Log in

No account? Create an account
Lessons from the failure of Livejournal: when NOT to listen to your users - Arvind Narayanan's journal — LiveJournal [entries|archive|friends|userinfo]

Lessons from the failure of Livejournal: when NOT to listen to your users [Oct. 4th, 2008|03:41 pm]
Arvind Narayanan
[Tags|, , , ]

The intended audience for this post is developers building/running websites. I'm sure most LJ users are perfectly satisfied with it. That's my whole point.

In the last couple of months, there have been multiple developments that seem to indicate that the combination of a blogging platform/commenting system and social network is finally becoming widespread: Google added the "follow" feature to Blogger, Facebook switched to a feed-focused redesign, and WordPress acquired IntenseDebate, a commenting system.

It is surprising that these changes took so long: Google's move has been seen as a response to a threat from Facebook, while Facebook itself was pretty clearly reacting to the Twitter/Friendfeed threat on the horizon. Finally, WordPress was arguably forced into action by the rise of third party commenting systems including Disqus and JS-Kit.

What you might not know is that Livejournal had this winning combination of features 5 years ago. Further, they had a significant technical lead over the competition: the now-ubiquitous memcached, for instance, was written to run Livejournal. The site had an API for most of its profile data at a time when no one else did; it wasn't even called an API. Finally, they had a huge (for 2003) userbase.

Here's how they could have capitalized on all of that: put their syndication mechanism front-and-center, which lets you import external RSS/Atom feeds into your LJ friends page, thus making LJ a fully capable feed reader. This could have been followed by the step of releasing either a Disqus-like widget, or a browser plugin that would in addition have made LiveJournal the default commenting mechanism for the entire web. The opportunity was just mind-bogglingly huge.

But none of that happened. The site hung on to its design philosophy of being an island cut off from the rest of the Web, and paid the price. Founder Brad Fitzpatrick has moved on, traffic has been flat for years, and last year they sold to a Russian company in the hope of making something of their Russian userbase. The site is now a sad footnote in the history of Social Networking Services. How did they do it? By listening to their users.

Let me describe four different circumstances in which you shouldn't listen to your users, all of which apply to LJ, and which explains why LJ went the way it did.

Skewed userbase. Livejournal attacted social deviants of all kinds. (I'm not judging; that's why I got on here.) The design of the site — little or no google indexing, no pingbacks, no comment RSS feeds, strong privacy features including granular controls, etc. — made the site an island. If you weren't on there, you wouldn't know it existed. Of course, many users wanted it to be exactly this way.

This is a problem that every website faces when it tries to go mainstream: your early adopters will try to keep you from doing so. You need to have the balls to bite the bullet and go mainstream anyway.

Vocal minority. Even within the deviant userbase, there was one subcommunity that was even more intent on keeping the site the way it was: the people who relied on LJ to exchange illegal shit. Naturally, this minority was very vocal in providing feedback. LJ should have responded by cracking down on the illegal material a long time ago; instead, they made a feeble effort last year by when it was way too late.

Be warned: the users who express themselves are likely to display varying levels of self-selection bias, and aren't representative of your overall userbase, and more importantly, your potential userbase. The only way to avoid it is to actually sit down in person with random people, have them use your site and listen to what they have to say.

Mindless change-resistance. Some users hate it when they have to change any behavior they are accustomed to. Other users protest simply because they like protesting; it makes them feel like they have some power against The Man. There were mass protests when LJ wanted to stop charging money to create accounts. Such users are best ignored.

Facebook pulled off their recent site redesign admirably: they did bucket testing for a long time, ironed out most bugs, and then rolled it out slowly. In spite of it, the anti-redesign groups counted millions of members, at least 10 million aggregated. The important thing to note is that these users weren't threatening to leave the site; they just felt like expressing themselves and their hatred of change. Facebook did what they should, gave them a couple of weeks to get used to it.

Fundamental conflict of interest. Most of the time, your interests are aligned with those of your users: if they're happy, your site grows. Occasionally, you do have to face the fact that you want their money and they don't want to give it to you. When LJ started to put ads on free accounts, users screamed bloody murder. You just have to ask yourself: do we want to stay in business or not?

[User Picture]From: mcfnord
2008-10-04 09:45 pm (UTC)
Good points.
(Reply) (Thread)
From: thedragonbook
2008-10-04 10:16 pm (UTC)
Nicely written, but its easier to point out the reasons for failure after the failure and difficult to see in advance what steps will lead to failure. Of course, thats what chooses a winner.
(Reply) (Thread)
[User Picture]From: arvindn
2008-10-04 10:27 pm (UTC)
true. it's also possible that by doing the things i said, lj might have alienated their users *and* failed to go mainstream. but hey, no risk, no reward, right? :-)
(Reply) (Parent) (Thread)
From: thedragonbook
2008-10-04 11:29 pm (UTC)
I agree that your points are valid. I am just saying that LJ might have found it difficult to see these in advance.
(Reply) (Parent) (Thread)
[User Picture]From: skthewimp
2008-10-05 02:31 am (UTC)
excellent analysis. another big problem I had with LJ was that they made it very difficult for me to add special features to my page - things as simple as google analytics.

a couple of my favourite features on LJ were threaded commenting and friends list. and there was the beautiful way in which private posts was implemented. it's incredible that they had such rocking features ages back and didn't capitalize.

also they allowed only paid users to syndicate non-LJ feeds. so since i'm a free user i've to wait for someone to syndicate a particular feed before i can add that. maybe this prevented LJ from becoming a full-scale feed aggregator.
(Reply) (Thread)
[User Picture]From: arvindn
2008-10-05 02:37 am (UTC)
tell me about it! if you look on my twitter, i can't believe how primitive commenting still is in other systems (both blogger and wordpress.)

as for syndication: i've been meaning to write this rant for a long time, but in the comments to my previous post about my new blog, ravi wanted a syndicated feed, but couldn't create one, prompting me to finally write it :)

i'm certain this is what kept lj from becoming a feed aggregator. they just didn't want it to. they wanted it to be about maintaining their "culture" and keeping the rest of the world away. this was a conscious decision.
(Reply) (Parent) (Thread)
[User Picture]From: skthewimp
2008-10-07 04:45 am (UTC)
yeah when I moved away from LJ, I put a post requesting some friend with an LJ account to syndicate my new feed for the sake of LJ users. one problme with that is that people are leaving comments on the syndicated LJ feed - which I get no notification of!

i think integration of comments into feed aggregators is the next big thing. if this can be done in an efficient way, it'll surely up the volume of comments significantly.

and on a different note, I'm told that Wordpress 2.7 is coming out with threaded comments (just like LJ). right now though, i'm using a threaded comments plugin on my site. just can't adjust to the flat structure.
(Reply) (Parent) (Thread)
[User Picture]From: mskala
2008-10-05 01:06 pm (UTC)
Thanks for the link - I spotted this item in my referrer log.

I'm not sure you're correct that the illegal segment of fandom wanted Livejournal to be an island. One of the points I made in the article you linked to is that most people in the illegal segment of fandom either don't know, or refuse to admit, that the illegal segment of fandom exists. Even if members of fandom acknowledge that illegal fandom activities exist at all, most members of the illegal segment don't acknowledge that they are themselves part of it; it always has to be someone else, not oneself. There's a lot of evidence for lack of comprehension in the comments made about my article - both on my own site and on Livejournal at the time my article came out. The level of denial, and ignorance of the larger world, shown by some of those people was amazing.

My position was, "Look, if you're going to do this, you have to do it discreetly"; the opposing position was "It should be perfectly okay for us to do this loudly in public." It wasn't "we want to be protected from the outside world" but "there is no outside world relevant to this situation." Then your suggestion that Livejournal remained an island because illegal fandom wanted it to be an island, doesn't ring true. The isolation of Livejournal had allowed that segment to become important, but it wasn't treated as valuable by the beneficiaries. Illegal fandom, on the contrary, wanted endorsement of its own mainstream non-island (and non-illegal) status.

I think you're right, though, that Livejournal is an island and that it remained so deliberately. One reason may have been because of LJ management's reaction to the existence of illegal fandom. The people in charge certainly weren't in denial about the need to keep some of that material out of the public eye - there were comments to that effect on my article from current and former staff members. One insider even speculated that maybe I was an insider too, with information on management's motives (I wan't - it was just reasonable inferences from the observed public behaviour). I think another reason Livejournal was promoted as an island was simply the site operators' idea of what market they were serving: they thought they were building a "diary," and they thought "diaries" were by definition private. Listening to the many users (not fandom users, but ordinary "diary" users) who wanted extra "privacy" features and the impossible "everyone except you" and "only cool people" security levels, probably kept "Livejournal should be an island!" at the top of management's minds and prevented them from building the kinds of interaction with the outside world that you suggest could have saved Livejournal.
(Reply) (Thread)
[User Picture]From: arvindn
2008-10-05 05:32 pm (UTC)
That makes sense.

Let me modify that part of my argument, then: the minority that exchanged illegal fandom was loud in expressing themselves, and this might have caused LJ management to get a skewed perception of the size of this contingent and the importance of keeping them out of the public eye, as opposed to purging illegal content.

At any rate, I guess we agree that keeping LJ an island was a deliberate decision.
(Reply) (Parent) (Thread)
From: ifatree
2008-10-05 07:47 pm (UTC)
that's why i recently came back to lj, actually. :) there was never a feeling that you need EVERYONE you've ever met to friend you into their personal world. you can usually keep your lj life insulated from outsiders better than the later web services ever allowed (without making multiple accounts). that's why they're now failing among the "long tail" crowd and losing its page views to "joe's private message forum" (which is where all of this really stared anyways).

and, philosophy aside, i'm getting tired of round corners everywhere. web 2.0 can kiss my lilly white arse!

ps. your island is gett raiddited. ;)
(Reply) (Parent) (Thread)
[User Picture]From: arvindn
2008-10-05 07:50 pm (UTC)
That's why I put a giant sign at the top of the article saying

"The intended audience for this post is developers building/running websites."

Apparently you missed that.

It's great that LJ works for you. I'm trying to point out that the site could have made a shitload of money, and they didn't.
(Reply) (Parent) (Thread)
[User Picture]From: mskala
2008-10-07 01:52 pm (UTC)
the minority that exchanged illegal fandom

Minor point, but it looks from referrer volume like this posting is getting read by a lot of outsiders and it's worth getting the terminology right for people who don't know: the word "fandom" means the cultural group, i.e. the human beings. Fandom is not the material exchanged by that group. Cognate with words like "Christendom."
(Reply) (Parent) (Thread)
[User Picture]From: arvindn
2008-10-07 05:28 pm (UTC)
oh, thanks for pointing that out.

yeah, i posted it to news.ycombinator.com and it kinda spread from there. just curious.. how many hits did you get?
(Reply) (Parent) (Thread)
[User Picture]From: mskala
2008-10-07 06:02 pm (UTC)
About 750 hits through your posting in the last couple of days. That's a drop in the bucket to the roughly 10K hits when the article was first posted last Summer, but a big enough blip to remove a garment on my Site Traffic Strip-O-Meter.
(Reply) (Parent) (Thread)
From: worldserpent
2008-10-07 02:04 am (UTC)
Very interesting post, but are you sure it is as much that LJ listened to its users as much as the LJ creators failed to foresee the potential of the service? Or are you saying that because they focused on the users alone, their vision was therefore constrained so they couldn't think outside of the box of the current userbase?

Also, at the time, it might have been hard to foresee the demand that LJ wasn't fulfilling.

BTW, this series of posts on social media and ads may interest you: http://synecdochic.livejournal.com/238398.html?style=mine (links to the other parts are at the bottom of the post)
(Reply) (Thread)
[User Picture]From: arvindn
2008-10-07 05:33 pm (UTC)
well, those aren't entirely different things. lj didn't foresee their potential because of how focused they were on their current users. they put "maintaining the culture" over growing the userbase.

i don't buy the hard-to-foresee argument, because every website creator should always be asking themselves, "how do i take this mainstream?"
(Reply) (Parent) (Thread)
From: (Anonymous)
2009-09-15 05:25 am (UTC)
While you analysis of fandom being to blame for Livejournal's downfall is interesting, honestly, the reason why it failed was because of Brad. He just wasn't interested in taking it in any direction other than an island. I FREQUENTLY complained about the problem of LJ's walled garden, but my concerns were always on deaf ears.
(Reply) (Thread)
[User Picture]From: arvindn
2009-09-15 05:30 am (UTC)
that's interesting. thanks.

i don't mean to imply that fandom was the major reason for lj's downfall. it is one of many; i see it as half-way between a cause and a symptom.
(Reply) (Parent) (Thread)
From: ext_255925
2010-09-13 02:59 pm (UTC)

when not to heed customer advice

Hello Arvind,
I'm late to this party, but still found your two year old blog helpful and interesting.

I didn't know the sad story of LiveJournal's wasted opportunity. Reminds me of an open/closed debate we had at SV-PAL, and early free-net site in San Jose. Thanks for sharing, and I do like the list you put together.

I'm most interested in your 4th point regarding conflict of interest. In my experience, this is where an honest dialog between company and customer in a transparent forum of blog is very powerful. Take an example where customers don't like to see ads. A proactive community manager couple seed an open discussion of ad-supported free service vs. premium ad-free by paid subscription. A poll of other mechanism could be used to survey the customers on potential price points. (Of course, we'll know the data comes from a skewed sample and the rest.
Thanks again.
Marc in San Jose

(Reply) (Thread)