re really motivated, we’ve really like that feedback on the relevant talk pages for those categories.

Get out the umbrella bat signal ™! Where is the Organization for Transformative Works?

November 29th, 1999

Fandom has activated its Batman ™ Umbrella symbol in the night sky and is desperately seeking the Organization for Transformative Works!

If you haven’t heard, the Obama administration is trying to pull a Joker and sneak through copyright legislation that would make it impossible for fansites, fan fiction archives and fan vidders to operate. Our fandom caped crusaders for justice are busy knocking boots with their main Catgirl like squeeze, Yuletide, wasting valuable time like they did with their old and forgotten Albert like friend, aka Geocities.

So we here at Fan History are sending out a Batman ™ Umbrella symbol to them to get their bat umbrella back on task because they can do what we at Fan History cannot. They can fight the evil Joker! They have a legal branch with lawyers ready to do battle. Help us Organization for Transformative Works! You’re only hope!

Largest fandoms by total number of wikis on Wikia

November 29th, 1999

We recently created a collection of articles about fandom related wikis on Wikia. The list isn’t complete as Wikia has over 24,000 wikis listed on their own internal list. Six to eight hours of data crunching left us exhausted and so we have about 8,000 or so wikis we haven’t looked at, couldn’t figure out how to handle the problem of ten wikis from the same fandom and language having the same name. That said, we can still give you an idea of some of the most popular fandoms to create wikis about. They are:

  1. Star Wars – 161+ wikis
  2. Pokemon – 135+ wikis
  3. Music wikis (not fandom specific) – 102+ wikis
  4. Movie wikis (not fandom specific) – 100+ wikis
  5. Runsescape – 91+ wikis
  6. Naruto – 88+ wikis
  7. Warcraft – 84+ wikis
  8. Nintendo (non fandom specific) – 75+ wikis
  9. Anime (non fandom specific) – 70+ wikis
  10. Halo – 66+ wikis
  11. Club Penguin – 62+ wikis
  12. Yu-Gi-Oh – 59+ wikis
  13. Dragon Ball – 48+ wikis
  14. Fan fiction (non fandom specific) – 46+ wikis
  15. LEGO (non fandom specific) – 45+ wikis
  16. Comics (non fandom specific) – 44+ wikis

One person cannot do it all… What we should have done differently…

November 29th, 1999

When looking at large scale preservation projects, a lesson I have learned from the Geocities situation, is that one person cannot do it all.  Two people cannot do it all.   Three people cannot do it all.  Four people cannot do it all, even if they are dedicated, well meaning, have certain skills and are willing to put the time into saving and preserving the history of an online community.  I’m ecstatic about what we accomplished but I’m also disappointed.  I feel like we could have done more, if there was more than a core team of a few of us at Fan History. 

In doing a preservation project, one person trying to save things means that things will be missed.  The process can be greatly aided by institutional help in determining the size and scope of what needs to be saved, and in developing a list of resources that need to be saved.  The following of institutional structures that would have been useful for us to have had relationships with:

  • Yahoo/Geocities.  ArchiveTeam tried to reach out to them and were rebuffed.
  • Open Directory Project.  We didn’t really need them because they helpfully provide a list of all the sites listed on DMOZ.  We were able to utilize this.  For that, we are happy.
  • WebRing.  Lots of Geocities sites hosted there.  Couldn’t scrape them easily to try to easily pull a url list.  Didn’t respond to our contact requests.  Couldn’t find anyone associated with them on Twitter to help us get that.  We probably should have called and called and nagged. :/
  • ArchiveTeam.  We used one of the lists they provided to get urls to scrape but most of what they provided publicly wasn’t that useful for our needs.  I did a comment on their blog.   We should have reached out more to them.  They had the technical knowledge to do things, had the computers, had the dedication to do this.  We have the structure to provide a historical framework for some of what they were doing.
  • The Internet Archive. They archive old copies of pages across the Internet.  
  • Organization for Transformative Works.   They are dedicated to preserving fandom history.  Repeatedly reached out to them on Twitter, on LiveJournal, via e-mail and elsewhere.  Never responded.  EXTREMELY FRUSTRATING. They have manpower and a shared purpose that would have been helpful.

We should have gotten institutional support.  They could have provided us with three things: 1] A (structured) list of fandom related urls, 2] Technical assistance, 3] Community support and manpower.  Some of that we got because these institutions provide some of this as part of their own efforts and mission.

After institutional support, we needed technical assistance.  We were lucky in that we got some assistance from Lewis Collard and illyism.   We just didn’t think to ask some people until too late, about two weeks before Geocities closed.  Asking two individuals for all our technical assistance and implementation of technical solutions isn’t fair, won’t result in a timely solution.  It doesn’t allow enough time for integration with institutional support.  Lewis scraped about 5,000 pages and screencapped them.  He downloaded about 2,000 files.  In doing all this, he blew through his monthly bandwidth allotment. Illyism developed a Firefox plugin.   There are several resources we could have, probably should have tapped in our own community, the wiki community and the fandom community.  They include:

  • #mediawiki, #wiki, #yourwiki, #recentchangescamp, #wikia, #pywikipediabot, #wikihow on irc.freenode.net.  These chat rooms have awesome people in them.  They do a lot of open source projects. 
  • Organization for Transformative Works.  They are training female programmers for their projects. 
  • ArchiveTeam and Internet Archive.  They already have technical people. 
  • Idealist.Org.  Listing requests for volunteers might have gotten some assistance.
  • Wikimedia related mailing lists.
  • Tech and programming people we met at RecentChangesCamp.
  • Relevant LinkedIn communities.
  • weget developer community.
  • identi.ca.  They have a lot of developers on the site who use it instead of Twitter.  It is smaller and more intimate.  It has a lot of open source advocates who like doing things for the greater good.

There are probably a few more places where we could have asked for help.  We could have used help creating a list or urls, screencapping sites, creating tools to automatically input the attained data into a usable format, and tools to make it easier for non-tech people to contribute.

The last component in the we cannot do it by ourselves and succeed is community support.  In our case, that community involves fandom.  We needed the community to help edit relevant articles to do deeper documentation on stub articles created by our core team and created with assistance from our tech team.  Sometimes in fandom, it is easy to get locked into a box of thinking of your corner as representative of everyone.  The more people we had, the smaller that box would have been.  We should have been more aggressive with our outreach to that community and to internet news and mainstream news sites that would have covered our project and helped us get greater exposure with our core audience.  Places we should have been more aggressive include:

  • wikiFur, Transformers Wiki, Futurama Wiki, Battlestar Galactica Wiki, Wikia, other fandom related wikis.  The content we created on Fan History that preserved the history of those fandoms could have been shared with them so everyone could have won.  If we had reached out.
  • Twitter.  We should have been tweeting damned near every fan person who mentioned Geocities on Twitter to beg for assistance.
  • Facebook, MySpace, Yahoo!Groups, Ning.  We didn’t reach out there at all.  Large fan communities exist on them and are ones that probably used Geocities.
  • Fanac Fan History Project and other science fiction historical projects.  They stood/stand to lose a lot of their own history too.  This includes convention reports, early histories of scifi fandom on Geocities, pictures, etc.
  • Fansites.  Lots of domain level fansites will plug meaningful projects.  They could have helped us get the word out, helped us to recruit people with this project.

We mostly tried to mobile through LiveJournal, through blogging outreach, on IRC, through contacts on the wiki community (Thank you ProjectOregon and AboutUs!) and our own personal network of acquaintances.  This wasn’t as successful as a method as it could have been.

Repeating: One person (or a small team) cannot do it all.  For this type of project to be successful, you need three things: Institutional support, technical support and community support.  If we hear that other sites are at risk in the future, we now know what needs to be done and better understand the consequences of failing.  We’ve learned an important lesson.

In the meantime, we’ve still got a lot of data that needs to be processed into the wiki into a usable format.  We could use some technical and community assistance in accomplishing that.  How do we get 3 gigs of screencaps uploaded?  What is the best way to upload 1,000 text files to a wiki?  Is there a way to scan those text files to make sure that they are what we think they are?  As a community member, can you help us improve articles we did create?  Can you help by improving descriptions on the screencaps we do have uploaded? We still need help.

Now, I’m going [maybe] to take a bit of a wiki break, data mining break, data processing break because I’ve pretty much been doing that straight through for the past week.

Canonical URL by SEO No Duplicate WordPress Plugin