How do I know this?  I got a copy of the website analytics.

The bounce rate for LessWrong's home page is 60%!

To be clear: Over half the people who visit LessWrong are going away without even clicking anything.

Yet how many NEW visitors are there?  Almost half of the visitors are new!

Granted, new visitor statistics aren't perfect, but that's a LOT of people.

Simple math should tell us this:

If we got the bounce rate down around 30% (a reasonable rate for a good site) by making sure every visitor sees something awesome immediately, AND make sure that each visitor can quickly gauge how much they're going to relate to the community (assuming the new users are the right target audience), it would theoretically double the rate of growth, or more.  There's a multiplier effect if the bounce rate is improved: you get better placement in search engines.  Search engines get more users if they feel that the engine finds interesting content, not just relevant content.

It's been argued that it's possible that most of the bounces are returning visitors checking for new content.  Well if half the visitors to the site each month are new, and we did a wonderful job of showing them that LessWrong is awesome, then the amount of returning visitors could double each month.  We're getting a tiny, tiny fraction of that growth:

http://www.sitemeter.com/?a=stats&s=s18lesswrong&r=36

Why did I write you guys so much in the home page rewrites thread?  Because I am a web professional who works with web marketing professionals at my job and to me it was blatantly obvious that there's that much room for improvement in the growth of LessWrong.  Doing changes like the ones I suggested wouldn't even take long.  Because I like this site, and I knew it had potential to grow by leaps and bounds if somebody just paid a little bit of attention to real web marketing.  Because I was confused when I first found this site - I had no idea what it's about, or why it's awesome.  I closed the home page, myself.  Another friend mentioned LessWrong.  Curiosity perked up again.  I came back and read the about page.  That didn't make things clearer either.  I left again without going further.  Friends kept telling me it was awesome.  I came back one day and finally found an awesome article!  It took me three tries to figure out why you guys are awesome because the web marketing is so bad.  The new proposals, although they are well-meaning and it's obvious that John_Maxwell_IV cares about the site, are more of the same bad marketing.

I've been interested in web marketing for ten years.  It's a topic I've accumulated a lot of information about.  As I see it, the way these guys are going about this is totally counter-intuitive to web basic marketing principles.  They don't even seem to know how harsh users are the first time they see a new website.  They tend to just go away if it doesn't grab them in a few seconds.  They're like "well we will put interesting links in" but that's not how it works!  The links don't make the site interesting - the site has got to be interesting enough for users to click the links.  Thinking the links will make the site interesting is backward.  If you want to improve your bounce rate, your goal is to be awesome immediately in order to get the user to stay on the page long enough to want to click your link.  If it wasn't usually hard to get users to click links, we wouldn't track bounce rates. These guys know this particular group of users better than I do, but I know web marketing principles that they're not even seeing when pointed out.  To me, they seem to be totally unaware of the field of web marketing.  The numbers don't lie and they're saying there's huge room for improvement. 

If you want to grow, it's time to try something different.

Here's a thought:  There is a lot awesome content that's on this website.  We need to take what's awesome and make it in-your-face obvious.  I wrote a plan for how to quickly find the most effective awesome content (the website statistics will tell you which pages keep new visitors on them the longest), and how to use them to effectively get the attention of new users - copy the first paragraph from one of those pages, which was most likely constructed by a competent writer in a way that hooks people (if it's keeping them on the page then it's essentially proven to!) and place that as bait right on the front page.  (There is also a wrong way to do this.)  Then of course, the user needs to find out why the LessWrong community might be a place where they belong.  I shared ideas for that in "About us - Building Interest".

Don't let's assume that growth is going to be good.  You're going to get more internet trolls, more spam, (there's a way to control spam which I would be happy to share) and more newbies who don't know what they're doing (I provided some suggestions to help get them on track quickly, preventing annoyance for both you and them).  There will be people with new ideas, but if the wrong audience is targeted... well.  We'd better choose what audience to target.  I saw an internet forum take off once - it seemed to be growing slowly, until we looked at the curve and saw that it was exponential.  That of course quickly turned to a dazzling exponential curve.  Suddenly the new users outnumbered the old ones.  That could happen here -- even if we do nothing.  YOU can get involved.  YOU can influence who to target.  They're taking suggestions on rewrites right now.  Go to the thread.  I invite brutal honesty on everything I wrote there.  Or pick my brain, if you'd prefer.

What do you want, LessWrong?  Do you want to grow optimally?  Who do you want to see showing up?

 

New Comment
107 comments, sorted by Click to highlight new comments since: Today at 9:32 AM
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

Discussion forum growth is not quite the same as general website growth. Having forums grow a lot while maintaining the culture that drew the initial contributors there is still something of an unsolved problem, and fast enough growth can kill forum cultures dead through the eternal September effect. It basically happens at the level of comment-response pairs. Things are good as long as most interactions have at least one side familiar with existing site culture, but once you start getting outside users talking with other outside users in volume, there's not much left maintaining the older culture. And if the outside users come from The Internet In General, the new forum culture is going to end up looking like The Internet In General.

Active and clueful moderation can help, but that requires moderators who can spend a lot of time daily doing active and clueful moderation.

Some forums make things work a bit better by managing to make their content interesting enough that people are willing to pay $5 or $10 for making a new account, and then asking that. Drive-by trolling and spam becomes harder, but regular users with various issues can still make lots of work for active and clueful m... (read more)

Things are good as long as most interactions have at least one side familiar with existing site culture, but once you start getting outside users talking with other outside users in volume, there's not much left maintaining the older culture.

Worse yet, the new users may comply with the culture in form but not in spirit. In the concrete case of LW, this means new users who are polite and non-confrontational, familiar with the common topics and the material covered in the classic OB/LW articles, making appeals to all the right principles of epistemology and logic, etc., etc., but who nevertheless lack the ability and commitment for truly unbiased and open-minded discussion at the level that used to be the standard. I think this is indeed what has been happening, and I don't see any way an open-access forum could prevent this course of events from taking place eventually.

(It's hard to make a point like this without sounding arrogant and conceited, so I should add that in retrospect, I believe that when I joined LW, at the time it probably caused a net lowering of its standards, which were higher back then.)

When I first joined, I barely commented, because I felt it would inexcusably lower the average comment quality. I still refrain on topics which I’m interested, but not competent in; but for the last 18 months or so I’ve felt more comfortable with the vector my comments apply to the site average.

Separately, I generally agree with Will Newsome about the high quality of your contributions.

8negamuhia12y
This being my first comment, I found LW through Google Plus, and my first reading was So You Want To Save The World, which took me a couple of weeks due to the prerequisite papers, plus a ton of other stuff i was reading at the time. One might say, that that might have been an unnecessarily daunting introduction to a community - any community - but I'm glad of the way it happened because I'd never have gotten as interested in cogsci and "A-grade+ rationality" as i have been ever since. That was just before Christmas last year. Now, I've read most of the sequences, and even used some of the tools recommended here, but constantly struggle to find a good place to start making contributions to discussions. So it's taken me roughly 8 months of lurking to do this post. Does everyone do this, or am I a unique point in newb-space?
9OnTheOtherHandle12y
This was my biggest fear in joining this community; I did not want to be the clueless kid who forced the grown-ups to humor her. I'm quite new, so I don't know how accurate this is, but I must say that the oldest comments on Overcoming Bias (around 2007-2008) were actually of a much lower quality than the 2009-2010 comments on Less Wrong. Maybe it was because OB, being sponsored by Oxford, had a higher rate of drive-by trolling? As far as I can tell, Less Wrong is still intimidating enough to deter well-meaning newcomers from saying too much, although nothing but boredom kills trolls. Still, anything that increases traffic will pull quality of discourse toward the mean, unless we can somehow accomplish the miraculous task of bringing newcomers up to the LW mean. Is there a way to quickly bring people up to speed on the spirit of the law, rather than just the letter of the law? Is there a way to make people want to display their curiosity, their ability to admit mistakes, their thick-skinned, careful consideration of criticism? Saying “I was wrong,” “Thanks for explaining my mistake,” “I’m confused,” etc is a sign of status among more established Less Wrong members (admittedly, the mistakes made shouldn’t be too elementary, and it definitely helps if you’re high status already). How do we get newcomers to care about this particular measure of status from the get-go? Make It Stick gives us a simple heuristic for getting ideas to burrow in people's heads and inspire them to action: Simple, Unexpected, Concrete, Credible, Emotional, Story. This says to me that we need to distill the spirit of truth-seeking and truth-speaking into something with the emotional resonance of a proverb - while avoiding the creepy cult vibe that the overtly religious phrasing of the Sequences tends to bring about. We then need to display this short credo prominently on the front page. Our current main page is a flood of information. I think it's quite elegantly designed, but the problem
4TGM12y
Have you adjusted for the likely event that you have become more rational, and what you have actually observed may have been LW becoming at a lower level relative to you, whilst staying relatively flat or even improving?
4Risto_Saarelma12y
I was thinking the same, but I don't know how to convince myself that there has been a clear decline or a markedly improved past state. My take is that LW has always had problems with user-generated content measuring up to the blog content core of the site. I don't think I followed the forum that closely during the first year or so after the OB age, and the conversation in the pre-LW OB comment system often seemed pretty dismal to me.
0beoShaffer12y
This sounds plausible, but it also sounds like something I'd expect at least one old timer to feel even if it isn't true. Furthermore, I've read many of the comments on the sequences, and they don't seem to support this hypothesis. Can you provide any specific evidence?

Its also worth noting that one of Hacker News' main strategies is flooding the page with advanced math and other highly technical topics whenever it gets serious mainstream attention.

6dbaupp12y
And disabling new account creation for the duration of the attention. (All that said though, HN regulars seem to often complain about the decline of Hacker News, how it is turning in to Reddit. (I'm in no position to judge if these complains are valid.))

I already have a solution. To reduce the bounce rate on Less Wrong, we just remove all the math and rationality stuff that requires high IQ and metacognition, and replace all that with lolcats.

7Viliam_Bur12y
Just don't forget to do A/B testing! More seriously: the goal of the main page should be to give honest image of the website. The main page is optimal when people like the homepage if and only if they would like participating in LessWrong. We don't have to attract everyone. We should just make sure that the main page does not send away people who would have stayed if they were exposed to some other LW stuff instead.
1A1987dM12y
Yes, with the caveat that a person who is captivated by the front page eventually deciding that LW is not for them is not as big a deal as a person who would like participating walking away upon seeing the front page: the former just wastes a couple of hours and is disappointed, the latter misses out the chance of finding valuable stuff and becoming a valuable contributor.
0Wrongnesslessness12y
That's a good point. However, I think there is not much we can do about it by refining the main page. More precisely, I doubt that even a remotely interested in rationality and intelligent person can leave "a community blog devoted to refining the art of human rationality" without at least taking a look at some of the blog posts, irrespective of the contents of the main page itself. We all know examples of internet sites with poor design, but great information content. So the question of refining the main page, I think, really comes down to selecting the right articles for the Recent Promoted Articles and Featured Articles sections. The rest is already there.
1Epiphany12y
Problem 1: If the person has already done work to refine their skill with rationality on their own, they may figure the blog is likely to be more of the same old stuff. LessWrong would love to have more of that type of person, don't you think? They're not going to be interested in the topic of rationality presented plainly all by itself. You have to immediately prove to them that the particular collection of information is fresh and exciting, or they slot you away in the "already read about it" category. Problem 2: A random post isn't necessarily going to be a good post. I did random post several times before I found one that excited me and made me realize that LW is awesome, and the ONLY reason why I kept trying was because so many friends had referred me to LW that I was starting to think I was missing something. Problem 3: A lot of users are busy and don't take their time with new sites. They usually will not read an entire front page. You have a few seconds to interest them. You either put something awesome in their face, or they're lost.
6EvelynM12y
That's an unhelpful, unthoughtful answer. You can do better. "Optimizing Reality"
4roystgnr12y
One distinction between criticism and mere insult is that a criticism always has a chance of helping the person criticized, by explaining a specific problem and thereby suggesting an avenue of improvement. I downvoted your reply, which was not a criticism. I translated his joke as "Optimizing for minimal bounce rate is an oversimplified goal, which could interfere with more important goals unless they are given more consideration." That seemed helpful and thoughtful enough to me.
1gwern12y
It also seems to have been made repeatedly in the comments already.

I'm all for trying to make the LW front page more engaging, but I am skeptical about an assumption pervading this post: that growth should be a major goal, and that faster growth is better aside from major pathologies like spammers and trolls.

For a website that exists mostly to make money (by selling things, advertising a product, gaining visibility for a person or company, etc.) this is a good assumption: all else being equal, you want more page views, lower bounce rates, higher conversion rate, etc. But LW doesn't exist mostly to make money; it exists (in high-minded principle) to refine the art of human rationality and (more prosaically) to provide interest and entertainment for its participants, and for these purposes some visitors are much, much more valuable than others and many have negative value. (This is true even taking into account the prospective value of LW to those visitors.)

By definition, LW wants (or should want) to "grow optimally" -- but that may mean "grow very rapidly" or "grow slowly" or even "not grow at all, at present".

6beoShaffer12y
I concur. I have implied this in my previous comments, but I will going come out and explicitly say it here. Caeteris paribus I want LW's growth rate to be higher than it currently is, but I am more concerned with keeping out the intellectual riff-raff than encouraging growth. Furthermore, I actively want to avoid the exponential growth scenario, unless we can create an effective user orientation or another way of reducing the time it takes for new members to acculturate.
1Epiphany12y
I agree that there's a lot of annoyance for both the new user and old users when new people join. A new user orientation is needed. That's why I wrote an outline for a New User Orientation full of suggestions that I personally would like to see included in the new user orientation. Please go and critique it - older members are surely going to have suggestions I wouldn't think of. I have nothing against writing one, but I can't write a good one with no input, I'm too new. Please distinguish "keeping out the intellectual riff-raff" from elitism. Are you talking about people who really aren't serious about rational though? Maybe you just mean you don't want more trolls? Or do you mean you want to outright make an IQ requirement? I have an idea for scaring off those who are not serious about rational thought that goes like this: From my post New User Orientation Do you think this would work?
1Epiphany12y
What are the reasons for not wanting to grow? You didn't state any reasons I haven't already stated. For the record, I am aware that it's not a good idea to assume that growth is 100% good. That's why I put that in the OP. I feel that LessWrong could be really important - and maybe it already is - that's why I want to see it grow (that much is obvious.) I started writing about why, but I want to hear your reasons for not wanting it to grow. You're going against the grain - not a bad thing but it means you're going to have to really lay out your reasons if you want to change the way the wind is blowing. Elaborate, please.

(I think you may have a wrong idea of what "pontificate" means. Either that or you're being gratuitously rude, which I'm going to assume you aren't.)

The following two propositions are different. (1) Growth is not always 100% good. (2) Growth is not always good. #1 is what you stated. #2 is what I'm saying. #2 goes much further than #1 does. The obvious inference from #1 is "make sure you grow, but take some measures to mitigate the possible downsides"; the obvious inference from #2 is "consider carefully whether growth is what you want".

I don't know whether I want LW to grow, or how fast, or when (which is why I said that optimal growth for LW could be any of several things, including "grow very rapidly") so I can't really give you "[my] reasons for not wanting it to grow". What I can do is to give some possible reasons why growth -- especially rapid growth, especially especially rapid growth by the mechanism you're implicitly proposing -- might be the Wrong Thing.

The main reason is this: It may not be possible to grow rapidly by the sort of means you describe without changing the demographics of LW in a way that would lower its qua... (read more)

1Epiphany12y
You had a good point in your suggestion so I changed my "100% good" statement. I also responded (different comment).
0Epiphany12y
Oops sorry. I removed pontificate. Thank you for not assuming I was being rude. "I mean that maybe only a small fraction of new visitors to LW are people whose presence would enhance LW, and that if that's so then anything we do to encourage a lot more new visitors to stay will make LW worse." Good problem. However, that can happen whether it grows quickly or slowly. "I am not optimistic about the prospects of keeping it if LW grows rapidly by keeping a much larger fraction of its new visitors." Yes! I did warn that this could happen, and that it could happen whether we like it or not. I wonder whether you think this idea would work: "http://lesswrong.com/r/discussion/lw/e5r/lesswrong_could_grow_a_lot_but_were_doing_it_wrong/78r2" I also had another idea: I like that things can be voted up and down, but I think it would be a heck of a lot more effective if specific feedback was provided. As a new user, I will adjust a lot faster if I know WHAT I am doing wrong/right not just THAT I am doing something wrong/right. Also, I wouldn't write off the votes to "bad attitude" and "trolls" when I don't understand them. That's what I'm doing now, lol, and I don't like that, but I don't know what else to think of them, lol. I think if the votes required a reason, which could be as quick as the word "Trolling" or the name of a logical fallacy, that would make a huge difference. That would ALSO force us to get conscious of our reasons for voting, which would provoke conscious review of the reason, which would probably result in better decisions. And to ensure the culture doesn't change too quickly, what if it took, say, three months and X number of posts before you're allowed to vote? That way, the oldest users get to influence the culture, and users who just aren't compatible with LessWrong will wander away before they've exerted an influence. This is basically the same as how you have to be a resident of a country before you can vote. From a developer's standpoint,
1gjm12y
Yes. It's an issue regardless. But if pursuing a strategy of rapid growth guarantees an influx of negative-net-value-to-LW new users, it's probably a bad thing. (We will all die eventually, but I still prefer there to be fewer mass murders.) I don't have a good enough mental model of the typical first-time visitor to LW to have a strong opinion. What I do think likely is that either it wouldn't work or it would work at the cost of getting rid of that rapid growth you were hoping for. Because I think it's likely that rapid growth by increasing visitor retention implies the sort of change in LW's demographics that I described. I have proposed something similar myself. The UI would be tricky to get right. It might well make LW a better place (or it might not; these things have a way of producing unexpected consequences). But I don't see how it would do much to solve the problem I described, if (as I suspect but don't claim to know) it really is a problem. Might work. Has anyone reading this got experience of such systems? (The obvious concern is that it would drive away "good" users as effectively as it would drive away "bad" ones, so that it would slow growth without actually making the overall pattern of growth any more favourable.) According to my quick count, 28 different people have commented here [EDITED to add: not counting "metatroll"], of whom 7 seemed definitely in favour of growth (perhaps having thought it through, perhaps just because it's a sort of default goal), one seemed definitely against growth, and 6 seemed definitely skeptical (i.e., saying "growth might not be what we want", basically my position). The other 14 expressed no opinion on the matter that I could detect. Make of that what you will. I think a lot of people have, but so far as I know no one (here or elsewhere) has a silver bullet that ensures that a community of unusual people will retain its distinctively valuable characteristics as it grows. (Or for that matter as it doesn't.)
0beoShaffer12y
I'm not sure if I like the basic idea, but tying it to the new users' karma would favor good users over bad (for certain values of good).
1gjm12y
(Your quoting is slightly broken.) Hacker News uses this sort of system: there are thresholds for things like being able to vote, being able to downvote, etc., and they are all based on your karma score. The same is true on Stack Overflow, though that's a very different kind of site. Both of them see frequent complaints that they're going downhill, but it's hard to be sure whether that's anything more than standard "the world was better in my young days" thinking (which I think results from a general tendency to remember good things better than bad things).
0beoShaffer12y
Frustratingly, the help doesn't say how to do nested quotes.
3gjm12y
Hmm, let's experiment. This line is normal. This line is normal. All those lines are separated by blank lines. (Not doing so produces bad results.)
0beoShaffer12y
Ok, I think I have it working.
0beoShaffer12y
Yes, but one of the things that effects wether someone is a good user or bad user is how well they acculturate. From what I understand, sites with a high old timer to newb ratio usually have an easier time acculturating the newbs.
0Epiphany12y
That's not what I meant... Imagine a curve that gets more and more steep as it progresses. In the first quarter of a year, say a website gets 1000 new users. In the second quarter of the year, it gets 1500 new users. In the third quarter of the year, 2250 new users. In the second quarter, 2/3 of the members are old members, 1/3 of the members are new. In the third quarter, the ratio changes because there are so many new members - nearly half the members are new. In the fourth quarter, an even larger number of new members joins, and suddenly there are more new members than old ones. If this keeps up for any length of time, the culture will be totally destroyed. Whether you start with 20 members and increase it by an increment of 1 and 1/4 the amount of users every quarter, or start with 1000 and increase it by an increment if 1 and 1/2 the number of users every quarter, if the number of users keeps increasing this way, eventually it will get to a point where the ratio flips and there are more new users than old ones. It doesn't matter whether that type of growth happens quickly or slowly. If the growth is exponential there's a good risk of that eventually happening. Slow growth might mean it happens next year, fast growth might mean it happens next month. Either way, it's important to protect the community from that problem.
1beoShaffer12y
I think you're missing my point. Does your model have any allowance for the idea that new members can eventually become old members, and that this process is speed up by new members interacting with old members? Because the main point that I was trying to make is that a large portion (though nowhere near all) of a user's desirability is how well they absorb the culture, and that culture can survive (more or less) intact through multiple "generations" if at any given point in time a large enough proportion of people have already acculturated. Consider a toy model were users can be discreetly divided into "old-hands" and "newbs". In this toy model newbs turn into old-hands in x weeks iff at least 50% of the other users are old-hands (who can set a proper example), otherwise the culture collapses. In this case the culture can survive indefinitely iff the doubling time is longer than x.
0yli12y
Of course it could still be a good idea to rewrite the home page, since maybe the sort of people who would be attracted to a site with LW's current home page have on the whole negative value for LW.
0Epiphany12y
Why is it attracting the wrong ones, yli? I have a vague sense that you might be right, so I'm very curious about your reasoning.
6yli12y
Well since the front page is so generic, it might attract people who react positively to the mere idea of "rationality comminity" who are more likely to be people like stereotypical Objectivists and entry-level atheism fans, as opposed to the kinds of critical and savvy people that LW would need to prove itself to by eg. showing them some awesome posts from the archives and not expecting them to be excited merely about the idea of a site about rationality.
1Epiphany12y
Yes, yes! I mentioned the same thing in a different comment and I'm glad to finally see that someone else gets it. (: Other people probably see this also, but it seems most are not really interested in analyzing how new people are reacting to the front page and about page and how it affects who does and does not join. Thanks, Yli. I'll definitely be interested in hearing any other theories you may have.
0yli12y
I'm happy you see where I'm coming from. Another opinion I have about issues like this is that to fix them, it wouldn't be enough to read complaints in comments in threads like this one and try to fix things you find people complaining about. You would actually need to find a person (or persons) with a good eye for and intuitions about these things, who has good taste and who knows what they're doing, and just let them take control of the whole design and let them change it at their discretion. I think there must be a few people on LW who would be both capable and willing to do it, though of course the site as it currently is would repel them.
3Epiphany12y
I feel capable of doing the intuition part because I'm a psychology enthusiast and have done a lot of free coaching and emotional support, so I'm good at intuiting things about people's feelings. It's a matter of knowing what these guys care about the most. Without a feel for their culture, I won't be able to target people who already have the same culture - this is true - but since this site is creating it's own culture, that's not the objective. Finding people compatible to the culture, since it's a rationalist culture, isn't quite the right problem to solve either. Firstly, these guys seem to want NEW information, not just hear the same old stuff - they're not looking for people perfectly following their conventions, conforming to a stereotype, the way that most cultures do - that would be boring. Secondly, it's a culture of rationalists. A true rationalist wants to be proven wrong and to change in the event that their ideas weren't as good as somebody else's. So what we're looking at here is not the challenge of "Find more people with the exact same ideas." but the challenge of "Find people who can contribute to this culture." Of course there is a fine line to draw when making the distinction between those who contribute to the culture and those who do not. The most important way to make the distinction seems to be "don't subject us to tedium". I theorize that boredom is the brain's way of punishing you for not constantly learning, which gives one a survival advantage since we live in a world where even one small spot of ignorance can cause a problem, yet we can't know everything. The only solution is: learn constantly. So, it's no wonder humans experience boredom whenever they're idle or doing something far too easy. Depending on one's intelligence level, one might be fully immersed by the human world at all times, because it's designed to put people at that level into a state of flow, or they may be totally left out and experience flow very rarely. I think a
2OnTheOtherHandle12y
Do you think it might be possible to make Less Wrong more interactive? There are a lot of simple rationality tests which most people get wrong; like the classic tests used for Confirmation Bias or The Conjunction Fallacy. Could we make fun versions of such tests for people to try out, perhaps with animation, etc? With a little work, we could even come up with little games that test Bayesian reasoning skills. This would select for active, interested people who like to try things out, and if we quickly explained mistakes as soon as people made them, the idea that "I, too, am irrational" would be much more salient. Rationality is a skill we have to practice, in the end, not just a thing we read about. (I apologize if this would be too hard to implement to be worth it; I have very little programming experience.)
1Epiphany12y
I've been thinking about this for a little while. This is a really, really good idea. It's not too hard to implement. I feel there are some other things to do first (prevent endless September for instance) but other than that I think it's a great idea and it's just a matter of constructing these with the right questions. Questions have to be worded very carefully, and tests constructed cautiously, in order to have the scientific properties that give test results their accuracy. On the other hand, fun internet quizzes can bring in users and do not necessarily need to be scientifically sound (though for this site in particular, I'd figure that wording even a fun quiz as scientifically as possible would be the way to go, as that would attract more like minded people and gain more respect). It would actually take a lot more time to think of all the questions, consider how scientific the series of tests was, and word all the questions correctly than to set up a script that allows you to take the tests.
1OnTheOtherHandle12y
At least to begin with, we don't have to come up with things on our own. There's a whole literature of psychological studies we can comb through to recreate. The Sequences cite a whole lot of iconic studies, and a sufficiently motivated person could dig up more obscure follow-ups, too. Converting them into a format that would work on the Internet is a bit trickier, but a lot can be done with Java applets.
0Epiphany12y
Ooh... But what about copyrights? And if they are copyright expired, would they be any good? Maybe. Hmm. Do you have suggestions for specific materials we could start with?
1OnTheOtherHandle12y
The first page I linked to shows a Confirmation Bias test designed by a LWer, based on the classic experiment. The Conjunction Fallacy has a simple written multiple choice test. Anchoring and adjustment could be tested for by providing people with high or low random numbers and asking them to answer a bunch of estimation questions, like the "How many countries in Africa?" test. It would work out best if we already had people take this test and had gathered data to show people upon completion.
0A1987dM12y
That's what downvotes are for, isn't it? OTOH, if a potential valuable contributor sees the front page, doesn't fully realize what the site is about, and walks away, that's too bad.

Don't forget A/B testing...

2Cyan12y
I think this is really important, so I'm just going to say it again, only louder. DON'T FORGET A/B TESTING.
3John_Maxwell12y
A/B testing is useless without variations to test. If you want to create a variation, go for it. Right now there is a lot of commentary in this thread but not much action.
1Epiphany12y
We're starting to come up with a lot of plans here and I don't want this to step on anybody's toes. I'm not sure whether you're the sole decision maker for the website, or if other people need to be told. Do changes like these need to be approved or is the website delegated to you?
-3John_Maxwell12y
I'm no decisionmaker. I just created that post because I thought things could be improved. If you or anyone else has text they want to put on the about page or the home page, send me a personal message and I'll tell you how. Right now things rely on security by obscurity. Edit: As matt points out, it's not security by obscurity so much as Wikipedia-style open collaboration.
8Eliezer Yudkowsky12y
Please don't. All edits to the about page should go through an editor. Random people should not be told how to edit the about page.
0John_Maxwell12y
Er, want to specify who counts as an "editor"? Don't worry, I would have politely turned down anyone who didn't meet some threshold of credibility...
2lukeprog12y
See Who are the LW editors?
3John_Maxwell12y
OK. This policy doesn't make very much sense, in my opinion. Based on the log, lots of people have already edited the homepage who weren't editors, and at least some of the edits they made were valuable. Asking an editor's permission to make changes to the homepage is an inconvenience, and it's also a little demeaning. I suspect that the (extremely small) amount of community effort that's been put forward towards actually making improvements to these pages will completely dry up if this policy is broadcasted. (I know my enthusiasm has dropped dramatically.) I could see why this policy might sense if spam or prank edits were a problem, but as it is it seems needlessly controlling. Bleh. Edit: Eliezer has communicated via email to Louie, Matt and me that he retracts his statement.
5lukeprog12y
The Singularity Institute has invested hundreds of thousands of dollars into LessWrong.com. I don't think it's too unreasonable that we'd like to have some quality control on a few central pages like the home page and the about page.
1Epiphany12y
Wait if they've got so much money to invest into this, why have users been allowed to edit the home page and why is the marketing bad? Might those pages have been done like that intentionally to throttle growth and filter people using an err on the side of caution approach to preserve the culture? Maybe they have plans in mind we don't even know about, and this entire discussion is irrelevant.
5lukeprog12y
Note that most of the investment was (1) web development donated by TrikeApps and (2) Eliezer's salary while writing The Sequences. We don't have much money to invest in further development, but we are currently paying Trike and an oDesk coder to add new features to LW and fix some old bugs. I have plans to revisit the LW design and marketing but... so much to do, so little time/money. If you'd like to help, you can (1) give earmarked donations or (2) find a professional web designer/marketer who is willing to look at this stuff for free.
9Epiphany12y
Volunteer web pro here: I'm a professional web developer. I primarily do CGI programming. I am also capable of front-end web design / graphic design, database work, and other related tasks. Related tasks: ensuring the host's sysadmins do things correctly, solving random technical issues, and various tasks that make use of cpanel / root server access. I've done a lot of work with a search engine optimization expert and a marketing coordinator, and have read quite a bit about web marketing over the last ten years or so. I'm not specifically a marketing professional, but I know a lot more than a non web professional and I'm willing to do the following things: Review the Google analytics data (requires access to the account), present LessWrong members with ideas, come up with a plan for growth that is popular and takes all significant criticisms into consideration, present the solutions to an authority for approval or changes, implement those solutions (assuming the solutions are not so time-consuming that I can't fit them into my schedule but so far the ideas are not too large... though I definitely want a good plan for preventing an overwhelming number of new users incompatible with the culture, for sure, before doing things that will increase the risk/speed at which one is invited, and in fact, this protection may need to be added regardless if there is not already a plan to prevent it, so that could easily end up to be the #1 priority were LW to accept this offer), review the situation and data afterward to see what the result is, and either put it back to the way that it was or try an improved plan. If there are unrelated requests, I may grant them, depending on whether they are within the realm of my capabilities and time constraints. Feel free to ask about them. I can provide one or two freelance references (I'm F/T so I don't have a long list of freelance references), a criminal background check, credit check or other reasonable checks to an official LW autho
1lukeprog12y
Thanks! Would you email me to discuss further? You're right that contacting Eliezer directly isn't going to be the most successful entry point into dialogue with SI.
1A1987dM12y
People able to edit other people's LW posts. AFAIK, Eliezer and Alicorn (and possibly someone else too).
-2[anonymous]12y
He has already provided enough information. It is trivial to figure out how.
6matt12y
Erm… that's security by obscurity in the same way that Wikipedia relies on security by obscurity, right?
0John_Maxwell12y
Fair enough.
0shminux12y
Current design is the control.
-1John_Maxwell12y
Hm. Do you see any advantages to formal A/B testing over just popping something in and seeing how stats like the bounce rate change?
-1Raemon12y
DON'T FORGET A/B TESTING!!!!!
0Epiphany12y
I've been thinking about that. Here are my thoughts: Obviously we will want to see if the new pages are getting more members than the old ones - we can look at the statistics for that. I'll definitely want to know if we're getting results if my idea was used, and I will be really curious about why not if that's the case, which will lead to speculation and new ideas, I'm sure. I am also really curious about which articles hook users most effectively. I'm thinking a random excerpt rotator would do the trick. There are ways to see, in the statistics, which pages are working the best. (By going to the stats for that particular page and see how many people were referred over by the home page.) This is going to be a lot of data to demand. So, either somebody with access to Google Analytics is going to have to get involved, or I will have to get access to it. I'll see what they want to do. Ah! Also! I'll need to get "number of members" statistics before we do anything, and then compare a before and after.

Epiphany, I'd love to see you actually take a stab at an introductory paragraph for the About page that you think would work well.

I don't think anyone has the budget to hire a professional web marketer. Additionally, the crowd we're "marketing" to is a bit atypical, so it's not clear how well standard principles would transfer. Less Wrong is full of long-form essays on rationality; if you don't like reading, you're probably not going to like the site.

I think it would be great for more people to produce variations and suggestions for the LW newc... (read more)

9Epiphany12y
I would totally love to take a stab at the About page. I want to get more information from these guys first. I want to know what sort of people they want to attract. I have a gist of the culture here, but my gist of it is new. I want to make sure that I'm targeting their audience. Really, we need to do that first. That's so important. Whether we do a good job of marketing this site or not, we could one day find ourselves so overrun with new users such that the old users are out numbered. That could happen so fast that the entire place seems to change in just a few months - that's how the exponential curve at another forum was experienced. We want to do whatever we can so that if we hit a curve like that, existing members are happy with the influx and the influence new people have on their culture. Maybe a new thread should be made "Who do you want to meet? Choose a target audience." - Do you want to do this one or should I? I am asking because I don't want to step on your toes - and there might be questions about the culture that only an older member would think to ask. Not sure.

I would like to see more diversity. Not just in terms of demographics (though that too), but in terms of fields and specialties. There is nothing inherent in rationality that should limit it to computer/ math/ physics/ philosophy types. There are highly intelligent people in other fields also, and I feel like people from other disciplines could introduce an influx of new ideas.

Also, I think recruiting new members is a positive goal, and fully support it. I would like to see the community grow.

There is nothing inherent in rationality that should limit it to computer/ math/ physics/ philosophy types.

Actually, I'm pretty sure there is.

It really really helps to be comfortable with math to do rationality, there is no way around it. The kind of people who have both the capability and interest to master things like programming or probability theory or Quantum mechanics will tend to be what you call "computer/ math/ physics/ philosophy types".

There are highly intelligent people in other fields also, and I feel like people from other disciplines could introduce an influx of new ideas.

But again how do you know we don't already efficiently mine smart people from other disciplines? Surely you don't expect smart people to be evenly distributed between professions? Consider the 2011 census:

In order of frequency, we include 366 computer scientists (32.6%), 174 people in the hard sciences (16%) 80 people in finance (7.3%), 63 people in the social sciences (5.8%), 43 people involved in AI (3.9%), 39 philosophers (3.6%), 15 mathematicians (1.5%), 14 statisticians (1.3%), 15 people involved in law (1.5%) and 5 people in medicine (.5%).

Computer scientists probably are... (read more)

1daenerys12y
Why do you think you need to be able to master quantum mechanics in order to "do rationality" ?
0CharlieSheen12y
I don't. However EY keeps emphasizing how crucial the QM sequence is to the other material, so I take his word at it. I do think probability theory and a lot of other math is a must.
1shminux12y
You guessed the teacher's password! Now, can you recite (and criticize) his reasons? Why? There is very little math in the Sequences, and almost none beyond the American grade 10 equivalent. Most is simple arithmetic and an occasional simple equation.
1CharlieSheen12y
How clever of you to share another one! A gold star for both of us! Can you now explain why trusting a sound rationalist's or specialist's conclusions based on their authority if one hasn't the time to investigate them oneself is wrong from a Bayesian perspective? I think it mattes for his arguments about us being the pattern in our brain rather than the meat of our brain. But again I haven't read all of the QM sequence, I don't recall claiming I was a particularly good rationalist, all I claimed was that: "It really really helps to be comfortable with math to do rationality, there is no way around it. " You don't need to be a great rationalist to see that. Please tell me how many Americans with 10 grade equivalent can read and understand any of the statistics used in papers LWers cite. How much of a gwern do they have in them? Those who can't and don't read the studies cited are taking EY's or gwern's or lukeprogs conclusions on various topics as much on authority as I am the relevance of QM to rationality.
0shminux12y
It helps, I don't disagree. Especially if you have to calculate some Bayesian thingies. But that's an advanced level. In a hypothetical RationalU it would probably correspond to the third year.
4CharlieSheen12y
To take a stab at that applause light. Ceteris paribus yes I can agree diverse contributors may be beneficial to our mission. Especially value diversity, since differences in desired conclusions may lead to motivated cognition being called out more. But I think when most people speak of diversity they don't have that kind of diversity in mind. So sticking to the other kinds, I have to note that I haven't seen a single data driven argument for this why this would be so in actual humans. It is simply assumed or asserted. I'm pretty sure this is so because it happens to be a sacred value of our society. While cyberspace is clearly different from meatspace, studies done in meatspace seem to show diversity has negative effects that are seldom talked about. But leaving aside such undesired consequences let me just point out that the ceteris paribus in my first paragraph is also unlikely. Very small differences can result in almost complete homogenization. Not only has it proven difficult, expensive and perhaps mostly ineffective to do this in meatspace, you would actually need to keep a delicate balancing act to keep a certain heterogenus mix together if your goal isn't to simply swap one group for another. I think we already spend insufficient attention spent on gardening and this would be just one more difficult task to add to the list. Once more I ask why no one ever subjects such proposals to explicit cost-benefit analysis?
2Epiphany12y
Hmm. That's an interesting idea - and you got a lot of points for that. This is a more complicated task than simply ensuring that users who are interested in developing their thinking abilities get a chance to see how awesome LessWrong is when they come to the front page. I will think on it.
5John_Maxwell12y
My answers: smart people, people who already have a strong technical background in something, people who already have a strong interest in rationality. This sounds like a good problem to have to me. If we later decide we don't like it, it should be easy to just change the about page back to the way it was before. There seem to be a lot of lurkers on LW, so a surge of new lurkers might be more likely than a surge of new contributors, with lurkers gradually moving to contributor status. IMO it isn't worth a new post, for the time being at least. I think the principle of be bold applies to this sort of collaborative writing process. In other words, throw something at us and we'll give you feedback on it.
0beoShaffer12y
I second daenerys and John's user requirements, but will add a few more restrictive ones of my own. I want people who are good at taking and giving constructive criticism. I also want people who are willing to read a fairly large back log of posts (at least a decent chunk of the sequences), and read more if people give them links indicating that we've already covered whatever it is they are talking about.
[-][anonymous]12y60

Potentially useful: When I first got linked to LW. the thing that hooked me was the "featured articles" thing (which prompted me to read "rational home buying", which was totally awesome) and the "yes, a blog" post in the about page.

EDIT: IMO, we should gather more stories of people getting hooked to see what actually works and what to emphasize.

Is it known what portion of new users land on the front page versus, say, being linked directly to an article?

Is 60% actually high for a blog/forum? I'm not a professional but I was under the impression that it varied drastically based on what your website is, and that blogs/forums are naturally really low. I'd imagine it would be especially true for LessWrong, simply because of the nature of the material.

Google Analytics Benchmark Averages for Bounce Rate

  • 40-60% Content websites
  • 30-50% Lead generation sites
  • 70-98% Blogs
  • 20-40% Retail sites
  • 10-30% Service sites
  • 70-90% Landing pages

This is what a cursory search pulled up. Is this way off base... (read more)

5Epiphany12y
If the majority of the users were returning visitors, I wouldn't be concerned about the bounce rate (but then I'd have to wonder why the site wasn't getting new visitors.) If I felt the front page did a good job of putting something awesome in your face (instead of hiding it behind a link) then I wouldn't be concerned. If the sitepoint graph showed the site growing as quickly as 50% new users per month implies it could, I'd assume the bouncers were returning visitors. It's that combination of factors that makes me think that new visitors are going away and that the web marketing has room for improvement. A lot of blogs have only a single page - that single page nature may be why they've got a higher bounce rate compared with other types of sites because new visitors get the content they were looking for on the first page, and because there's not anything else to do but read. Yes, you can read every post if it's really interesting (Hyperbole and a Half was like that for me.) but 90% of the time I'm just grabbing a recipe real quick or trying to figure out how to fix my blender and the information I need is right on the first page, then I'm done. These single page blog sites aren't a gateway into a community. And LessWrong isn't a place with practical information where you're doing a quick question run. Correct me if I'm wrong, but this place attracts intellectuals who love reading. Assuming that's the sort of person coming here, we have to ask why new visitors are not clicking. It's not because they're satisfying their reading fix on the front page as with a single-page blog. The front page doesn't have a lot of text. If they have an intellectual nature, and are expecting to read something interesting, they should be showing more curiosity about other pages. That they're not means the page failed to interest them. Why else would a new visitor leave without clicking on something when you think about the current home page design? When it comes down to it, there's no
2gwern12y
FWIW, for http://www.gwern.net since the beginning, the bounce rate on my main/index page (as a landing page) is 31% on 47% new visits. Whether that is very good or very bad compared to LW, I will leave it to experts like Epiphany to say. (If it's very good, I think it suggests a design strategy like 'include a list of LW Greatest Hits/mind candy with brief tantalizing summaries'.)
2beoShaffer12y
As a (typical?) user of gwern.net I'll point out that I had was already familiar with your work from other sources. Also, if many of your new users com from LW there is likely a pre-screening effect.
1gwern12y
I do get a lot of traffic from LW, yes, but generally they're going straight to specific pages like the modafinil page. The frontpage gets much less from LW according to the Analytics, although you're right that the bounce rate for LWers is one of the lowest.
2John_Maxwell12y
For bullet points, what works for me is to either use asterisks and separate the bullet points by blank lines, or add 2 spaces to the end of lines that I want to break.
1Xachariah12y
That usually works for me, but the quotes just seem to mess things up. Specifically quotes with bullet points after a quote without a bullet point seem to cause troubles. * Bullet point Edit: After testing, it seems that the first item on the list controls whether anything can be a bullet point within a quote. Odd, but at least I understand it now. Editx2, Ah shoot. Now I can't figure out how to escape quotes properly. I think I'll quit while I'm ahead.
2loup-vaillant12y
Let's see… Source code (indent with 4 spaces): > Begin quote > > * Try bullet point… Now… * Bullet point Source code: * Bullet point > * Bullet point inside quote inside bullet point > > Quote inside quote inside bullet point. > > > > * Aaand the bullet point! > > > > [non-breakable space (the actual one, not '&nbsp')] > > Looks like it works.

I want to prefix this by saying I am interested in improving the front page anyways and that I'm not experienced with google analytics, but did check its help page before posting.

As far as I can tell the bounce rate includes return visitors. This leads me to suspect that the high bounce rate could be from regulars checking LW and leaving when they don't see any interesting new posts or recent comments. Also, I think we can do a better job of meeting both goals, but I for one am willing to accept a fairly high bounce rate if it keeps the quality of newcomers high and the current system seems to be doing a passible job there.

Sure, the bounce rate may include returning users, but you seem to be missing this really important point:

When I first saw the site, I had no idea what it was about. I left. Without clicking anything.

When I saw the site the second time, I read the about us page, still didn't know what it was about, and left. Without clicking anything.

A web professional who has worked with marketers (me) is telling you "the web marketing sucks".

We really don't know what percentage of the bouncers are new users from this page. But if we want growth, we can't just tell ourselves that most of the people who are leaving are returning users and brush this off. Consider this:

http://www.sitemeter.com/?a=stats&s=s18lesswrong&r=36

If the new users were all registering, the growth should be how much faster than that?

(I'm hereby updating the OP with this. Thanks for pointing it out.)

[-][anonymous]11y30

Something just clicked in my brain... As far as I can tell, the LW front page doesn't have a link to either the Welcome thread or the FAQ. Instead, the intro section alone has a dozen links, and even the About page doesn't get mentioned until the end, and is perhaps awkwardly prefaced "For more..." as if the previous reading material wasn't enough to start with. (The FAQ is linked on the About page though, albeit hidden in the middle.)

No idea if this comment will even get seen by anybody, but I felt like giving some anecdotal insight since I have an oddball curiosity in how website communities grow and how webdesign can make that growth positive or negative.

I've never heard of Less Wrong before 20 minutes ago. Found it randomly via a video game website where someone had linked to an article here, took a few moments to look around...

And I almost bounced from the main page. Why? My guess is jargon. There's a lot of it on your main page, and the first link in the "Welcome to Less Wron... (read more)

I'm not sure growth is necessarily a thing LW needs per se, as some of the other commenters have pointed out. But I do think there is scope for improving the landing page and decreasing the bounce rate of people we want to be here. That last is crucial.

In my case, I still only have a handful of comments due to not having much time to post at the moment, but I got here through HPMOR and read the Sequences... and still had problems sticking around to post anything as I was somewhat underwhelmed by the design and ease of navigation. I think things can definitely be made a little more user-friendly without ruining the community.

There is a additional responses that people go out for reasons than bad desing or misinformation. The goal is not "outsiders completely agree with the core mission in the first phrase", but a broad idea. Sites like facingthesingularity.com and intelligenceexplosion.com seems like paradigms of good exposition.

Highly support this initiative.

I went ahead and updated the homepage, about page, and FAQ in line with the voting here, mainly so Google would stop penalizing us for duplicated content.

I don't know if we want more visitors as much as we want more converts. We really do have a lot in common with a phyg, in that we have a particular worldview that isn't mainstream and that we're trying to spread.

4John_Maxwell12y
Well, if they succeed in "converting" us to their way of thinking, because it's correct and they can persuasively argue for it, that's even better. Right?
0CronoDAS12y
Hopefully!
0Epiphany12y
What is a convert? An irrational person who wants to become rational, or a more or less rational person who wants to become better?
3CronoDAS12y
Either. More specifically, our worldview includes: Atheism Physicalism Transhumanism, including radical life extension (or at least the belief that it would be a good idea) Empiricism Ethical naturalism and some other things, including rejection of most of the kind of nonsense the Skeptic movement complains about

If you could find more detailed information on referring sites or searches that are driving people to the site, that could be helpful. I don't know squat about it, but I'd think on site web analytics would provide direction on improving retention. I do hope that the keepers of the site avail themselves of your web marketing experience.

0Epiphany12y
I have requested access to the analytics. (: Thanks for your encouragement. (: