Azz (bolt of blue) Обещанного три года ждут. (azurelunatic) wrote,
  • Location: 85225
  • Mood: ...okay, maybe more like 10k words. Heh.
  • Music: Cowboy Mouth, "Jenny Says"

Turn off the radio; turn off the lights

Being a few words of advice to anyone stepping in to a LiveJournal management position.


I have used LiveJournal for over seven years now. I signed up in May 2001. I use LiveJournal as a journal, as a communications service, as entertainment, and much more. I have dedicated countless hours to not only using the service, but also helping make it better.

It is a nightmare to think of LiveJournal in the hands of people who drive it into the ground in a misguided effort to make it succeed. It is easy to make management assumptions that are completely rational given a generic online service, but are entirely the wrong thing for LiveJournal due to its culture and history. This document covers not business management, but user management, as LiveJournal's users are an idiosyncratic group with strong opinions, a history, and disparate factions with different goals and different uses for the service. The unifying wish of all groups of users is for LiveJournal to succeed, still be around into the far future, and to remain a fun and free-thinking place, in the spirit of the way it was under the watch of bradfitz.

To manage LiveJournal properly, you must be:
  • familiar with the service yourself
  • invested in nurturing and maintaining the current user experience as well as developing in exciting new directions
  • willing to engage with the userbase as people who hold a stake in LiveJournal's development and future, not just content providers or visitors
    (There will be members of the service whose input is misdirected, or who do not have the background to give useful advice, however, this is not the document for covering those issues.)
  • aware of the history that your users will be aware of, even if not in the same detail
  • able to realize that there is no one true way to use LiveJournal

Do not underestimate the amount of time and energy that people put into LiveJournal. There are people who barely use it. There are people who use it a reasonable amount and have no more attachment to it than they have to any other company providing a needed service: SRP for power, Cox for internet, Gmail for mail, LJ for blogging. The real strength of LiveJournal, however, is that the most active users have formed an emotional attachment to the community above and beyond their need for the service that LiveJournal provides to them.

Brushing off user complaints with "It's just a website on the internet; it's really not that important," is dead wrong from a management standpoint, and worse, incorrect as far as active and passionate users consider the site. There are people who will become exceptionally upset if the service becomes unstable or their site experience alters substantially. (Whether or not this is healthy is a question for their psychotherapist. You still have to deal with the results.) You will not need to explicitly seek out the feedback of the most passionate users; you will have it, whether you ask for it or not. If your passionate users are angry, you will hear about it, often through cat macros in news posts. If your passionate users are happy and dedicated, this may not show up directly, but accounts will continue to sell, people will express their devotion and love for LJ, and send cookies. Real, tasty ones.


Familiarity

Every staff member joining the LiveJournal family can be issued a permanent account. You are automatically given the first-class experience, whether or not you have any interest in taking full advantage of it. At the time of this writing, Permanent Accounts will soon be going on sale for $175 (USD). I recommend you make use of this, even just to play around every now and then and see what can and cannot be done. Every user won't expect you to automatically know everything about everything, but employees are generally expected to have demonstrated that they know something about something.

Possession and frequent use of an off-LJ blog may gain you more credibility with the rest of the blogosphere, but will not count for particularly much with LiveJournal users. If you want the attention of the average LJ user, you will need to update your journal at least once (even if it is only a notification that the journal does not provide public content if you're using it locked down, or a test post, or something saying that there's not likely to be much here, check there), upload a userpic or two, and use it to comment as appropriate. No one's really going to expect the juicy details of your private life, but the odd professional or silly linkblog entry wouldn't go amiss.

Syndicated feeds can help if you'd like to use any offsite blog to gain you credibility on LJ. To do this, mirror or syndicate it to LJ, and make sure that it is listed in connection with your Official Work journal, ideally explicitly stated in your profile as well as listed as your website. (People's listed website may fall under LJ-specific banner blindness and just plain not get seen. There's also a very real unwillingness to leave the site for some people when exploring around LJ. You don't have to understand why, just realize that it exists.)

Familiarity with the day to day actions of using the site will give you a better appreciation for what your users are doing on a daily basis. You may soon come to form your own opinions about how the site should work technically, if there are aspects of it that you love and would like to see more of, or that you really cannot stand. Share these thoughts with the appropriate people in development. These are the things that your users are going through, except you're in a position to say "Please, sir, may I have another?" and "Stop the pain!"

While it is not mandatory that you do so, I also highly recommend creating a journal that remains Basic or Plus, and can be switched back and forth, to get a picture of the site experience for someone who does not have the advantage of a paid or better account. While people using free accounts of either type may not be contributing directly to the revenue of the site, this does not mean that they have not paid in the past or will not pay in the future. Furthermore, free users who actively participate by posting and commenting, or by recommending LJ-hosted content elsewhere, contribute to the site user experience and draw traffic who in turn participate and bring revenue.

Exploration around LiveJournal above and beyond the official communities will give you a sampling of what people are really here for. There are many different and splendid things that LiveJournal has to offer, and you will get a chance to see what keeps people here. Looking at comments to official communities to try to get a picture of the site as a whole is about as accurate as attempting to suss out the culture of the United States from visiting a political rally. (Sometimes it's more like visiting the kind of demonstration where the police break out the riot hoses.) It's neither an accurate nor a fair picture.

Go add a comic feed or two, go check out a community that caters to your interests, go read a few journal entries that are heavily linked, help contribute to dragging down the cluster upon which ohnotheydidnt resides (and then get out while you still have your sanity if you don't like celebrity gossip). Find a journal at random. Find someone whose writing you absolutely love. Find someone whose journal is *not* full of quizzes. Find someone whose style makes your eyes bleed. Find someone whose style you want to steal. Find someone whose content appalls you. Interest-surf. Play profile-surfing where you start out from a random journal and see if you can get back to places you recognize by only clicking on links to the profiles of other users. (Using one of the default communities is cheating.) Cruise by the_lj_herald. Find a cat macro you like in news comments. Cruise through http://www.livejournal.com/support/help.bml and poke around in the journals you find in http://www.livejournal.com/support/highscores.bml.

Use of the site will cultivate the users' familiarity with you as well. Unlike many other online services, where the administrative team is essentially faceless and interchangeable, LiveJournal users became accustomed to looking to bradfitz as not only the founder of the site, but as a real live human being who was actively engaged in doing work on and for the site.

It behooves you to be very careful with your interactions with users if you will be interacting outside of an official context. Anything you say while under the login name of your official employee journal will be taken as gospel, and a promise, unless you are very careful to draw a line between what is an official statement and what is a statement of opinion. Any user-facing position can quickly become an exercise in politics. A few people of note have encountered problems with off-the-cuff statements being taken wildly out of context and drawing fire. You may wish to consider using that second journal to interact unofficially, to better draw a line between your official presence and your presence while off-duty. Brad maintains both bradfitz, the journal he used for official functions, and his first journal, brad.


A Question of Character

bradfitz founded LiveJournal, managed LiveJournal, and was LiveJournal until he handed it over. Brad had his faults, as everyone was quick to point out during his ownership and administration, but there was rarely difficulty in empathizing to at least some degree with the guy whose small project blew up enterprise-size.

There are other colorful and well-known personalities in LJ's history, although there is not necessarily any correlation between internal and external fame. Anyone who says something sufficiently notorious in public can get a reputation. Someone who is well-known and well-loved inside the company may be a complete unknown to the general public. Someone who has fallen from grace internally can still enjoy a soaring reputation with the userbase. However, anyone who does any degree of interaction with users in any kind of official role will develop a reputation.

Over the years, blame has fallen upon the head of the messenger, so that making an unpopular news post reflected poorly on the person who posted it, no matter who actually made the decision. While I am not privy to any of the discussions behind the scenes, I seriously suspect that there were some good cop/bad cop games being played with bradfitz and rahaeli's reputations. The recent creation of theljstaff to post in news and official communities eliminates blame being pinned on any one person for a news post that goes over like a lead balloon, but also removes some of the personal element from LJ management. The inability to associate a particular manager with a decision leads to a certain feeling of insecurity in the userbase due to the loss of transparency. The feeling of personal interaction is an advantage in the age of large and faceless corporations, but does have to be carefully maintained by actual interaction. The shift away from personal involvement has lately been mitigated somewhat by tupshin's dynamic presence in the development team, and marta's role in user interaction, among others.


History

It's necessary to be conversant with the history of LiveJournal's users versus its administration if you are going to be administrating the site. It's not as complex as, say, walking into the Middle East and expecting to make peace between the embattled warring factions there, because you're dealing with generations of ingrained distrust and hatred. However, it's not going to be as simple as deciding to make a fresh start. The collective memory of LiveJournal users is incredibly long. You'll be making a relatively fresh start with new users who are not aware of the site's history, and users who are aware that there is a new administration in town and have decided to give the new administration the benefit of the doubt. However, other users either may not be aware that this is new administration, may simply not care, or may be swept along in the current momentum of the opinions of their friends who are already established users.

Users who have been around since the early days (or have been through enough time to have a rosy view of the past) hold bradfitz in high (if highly questioned) esteem. However, the more things change, the more they stay the same. The same issues are being raised in news comments currently as they were six years ago, forever and ever amen.

The userbase has a remarkable talent for analysis and communication. This shows up especially in issues where they feel that proper attention has not been paid to an issue, or that they were wronged, or that they have been lied to. Unfortunately, this talent for analysis is not always correct, and is better for sniffing out the slightest hint of bullshit or snowjob than it is for clearing up controversy. No matter how thoroughly you saturate the site with official communities and links to them, the user gossip mechanism will always work faster, and bad and mistaken news travels faster than good news or the real story. Site rumors start off in the direction of conspiracy theory, and get bizarre fast.

Sometimes this means users will hear of an issue, and never hear how the issue was addressed or if it was resolved. This is exacerbated by a history of issues going un-addressed, and either un-resolved or resolved in a way that did not make the gossip rounds. This means that people will sometimes cherish a grudge over the lack of resolution of an issue that was resolved ten releases ago. Ideally, any problem that crops up should be addressed immediately upon it cropping up, and resolved as soon as practical, rather than waiting to be addressed once there is a resolution.

LiveJournal issues and users do not operate in business time, when decisions can take days, weeks, months, to come to a consensus and get signed off. LiveJournal operates in internet time, where a split-second can make the difference between an awesome comeback and a lame one, and two hours is more than enough time for any business to have noticed that something is wrong and put up a quick message saying that they're on it. Overnight is enough time for the story to get picked up by five tabloids, including Slashdot, the Valley Wag, and el Reg. An entire weekend, or an entire holiday weekend, is simply beyond the pale. If this is important to a user, how much more important should it be to an employee? Days and weeks of edits will save a foot in the mouth, but at the expense of catching issues while they are still relatively small.

A geek aphorism is to "Fix early; fix often," and I cannot recommend that strongly enough.


Major Controversial Events
  • The sale of LiveJournal to Six Apart

  • Ad-supported (Sponsored / Plus) Accounts, and the removal and subsequent re-animation of Basic accounts

  • Assorted notable incidents involving the Abuse Prevention Team

  • Assorted notable technical changes

  • Strikethrough

  • The sale of LiveJournal to SUP

  • The Advisory Board


Wikipedia has a Timeline of significant LiveJournal events; I'm not planning to cover all of them, however, the ones I do cover should provide a basic grounding, and many of the smaller events fit in the general categories of APT/administrative decisions and technical changes.

Almost any change whatsoever to the site or administration will cause a certain amount of dissatisfaction amongst the people who want it to stay the same. Some cause more dissatisfaction than most. While presentation of any change is key, even a well-presented change can fall badly flat if it comes too soon on the heels of a previous catastrophe, or fails to address the issues brought up in the previous catastrophe first. Even if there is no resolution, addressing the fact that there are still unresolved issues from a previous catastrophe demonstrates that the issues are not being ignored. You are dealing with a userbase full of people who will share incredibly intimate personal details with the general public on a website. These are not people who take kindly to feeling as if they've been ignored.


The sale of LiveJournal to Six Apart

As much as Brad made us sometimes want to bap him strongly with a clue-by-four, he was still our Brad, better-known and better-loved than MySpace's ubiquitous Tom. The news that LJ was being sold to another, already established, company, with little to no notice, was the cause of considerable alarm. While users have been inclined to trust Brad, or at least trust Brad to be Brad, Six Apart was treated as dangerous interlopers from the start.

Non-acceptance of Six Apart staff may have been in part because the people working day-to-day on LJ were not as well-known on LJ. The invading Six Apart personnel were not particularly seen to create a strong LiveJournal presence for themselves, preferring instead their established blogs offsite, using their existing products.

There was also no guarantee that Six Apart would keep the same practices and promises that Brad had, although it has always been the general expectation that any administration ought to follow through on the promises of the previous administration, even if they probably will not. The reception by the userbase was uneasy, although possibly accepted best by those people who loaded the news announcement in the few minutes where Brad had the ill-considered commentary that this was a financially necessary move.

Where Brad's changes could be interpreted in the best light by people feeling a personal loyalty to Brad, people did not collect the same feeling of personal connection to Six Apart. Even as the LiveJournal staff reassured the userbase that there was no "us" and "them" after working together, just "we", the userbase was not reassured, and still perceived a difference in ideas between what they thought of as original LiveJournal staff, and the new, conquering, Six Apart staff. The perceived origin attributed to the features was based on function -- features with a high bell/whistle/web 2.0 content were seen as coming from Six Apart, while low-impact improvements to existing features and any long-requested features were seen as inherently LiveJournal -- static, uncontroversial, practical.


Ad-supported (Sponsored / Plus) Accounts, and the removal and subsequent re-animation of Basic accounts

Ads were one of the first things predicted by the doomsayers when Brad sold LiveJournal to Six Apart. These people were immediately vindicated, albeit bitterly. Brad had formerly promised that there would be no ads on LJ, ever. The userbase took this as a strong promise when it was made, and immediately became concerned that this was going to lead to a lot of other changes that Brad would not have approved of. At this point, Permanent and Paid users were never supposed to see ads, and Basic and logged-out users would only see them on the journal spaces of Sponsored/Plus users.

Gradually, more and more ads crept in, in more places, shown to more people, just as predicted by those who were claiming that the presence of any ads at all was a slippery slope. Ads are a deep and abiding source of resentment for many users, even those who have accepted that it is financially necessary.

However, there was still the ability to create a Basic (ad-free) account. Then some genius decided that it was necessary to get rid of Basic (no-ads) accounts. This went over really badly. Really, really, really, really, horribly badly.

Jason Shellen honestly did not see what the problem with this was. After all, it was not impacting the experience of existing Basic users, and any account created before the cutoff date could switch to Basic. He did not reckon on the propensity of users to create new accounts (for projects, to commemorate events, in lieu of purchasing a rename, to better separate disparate parts of a complex life, and so much more), nor the principle of the thing.

Look, when the original owner of the site loathes ads with all his being, and the userbase generally agrees with him, and then ad-supported accounts are instituted because it's financially necessary, and the userbase loathes it, taking away ad-free accounts is going to be a really bad decision from a public relations standpoint. This would be enough to get even a pretty rational userbase angry. Had it been presented as a question of "Would you prefer ads, or no ads at all ... on no journals?" a rational userbase would have to admit that the priority is keeping the lights on in the data center. Had it been presented in this fashion, the userbase would have likely presented a wealth of alternative fund-raising ideas, and perhaps even embraced ads.

However, this was presented in a way that still leaves me boggling. I understand LJ pretty well, I think, perhaps too well, in ways a manager might never manage. So I decided to get an outside opinion. I have this best friend. He avoids LJ as completely as possible. I described the ad-related history in under a minute, and asked him how he would present the decision to disable the creation of no-ad journals to the userbase. "In a concrete bunker?" he guessed. Wise man. (Smartass, too.)

"Let's try spun as an improvement to the account signup process," I said. "Streamlining it, because it was too confusing, making fewer options."

Neither my friend nor I were able to find a reaction to the announcement that was not profane in some fashion. A sadder but wiser Jason Shellen left within two weeks of the resulting hail of quite justified negative feedback.

The ability to create Basic accounts was later restored, albeit with an ad presence for logged-out visitors viewing all Basic accounts, instead of their prior completely ad-free status, but the original decision and the manner of its presentation have not been forgotten.


Assorted notable incidents involving the Abuse Prevention Team, including the nipple/breastfeeding incident

Every time the Abuse Prevention Team makes a controversial decision, carries out a controversial decision made by people further up the management chain, or riles up a particularly touchy or well-connected user or group of users, cat macros fly. The issue is remembered by people who are directly affected as well as those indirectly affected, provided it's an issue close enough to their heart. This issue will return in the comments of news posts in the future, either until it is resolved to the satisfaction of those users, or even after, so it is not forgotten.

In some cases, the Abuse Prevention Team has appeared to be working independently of the administrative team, without their backing. This causes paranoia and woefulness in the userbase and most likely in the team as well. A (perceived or real) team of vigilantes with the power to suspend is not good for any site. An internal perception of not having the backing of the administration would assuredly cause morale problems in the team. While from my perspective as a volunteer it is certain that they have oversight that is not apparent to the casual user, the casual user does not have this reassurance. In other cases, fractions of the userbase have perceived or feared that site administrative changes, such as changes in ad sponsorship, could cause site administration to wield the Abuse Prevention Team at targets that do not appeal to an ad sponsor, against precedent and possibly against internally documented Abuse Prevention Team policy.

The Abuse Prevention Team is largely not trusted by the userbase, as they are the executioners, even if they are not the executive, legislature, judge, or jury. They must be seen to act consistently at all times, however, this is complicated by the requirement of privacy for all parties involved in an Abuse incident, by all situations being different, by people misunderstanding or incompletely understanding LiveJournal's Terms of Service, by people who have violated LiveJournal's Terms of Service misrepresenting the situation when complaining in public, and possibly even by policy changes that are not publicly documented. There are obvious problems with making the workings of the Abuse Prevention Team more public.

It is vital to maintain open and helpful communication between the Abuse Prevention Team and the administrative team, as the Abuse Prevention Team has valuable insight on the actual views and reactions of the userbase, and it would surely be a bad idea for the Abuse Prevention Team and the administrative team to be unaware of what the other is doing. (As I am not a member of either, I have precious few concrete examples of times this did not happen and Bad Things resulted. Please consult your local Head of Abuse if you would like more information on the ways this can go badly wrong.)

There is very little public buzz about the positive actions of the Abuse Prevention Team. Even actions such as the Abuse Prevention Team attempting to track down and contact the local emergency services of users who are reported as likely to be in immediate physical danger (attempted suicide, voice posts interrupted by violence, sudden illness) for welfare checks is either very little known, or seen in the worst possible light (how dare a web service attempt to interfere with someone's personal choice to kill themselves, and/or send emergency services needlessly), despite the Emergency Contact Info First Post meme.

If it is ever necessary for the Abuse Prevention Team to go against any policy or precedent that has been established for them, either in the Terms of Service, in the interpretation of the Terms of Service, or in any other way, someone must be accountable for that change. Not just privately accountable. If an action was dictated by management, there had god-damn-better be a public statement by that part of management that the actions of the Abuse Prevention Team were dictated by them, and for good and necessary reasons. If it would not be wise for that member of management to take ownership of that action, then that action had better be reconsidered.


Assorted notable technical changes involving inadequate notice and insufficient opt-out options

No matter how absolutely awesome a particular change is, you are guaranteed that someone is going to really, really loathe it. No matter how awesome and bulletproof the entire team of developers thinks something is, both ordinary users and people interested in exploiting it will get more creative. There is also a significantly vocal faction of users who would really prefer that any and all bells and whistles never be allowed to cause any ruckus on their end of LiveJournal, and would like not just an opt-out from each and every new bell and whistle, but a permanent opt-out for all bells and whistles, ever. Yet other users have grand desires for bells and whistles on such a vast or risky scale that, just as bagpipes are inevitably shared with one's neighbors, the effects of these bells and whistles might be too far-reaching to implement safely.

Any change must both be technically feasible, and in line with the gestalt of LiveJournal so that it would not put a crimp in the style of any common use. (Remember, everyone uses their LJ differently.) Here are some general classes of changes, and some of the issues that will arise from them.

  • User Interface Changes

    Any change to how users view the www.livejournal.com/ areas of the site will be controversial. Period. Even if it's the greatest thing since pre-buttered cheesy toast, someone is going to be allergic to wheat, dairy products, and on a low-sodium diet. Take this as a given for all changes. The existing developers and user-interaction specialists will have practices in place for being able to tell what an acceptable tradeoff is.

    Any time a new site scheme is put out, this is going to be notable. Any time an old site scheme is retired, this is going to be especially notable. Include a link to the master override site scheme changer any time when announcing the retirement of a scheme, so the people who love it won't hate you forever.

  • User Profile Changes

    These are a fun grey area, because they have historically been considered part of the www.livejournal.com/ area, although they are in fact under the user's subdomain. This means that users have more of a feeling of ownership over the profile than other, non-subdomain areas.

  • Journal Display

    Anything that affects the way a personal journal displays for the owner had better be carefully thought through. People get really attached to their journal styles, and spend many hours getting things just exactly right. The esteemed rahaeli joined Support to learn more about styles. If a change knocks some alignment a millimeter out of place, people will be upset.

    While styles may come and go, anything that messes with a user's actual journal entry is cause for outrage. The Snap.com graphical previews were an example of this -- messy javascript was being inserted into the text of a journal entry when the user did not put it there. To people who are sensitive about tampering with their entries, this was equivalent to having a "friend" scribble on them with permanent marker at best, and having the "friend" give them an unwanted tattoo on a very public part of the body at worst. Any proposed change that involves altering the display of the content of a journal entry (markup too, not just words) should be treated with caution. Any change that proposes to do so by default, to the journal owner, in a way that does not have a strong opt-out option, should be greeted with a warm and friendly trout across the face.

  • User-Initiated Display Changes to Others' Journals and Entries

    Some people are really attached to the idea of their content appearing in the exact format that they have written it in, and can get grumpy if someone else alters the display of it, even if it's for that person's own reading convenience. The idea that other people could flag a journal entry as containing adult content and the adult content status of that entry could then be changed by someone who was not them is anathema to these people, and not especially comfortable to others.

    Some people should not experience, do not like to experience, have difficulty reading, or physically or technically cannot handle some of the gawdawful things that people do to their journal styles and entries. These people need things like ?style=mine, ?format=light, content screening, login-based preferences over cookie-based preferences, and strong, persistent but reversible control of what is displayed to them. A subset of this faction still actively loathes the navigation bar, as it is clearly control-content, rather than user-content, but can be forced upon them by someone who wants to display it to all visitors to their journal.

    These two factions will always be at war, however, it is possible for someone to be in both camps on different issues, and in the case of preference over accessibility, accessibility should always win.

  • Easy Subscription vs. Stalking

    Any time any event, no matter how public, is converted into something that can be tracked with an LJ-side subscription, people will get upset about how it is a tool for stalkers. Have fun with that.

    Practically speaking, many things should come with the ability to make events private or protected, even if it would be really awesome to subscribe to for socially-networking-minded users. Privacy-concerned users are more worried about anonymous people or non-friends subscribing to events than their friends subscribing to events, in general.

    Particularly privacy-minded people freak out if people are seen to be watching something, even if it is private or otherwise locked, although this is one of the things that may have no actual solution other than vigorous user education. Unfortunately, LJ-side security events such as anything that displays a secured entry to someone who was not meant to see it gives credence to these fears.

  • Journal Owner Access to Visitor Data

    This is the other side of the privacy coin. Any tool giving journal owners more information about their visitors is going to raise an uproar among people who would prefer to not have their non-commenting surfing in arbitrary journals detected by the journal owner, for reasons both good and bad. Some of these people are aware that a server admin would have access to this sort of data. Some of these people are not yet that aware of how websites work, and think that one can actually be truly anonymous on the internet.

  • New and Inventive Means of Inter-User Communication

    Any means of inter-user communication brings with it the ability to be abused by trolls and other forms of lowlife such as spammers. The lower the barrier to communication, the more likely it will be used for spamming; however, no barrier, no matter how high, should be considered troll-safe, because there are some pretty determined trolls out there, and the people who have a devoted following of trolls will be sure to think of the ways that any given feature could be abused. (If they fail to think of it, the trolls will, and then the targets will report it in the hopes that LJ will fix it.)

    Any means of user communication should have an off-switch for the recipient. Any means of user communication should respect a ban. Any means of user communication that can, should respect the basic security settings of LJ, at a minimum private (none), friends-only, and registered users only (non-anonymous). Ideally, it would also include custom friends groups.

    Any means of user communication that allows a user who has been banned from a given journal to contact that user should be considered a security hole.

  • Any feature that has been previously seen on MySpace or Facebook

    No matter how awesome a feature actually is, if LJ does not have it and MySpace or Facebook had it first, someone will complain that the management or developers are attempting to turn LiveJournal into one of the above. In general, any feature that is primarily intended for social networking should also be able to be deactivated or ignored by someone who is not on LJ to collect the whole set of the people who they may have once punched in kindergarten.

  • Ad-Related Changes

    Any change that is primarily to benefit an advertiser or other source of external (not user-paid) revenue is going to create an uproar amongst the users that it inconveniences, especially if it is billed as a user improvement. Any change that is primarily to benefit a partner, even if it is not a partner who generates revenue, will be treated the same. See: Snap.com.

    A site that had ads from the get-go would not face this kind of issue. As LJ was created ad-free, any change in the direction of favoring advertisers over users will be met with open hostility. Don't look at me for sympathy either.

  • Administrative control of user content

    Content flagging. Interests censorship. (To a lesser degree, controls on content of default userpics, although this is a pure administrative issue, not technical.) Any time a new tool is created to allow administrative control or censorship of any user-generated content, there is a strong reaction from people who are perfectly willing to self-police, but resent any form of control imposed on them from an administration that they do not trust or particularly like.

  • Ass-Covering for Prior User Relations Fail

    If you're ever tempted to make a technical change to remove access or make data less available as the result of a dumb move you made that was called out using some of LJ's features ... don't. Censored interests and boldthrough, both of which have since been removed, are shining examples, although the shine from those can be better likened to the gloss measurement of polished turds. Please see your local damage control experts, and hope that they take mercy upon your soul. Generally, effective damage control to a technical crowd will involve owning up that you made a mistake, and detailing what is being done to a) fix it, and b) make sure it doesn't happen like that again. This may be counter-intuitive to catlike marketing instincts that insist that no, really, you meant to do that, but as much as geeks and LiveJournal in general love cats, they like their cats as is cats, not as is managers.

    Run any proposed technical changes at all that are planned for the near future of a user relations fail past an expert in userbase reaction. No matter if it was something that has been planned for three years, if it is released in close connection to user relations fail, it will be seen as connected and a reaction on the part of LJ to that failure. You may not want to connect these two things in the eyes of the userbase. If it is necessary to roll this out right then regardless of the connection, consult an expert in userbase reaction anyway, in addition to whatever spin doctor you're about to consult.

The current technical team has learned the hard way that no changes should be made right before people will become unavailable, and that fixes should be applied early and often. Some marvelous technical creations have come from fixing the problems inherent in LJ, and it's exciting to the developers to find these fixes, and it's exciting to technically-inclined users to see so much of the nuts and bolts of what makes the site work.


Strikethrough/Boldthrough

Any service that allows user interaction will inevitably gather its share of creeps. This is a given. The creeps on any given service will start out at people that no one really cares to be around (but who aren't really doing anything wrong) and go all the way up to people who are genuinely doing things in real life that are so vile and loathsome that they have no business on any internet service, ever. It is the job of the abuse prevention department of any service to separate the people who are unpopular but doing nothing wrong from the people engaged in wrongdoing upon the service.

The LiveJournal event known as "Strikethrough" is the result of a massive, clustered, and cascading failure to accurately tell the difference between unpopularity and actual wrongdoing. (This is also referred to as "Boldthrough", after technical changes were made to attempt to dampen the hysteria. That was a bad move that did not work.) barakb25 featured prominently in the chain of fail, although he appears to have spun his news appearances to lay the blame at the feet of anyone but himself, despite having been advised in exhaustive detail at many points, but not choosing to take the advice that would have stopped the clustered failure from failing further.

There are a couple LiveJournal-specific things that you are going to have to understand before you even try to wrap your head around Strikethrough and the attendant festivities.

First, the interests list. While it may originally have been intended to represent things people like, and the suggested format is that things for the interests list fit in the sentence "I like ___", the presence of something on someone's interests list does not automatically mean "I like ___." Rather, in context of any given journal, it means closer to "I like discussing and/or reading discussions of ____", or even, "I would like to be found by other people searching for ____ in LiveJournal's interests system."

Unfortunately, there was a disconnect in management's understanding of the feature and actual users' use of the feature. This was compounded by ignoring the advice of those who knew better.

Second, not all journals represent actual human beings. Perhaps the majority of journals are of human beings. However, some journals are complete public works of fiction, with LJ as a medium. Some of these works of fiction are kind enough to state publicly in their journal that they are, in fact, not real, and directions for how one can reach the author (or player, in case of an LJ-based role-playing game). Journals representing fictional human beings who are very vile human beings do not write under the same kind of restraint that an actual human being would write as. This includes their interests. Most real people who like beating people up are smart enough to not go online and state in a public and indexed forum, "I like to beat people up."

Third, LJ has become home to an online community of predominantly women who have open and frank discussions that involve unabashed sexuality. It is an online "safe space" that, regardless of its internal politics, is a widespread community drawn together by a common love for the creative arts and a wholesomely hedonistic wish to share resources in a common pool for the private pleasure of each, and defended vigorously against outsiders. In the relatively anonymous environment of LJ, women can safely explore fictional scenarios that they find scorching-hot, without fearing real-life retribution or exploitation for their exploration of their sexuality. Women who might feel uncomfortable visiting a sex shop or renting an adult movie (one that might not be suited to their tastes, anyway) participate in public and cheerful exchanges of customized porn that is in its own way as hyperspecialized and practiced as the adult film industry. Unlike the adult film industry, there is no budget, no actors, just a lot of people writing, drawing, painting, and then sharing. This is a gift economy where "bake you cookies and write you porn" is a standard gesture of comfort for someone who is going through a tough patch in their offline lives.

The community is largely fan-based, and highly networked, adapting to the LJ environment by creating communities, creating separate journals devoted to their life as it affects the community (as opposed to their regular online life or offline life), linking communities together by affiliations posted in the profile, linking together by specialized interest keywords, and integrating off-LJ elements for things that LJ is just not built to handle.

Some of the vast quantity of porn that is produced runs into some areas that would be exceptionally sketchy if someone wanted this in real life. This, however, is a non-issue for any member of the community who can successfully separate fantasy from reality. Prior to Strikethrough, no one thought twice about including any of the specialized keywords related to their happy erotica community interests into their searchable LJ interests, because everyone who would be searching for other users interested in that same keyword would be part of the community. Any member of the community would take as a given all the assumptions that the community held when viewing the interest keywords. Right? Right?

Fourth, Abuse had historically not done several things that were done during Strikethrough. First, to preserve LJ's common carrier status, Abuse did not seek out content that was potentially violating the Terms of Service. Second, upon receiving a report of violating content, before making any suspension, Abuse would carefully review the matter, and dismiss the report if there was no violation found. However, even prior to this point, Abuse was feared and mistrusted by the userbase in general.


Prior to the mass suspensions, there had been rumors floating around the fan community that there was going to be a crackdown based on interests. People who were paranoid or merely cautious heeded the warnings, removed interests from their profile, locked down their journals, and passed on the warnings.

The fact that the Abuse Prevention Team was forced to suspend journals found through a search rather than an individual report with evidence condemning each journal, was forced to suspend solely on the strength of items entered into the interests area, and was not given the opportunity to individually review those journals for evidence of wrongdoing above and beyond listing a sketchy interest, was a violation of the already-shaky trust that the userbase had for the Abuse Prevention Team.

It was also a profound and knowing betrayal by the management who ordered this action: against the people who had directly advised that this not happen and cited the reasons that it should not, against any member of the Abuse Prevention Team who was forced to enact suspensions against all precedent and their better judgment, and against the entire userbase.

After the suspensions started hitting, the entire fan community of LJ got into a panic state comparable to the panic state created by a massive, unplanned, persisting outage of LJ. Unlike an outage, LJ was up and running, so rumors could spread faster. Also unlike an outage, there was no authoritative source of information or comfort (such as the Status page). By the time any official statement was made, fears were running rampant that this was the beginning of a new era of censorship and control of LJ by external special interests, and any misstep, no matter how slight, could lead to the irrevocable destruction of what was in some cases years of creative work and hundreds of socially networked contacts.

After Strikethrough, some nameless genius decided that the best way to treat suspended or deleted users was not to show their usernames struck through, but still as a clickable link, but as a bolded, but not linked username. This made it more difficult to access the profile of a user that had been deactivated to tell whether they had deleted themselves, or whether they had been suspended. This led to more paranoia and hard feelings.

There were "only a few" journals suspended as a result of this action. However, it is not fair to say that they were the only people who were affected, as the suspended journals had friends, and people who had no reason to be suspended were nonetheless fearing that they, too, would be suspended. This set an alarming precedent. If some special interest could put pressure on the administration to remove journals that had an interest in content that was fictional, and legal, although edgy, what if another special interest put pressure on the administration over other legal but edgy content? We had thought that something like this could never happen, but then it had. That meant it could happen again.

The community fandom_counts has at the time of this writing 34966 members. It was created on May 30th, 2007 to attempt to provide a rollcall and count of the people who self-identify as fans. By June 3rd, 2007, it had a membership of 32,000. There are people who are not part of the community who identify as fans, and I am among them. Not all of the members of this community would have been affected by Strikethrough, but this provides a better sense of scale than the numbers of the suspended only. The reinstatement of struck-through but clickable deleted/suspended usernames after SUP's advent was welcomed, but in no way mitigated the effects of the event.

The exact chain of decisions leading to Strikethrough is still not widely or at all known, so there is little to no faith among many affected users that it will not happen again someday. The direct result of this is a large shift in activity from LiveJournal to other venues. Some people have moved entirely. Some who have mostly moved to other venues remain active on LiveJournal, sometimes getting involved in current issues. You would be hard-pressed to find people more angry than those who feel that they were persecuted and forced from the community.

I did not believe the rumors prior to Strikethrough, since they were so obviously hysterical and stupid, since they went against all precedent and sanity. I believed, with the insight that I had gained from volunteering, that there was no way that the Abuse Team that I knew and trusted could possibly do something so profoundly stupid and punitive. I knew that there were members of the fan community on the Abuse Team. I knew that all the team members that I personally knew were good and thoughtful people, and would not take ill-considered action. I have never doubted or hated my friends, but I will never forget or forgive the betrayal that led to the actions they were forced into.


The sale of LiveJournal to SUP

LiveJournal's sale to SUP came at a point in time when the current Six Apart-headed directorship had made enough blunders to permanently alienate a significant chunk of users, but after the departure of the key player in the Strikethrough imbroglio.

SUP had already gained notoriety among Russian users who wished to opt out of SUP services.

Reception was very mixed. On the one hand, this was a Russian company that might or might not do worse to LJ, and it looked like Six Apart was ditching LJ like a redheaded stepchild. On the other hand, some of the same daily operations staff would be staying, which could provide either stability, or continuity of bad decisions, or both. On the gripping hand, this was a company that was not Six Apart, who were largely blamed for most of the random crazy and unpopular decisions. Some people had already been lost, but some people who had been on the verge of leaving stuck around to see what would happen. Some people who really did not like SUP may have had this as their final straw.


The Advisory Board

So LJ created an Advisory Board to be seen listening to in the absence of the previous administration's appearance of doing anything that resembled listening, and promptly managed to not consult it when making drastic changes, and ignore the advice that was received, not even acknowledging that the Advisory Board brought up issues X and Y, but LJ did N anyway because of Z. Then there were the elections, which quickly turned into a trollfest as hotly contested as the 2004 US elections on LJ, in a close race between a known instigator with a large trollish following, a member of the Abuse Prevention Team, and a very competent regular user as a favored third. The elected board member received death threats from a party or parties unknown. And then, after all the hype, while there was a community ready and waiting for the representative to set up office, the issue was frozen for externally and perhaps internally unknown reasons, leading pretty much everyone to conclude that it was all an over-hyped scam to make the userbase think that LJ was listening, and causing the elected representative to have to lock down completely and not show her userpic in public lest she be jumped upon and accosted for not fulfilling her duties.


No One True Way

People use their LiveJournals in all sorts of bizarre and brilliant different ways. There are, I have been assured, internal studies on how people do the stuff they do with their LJ. It can be a completely private journal, a personal publishing platform, an interaction hub between you and your admiring fans, a salon with a hundred different unrelated conversations between an array of people, a closed meeting place for just a handful of friends, an open letter to the world, an update point for far-flung associates, a work of fiction, an experimental art project, a comic, even an adventure game.

But the point is, the way that you use your LJ is not the way that everyone uses it, and using that assumption in making business decisions and decisions about what the userbase will and will not put up with is doomed to be wrong. Keep tabs on how people actually are using LJ, and make sanity checks with the people who are in touch with the userbase to see what any change will fuck up.


On Bullshit

Please don't attempt to lie to, bullshit, snowjob, or blatantly spin LiveJournal users. People have done this with varying degrees of success, but in general it does not tend to go well, and rats get smelled out fast. If the primary purpose of a feature is fundraising, with some bells and whistles to make it desirable, call it like it is. If something is a development or financial issue, sure, put a good face on it, but if people are going to take it badly, acknowledge that, don't try to ignore that and pretend like everyone is going to be happy.

Brad's habit of blunt and developer-flavored speech was familiar and direct, and the contrast between that and smooth marketing speech was taken ... poorly. Brad provided competent technical sense. LiveJournal desperately needed people more inclined to business. However, the userbase is still not comfortable dealing with the results of cold hard financial calculating without knowing why things that are Not Fun are happening.

Virtual gifts were taken particularly poorly upon their introduction. They fill approximately the same niche as US-based schools selling flowers and candy to be delivered on Valentine's Day -- a fundraiser for whatever group is selling them, you could get pretty much the same things on your own cheaper, but people choose to buy them because it's a status symbol to be seen to get them and it helps support the good cause in question, namely the school. They were oversold at a particularly cranky time.

The removal of Basic accounts was a major, if not the major, spin incident. This did not go well at all.

Each instance of blatant spin that does not at least take dominant userbase feelings into consideration is marked down and noted by the people keeping track of these things, and future announcements are trusted that much less.


Technical Transparency, Whimsy, and General Geekery

Geeks are a tremendously whimsical lot, and require a certain sense of play in order to stay happy. LiveJournal appeals to the geek crowd not because it offers the amount of control that the average geek needs. In fact, it's far less flexible in some ways than other blogging tools to be installed on your own server. However, LiveJournal is open source, was developed communally, maintains an active suggestions community to take input from the peanut gallery, maintains an active feedback line monitored by real people, and has things like Frank the Goat, the named clusters, the 404 pages, and a thousand little touches around the site that say that this place was not set up prefab by some place that doesn't want to let people see inside, it was built from the ground up by real live geeks, and is maintained by real live geeks, and you are given enough basic information about how things work that you too could hop on board and help keep LJ running. The support board has always been volunteer-powered. People have been hired from volunteer positions. There's a very real sense that you can get involved and make a difference. People who aren't quite technically inclined enough to run their own blogging installation, or don't care to go to the trouble, or who prefer a larger community over more control, are happy here.

People who aren't technical at all come to care about the little touches. Someone who had mail with Yahoo does not care what cluster their emails are stored on, so long as they can get to them without delay. People on LJ started out wanting to know what cluster they were on when LJ was having growing pains, but came to love (or loathe) their whimsically-named clusters, even people who didn't really care what a cluster was, but did care that they were on Cartman. (People got upset at the switch from a South Park naming scheme to a meat-based naming scheme, even though nothing about that switch would affect the technical operation of the site. Go figure.)

The more technical and fun details that can be made available to the general userbase without compromising any security, or revealing any internal business information, the better. Granted, this is not going to be front-page breaking news if there is now a blood/sausage cluster, but it is the sort of thing that will make people smile when it's included in a news post as a fun fact to round out an announcement.

We want to be for the users, by the users. LJ users want to know that we're being managed not by a bunch of suits who are going to take away our playground and make things Not Fun Anymore, but by people dig what we're doing here because they are users themselves, and will do their best to keep glass bottles out of the sand and unbroken, drinks in the cooler where they belong, and are not averse to kicking back, enjoying a Cold One, and watching the kids run around and scream and throw sand at each other and play.


Trust Your Experts.

When in doubt, review rahaeli's Painful Lessons I Have Learned About LJ. Read it. Love it. Learn it. Live it.
(tupshin should have a copy. Poke him.)


Us and Them

When you are tempted to throw it all in a ditch because you are being bombarded with cat macros left and right, your fellow employees are possibly not talking to you, your volunteers are screaming, and you are going to need the company to either pay for the blood pressure medication you did not need prior to this position, or the vodka you are drinking to help cope with said high blood pressure, it is time to take a step back. You may think now that the day will never come when it is this bad, but the day may well come, and on that day, it may be cold comfort to know it, but any job that involves relating to the userbase in general is a high-stress job that can be quite frakking thankless, and even when you are doing your very best, sometimes it does not seem like enough.

It's very easy to wind up in an embattled us vs. them mindset when the stress goes up. If you possibly can, see if you can take some time for yourself. If you can't get that, or even if you do, seek out someone else who has been through the same kind of firestorm and let them know how you're taking it. They may be taking it pretty hard too, but they will probably have a few words of wisdom, and at times like those, one hangs together rather than hanging apart.


Care and Feeding of your Volunteers

LiveJournal's technical support and several other departments are handled through some employees managing a network of volunteers. There are quite a few trusted and seasoned volunteers, and new volunteers, and come-and-go volunteers, and people who have volunteered in the past, can't make room in their lives to do so now, but still stick around. You can meet them on IRC if you'd like. This is also the hot place to be when anything interesting is going down on LJ. There are several main channels, and random back-channels get created based on current need.

Volunteer culture has its own slang and history.

In any issue where site management conflicts with the majority views of the userbase on an issue, the volunteer team is likely to be caught square in the middle. The volunteer team, given that they interact so closely with the userbase as a whole, is also likely to have a feel for what reactions might result from any given change, and what other things might be affected.


Us and Them Redux

It is tempting to look at the employees and long-term trusted volunteers, look at the userbase, look at the history of conflict, and huddle up with a small circle of like-minded people and wonder how it is that so many people can be so very wrongheaded to have gotten on the wrong side of whatever major controversy is going on. That kind of viewpoint can save your sanity if you're in a tight and embattled spot, but it will not have good results in the long term.

By all means, pick out your allies and the people you work with best, but never stop listening to the userbase. Turn off comment notifications when making controversial posts in official communities, by all means. Skim over anything with a font size more than one point larger than default, and anything with a cat macro. Get a minion or someone not as emotionally invested to sum it up for you. Take the time you need to work on something that doesn't require interaction.

But don't stop listening.

Blast scathing commentary in a secure back-channel, don't let them see that they hit a nerve, and figure out what nerve of theirs this hit, and how to not hit that one in the future. Figure out the genuine business need behind the technically appalling implementation request for what someone fondly believes to be a shiny new feature. Ask why. Ask what could be done better. Apologize when something got done wrong. Apologize at least that someone was hurt and upset, even if you can't apologize for the decision that led to them being hurt.

Listen to industry experts. Listen to your co-workers. Listen to your volunteers. Listen to your users. Listen to people who never have used your service and never will. And don't stop listening.


Exported to http://azurelunatic.dreamwidth.org/6155829.html; further comment here may be ignored.
Tags: lj
Comments for this post were disabled by the author