How do we grow the Server Fault and Super User communities?

Question

Here are the last few weeks of the "visitor" metric per Google Analytics, on each Trilogy site:

meta.stackoverflow.com

http://img30.imageshack.us/img30/8907/analyticsvisitorsmeta.png

serverfault.com

http://img200.imageshack.us/img200/503/analyticsvisitorssf.png

superuser.com

http://img30.imageshack.us/img30/6666/analyticsvisitorssu.png

stackoverflow.com

http://img269.imageshack.us/img269/7807/analyticsvisitorsso.png

They are ordered in terms of least visitors to most visitors. Stack Overflow is the blockbuster here. Which makes sense, as it's the oldest, and the site that is most in tune with the Joel on Software and Coding Horror audiences we originally started with.

Meta is only included for completeness; we don't actually need any particular traffic level there (aka, "here") due to the nature of the content.

I am not showing the larger trend graphs here, but the deeper concern is that Super User and Server Fault are not growing much (if at all) over time. Google traffic levels are one sign of a site's maturity:

• SO has 90% google traffic
• SF has 77% google traffic
• SU has 67% google traffic

So we're not quite "mature" yet, but we're getting close.

I am not sure we've done a good job of reaching out to the power user (superuser.com) and sysadmin/it pro (serverfault.com) communities to make sure they're at least aware of these sites.

Why this matters: I worry that if we don't attract a critical mass of users there, questions and answers will languish. They don't need to be enormous, I just want them to be useful. And that implies some growth, at least!

So, this is my question to you: what should we be doing to reach out to the power user and sysadmin communities? How can we attract more of the kinds of power users, sysadmins, and IT pros that we'd like to see asking and answering questions on superuser.com and serverfault.com?

Suggestions welcome.

Answer 1

The main problem with SU is that people usually have one problem, they go to the site, post their question, get an answer (hopefully) and they're off. It's not like programming, where you always try to gain knowledge by answering questions, by reading more stuff, by asking others for help. I remember someone saying on meta that SU is a sort of tech support; in some ways it is. People usually need it when something's broken.

However, some (broader) topics might gain an audience. Gizmodo, Engadget, Lifehacker and Ars Technica would be good places to look for a potential audience. A lot of answers on SU are based on suggestions got from one of these sites. It would be nice if some of the editors on these sites could be engaged into becoming active members of the community. John Siracusa, an Ars Technica columnist, was, at one time, quite active. If you could convince these guys to play an active part in the community, it could really help.

There's another problem on SU: a lot of the top users seem to lose interest. Out of the 35 users on the first page, 16 have had almost no activity within the last week. And if you think it gets better with the second page users, it doesn't; it's 18 almost inactive users there in the last week. I believe a community is driven by its top users. If they can't be kept engaged, them it's a major problem.

Answer 2

Apparently this issue needs to be discussed here.

Specifically dealing with Server Fault, I can see why the site is not growing. Consider that you are competing against Experts Exchange, news groups, and many different vendor forums/communities. Server Fault really has some faults to overcome.

Problems

1. I think the primary problem is that the Server Fault community is not made up of professional system/network administrators, but is instead developers asking questions about their server environments. That may be great for a PHP developer wondering why a module is not loading in Apache, but the result is a bunch questions that are not relevant to most administrators.
2. A second result of the make up of the community is that extremely technical questions do not get answered because there are very few people to answer them. Also, when they do get answered, most of the community does not understand the topic enough to vote. Notice how low the vote counts are on most questions compared to Stack Overflow.
3. The site is not very friendly to newcomers. Put yourself in the shoes of an Exchange administrator and then look at the front page of SF. What on that page even hints that Server Fault is a good place to ask Exchange questions or, gasp, stick around and help others with Exchange questions? Compare Server Fault to Technet or Experts Exchange.

Solutions

1. First, I do not think you are helping Server Fault (or Super User) by tying the trilogy of sites so close together. You seem to grasp that these are separate audiences; however, you only advertise the three sites to developers. Of course, SO is going to take off while the others flounder.
2. Server Fault needs to be much less accepting of off-topic and Super User questions. Just because a question references RAID or Linux does not mean it is appropriate. Maybe give some direction to your moderators to clean up these questions to make the site more appealing.
3. Change the reputation system to reward those that answer difficult technical questions. Unfortunately, I really do not have an idea on how to do that. The current system rewards those who ask and answer easy questions making it just as easy for non-administrators to gain reputation.
4. Consider creating Portals on this site which target specific products or manufactures. It may be as simple is grouping certain tags on a portal page and changing the branding a little on each portal. Maybe this could include corporate sponsorships.
5. Seek out the experts! I would be contacting every MVP I could find to ask them to visit the site. And yes, there are tons of admin-related bloggers too but they tend to revolve around certain products (like Ben Armstrong and Microsoft Virtualization).

One Final Consideration

It may be time to step back and consider whether or not the Q&A format is best for Server Fault. There needs to be a better way to format a back-and-forth discussion between users which is required to troubleshoot and solve a complex problem. This isn't Stack Overflow were you can simply post a snippet of code and someone can point out your problem or a different way to attack the issue.

The current format lends itself towards simple questions and simple answers. The end result being a site that may not be to useful for professional system administrators. Evan's comment on a different question reminded me of this - none of the admins I have told about Server Fault have stuck around. Figure out why that is, solve it, and then your community will grow.

Answer 3

The same way you built StackOverflow. Find a blog that is very popular with those demographics and convince/Cajole those bloggers to review the site.

I'd think Superuser would be a good topic on LifeHacker.

Don't know many blogs focused on the admin types though.

Also, maybe if you posted some of the more popular (and general interest) questions to the appropriate Reddit, Digg, or Technorati subsites you'd get some traffic from there.

I tried that once for one of my questions on SO (Tells of a Novice Programmer) and got a decent bump in views the next day.

Answer 4

Well, since SO has all that trafic. There must be some ways to make use of it.

For example, how about changing the "first time visitor" message to give more information about the trilogy?

Instead of

First time here? Check out the FAQ!

You can have this

First time here? Check out the FAQ!
We also have sister sites focused on system administrating, serverfault.com, and super users, superuser.com.

Because I doubt the casual visitors of SO see the links to SF and SU at the bottom of the page.
And you could also try mentioning those sites to the users with 200 or less reputation, under the ad on the right column.

Answer 5

For SuperUser, the various wrt related tags have a fairly small number of hits (18, all combined). But router has 118.

I would suggest approaching the various wrt related projects out there (openwrt, dd-wrt, tomato, etc, etc.) for advertising or some sort of partnership. Perhaps offer tag sponsorship (with logo) to them (which I really don't think they'd ever be in the position to buy anyway), in exchange for them encouraging their users to seek help here.

I went looking at the wrt stuff since I'm familiar with it; but I expect you could apply the same idea to a number of other tags.

Answer 6

Perhaps it sounds a little obvious, but I'd recommend putting out an overt call to the community at each site asking everybody to spread the word. More eyes on the site is better for everybody.

I have yet to show Server Fault to a sysadmin and not received a favorable reaction.

Getting a few of those people who regularly attend local sysadmin SIG meetings, user groups, etc, to spread the word would be helpful, I'd think. As obvious as it sounds, I'm betting that most people don't even think about spreading the word.

(If nothing else, I want more questions on the site to answer... >smile<)

Answer 7

I have purposely not replied to this until now, as I wanted to see what the discussion would lead to. I must admit I am somewhat surprised at some of the responses and although I agree with most I also have my own thoughts on the subject.

Google is the Trilogy Home Page

One of the biggest issues I see coming forward from the community on a regular basis is the lack of quality questions and answers. However it seems that unless you were part of the original SO beta group, no one has realised the purpose behind these sites.

The trilogy is meant to be a canonical collection of questions and answers continually updated with the latest information.

Therefore to drive traffic to the site, the easy "Google It" questions should and must appear on SU and SF. With the ability of 2K+ users to edit posts, the idea is that these questions, if bad, gets edited into properly indexable questions with up to date answers. This is what makes the Trilogy stand out from the rest of the forums and sites out there. If someone types a question in Google relevant to a topic on one of the trilogy sites, that site should be the first 10 links that come up.

This means that the current user base needs to be guided and taught what the purpose of the site is, and the 2K+ users must take on the responsibility given to them by having the abilities, as they have a vested interest in the success of these sites. The development of the Stack Overflow Careers should be more then enough motivation to drive this change in attitude.

Blog, Blog, Blog

Each person that has a blog has a readership. Whether that readership is 1 person or 20, the more the sites are exposed in blog entries, the more they are indexed by sites like Google as the canonical source of information. Between the blogs and twitter, awareness needs to be spread irrelevant of the smaller communities it represents. Moms4Mom's recent explosion was caused by the twitter exposure when people like Scott Hanselman suddenly sends a tweet about it.

Both Joel and Jeff have access to some of the biggest names in IT, not just development. Get Scott to evangelise this in his podcast, even consider sponsoring his podcasts.

Advertising

Find the right podcast and become a sponsor. This could also potentially open up a huge traffic boost, and also reach a broader audience. Most users ignore internet advertising, we know this. However podcasts, RSS feeds and others can't be ignored. I have to admit creating a twitter account for each of the sites which people can follow is not a bad idea, maybe listing the top questions for the day.

Final Thoughts

Speaking specifically around the MVPs, involving them is by no means an easy task. They are often held back from involvement due to other commitments, and the never ending list of NDAs put in place by Microsoft. As a VSTS Ranger I have learned this the hard way. Not too long ago I was asked to remove answers of SO due to me violating the NDA, and I wasn't even aware at the time I was.

They are also by no means know the answers to every question, since at the end of the day received an award from Microsoft for an achievement, not because they know every answer. I am by no means saying they are not good, however they are human, and often not the right people to answer questions.

Summary

We need to ensure that we meet the initial goal of the site, which I have already listed. Be the canonical source of information where people find the answers to their questions. Not only by driving new users, but engaging and keeping the existing user base.

Answer 8

Here are my thoughts on SF's "problems":

1. There aren't enough experienced admins contributing. I don't think SF has reached a critical mass yet.

2. There's a need for some meta about SF, but meta.stackoverflow isn't working well for it. There aren't many of the higher-rep SF'ers here, and I don't think they're likely to come here. I think most people who are active on SF like the way the site works so meta about details of how the sites operate isn't very interesting and doesn't draw admins over here.

UPDATE: Just had an example of this, I asked a meta-question about how to deal with a certain type of "bad question" that shows up regularly on SF. Many of the answers weren't helpful because they're from people who are mostly on SO and didn't understand why it was a bad type of question. One guy gave a list of questions he thought mine was a dup of, but none of them was really the same as my question.

3. When you ask a question, one of the hints on the side is: "We prefer questions that can be answered, not just discussed." That just doesn't work well for a lot of issues sysadmins deal with.

   While it's certainly up to Joel/Jeff to decide what sort of site they want, I think if you rule out discussion questions, it makes it less interesting. I'd like SF to be a place where I could go (and I can suggest other people go) to get answers to all sorts of sysadmin questions, not just ones that have clear-cut, step-by-step answers.

4. I haven't thought about this one much, but it might help if there were a way to quickly filter the 0 votes/0 answers/minimal views questions. I think a lot of these are the "barely sysadmin" questions that people want to ignore. There'd have to be a time factor in there as well, since all questions start as 0 votes/0 answers/minimal views, but maybe if you could hide them if they're more than an hour old w/ low interest.

Answer 9

After advertising, I'd want to say that there are a LOT of fora for questions like the ones that show up on Super User.

And sysadmins tend to talk to their LUGs for feedback, or spend uncountable amount of time in man pages or reading InfoWorld.

I'm active (to some level) on all three sites, but find the Server Fault one to be most useful, as it's what I do for work most.

I also think (but have no statistics to back this up), that there are substantially more "programmers" than there are sysadmins - how many homework questions are on SO, vs. either of the other two? (1589SO vs. 16SF vs. 8SU.)

Lots of folks have to do programming for school/work/etc. and then move into something else later.

Personally, word-of-mouth communication is how I think this is going to grow most.

Answer 10

I totally agree with Josh, but let me add to his answer.

I didn't find out about SO until the day I noticed those links on the side of your blog (which BTW are kind of small (also why isn't your flair on there?)) and I probably never would have otherwise.

SOLUTION: Advertise! I would suggest starting with sites that have high traffic that are geared toward computer enthusiast and SysAdmins. How about Everything SysAdmin for SF and ExtreamTech for SU.

Seriously make those links bigger on your blog, and maybe show the favicons.

Answer 11

There are already some well known users on Server Fault.

For example the "Standalone Sysadmin" Matt Simmons

---

He has also put up a blog post that he has put up an OPML file of his Sys-Admin subscriptions.

Answer 12

Regarding SF, have you tried working with/cross promoting the various professional bodies, like SAGE, LOPSA, etc. (list here)?

Answer 13

House ads would help, but I suspect that part of the problem is the nature of those audiences.

---

SF:

It has been my experience (which may or may not be typical) that more Sys admins are of the "I had to learn it on my own, so you need to as well" and/or "I don't want to talk about x lest someone use it to penetrate the system" - resulting in the Admin community being more a scattering of islands with a few kind souls trying to connect all the little silos of information.

Contrast that with the programmer community, where sharing and mutual growth are expected and encouraged - and you see where you'll get more support for a community Q&A site.

You're probably getting a lot of programmers (like myself) who end up doing double duty as the server admin and the programmer.

---

SU:

It seems we haven't really tapped into the real SU community as (from a semi-regular visitor's standpoint) we're getting a lot of "tier 1" type help desk questions - not a group of enthusiasts that like computers, are not "professionals", but have a desire to learn more about computers, scripting, etc.

Don't hear what I'm not saying, they are there, just not enough of them. As others have suggested, audiences like Lifehacker and Gizmodo would likely be good places to post a few ads and/or contribute a few "guest" articles to bring some attention to the sites. The latter might be better received.

Answer 14

It's forum fatigue (community fatigue?)

If they were just sub-forums of Stackoverflow you wouldn't really have that problem.

Answer 15

Specifically relating to serverfault:

As has been mentioned, reach out to some members (and non-members) who have well read blogs. Matt Simmons is an obvious candidate; Tom Limoncelli is someone else who's generic 'sysadmin' blogs I read.

Now on to an issue that may be hindering serverfault; not sure how to address it, but you should be aware of it. The term 'sysadmin' covers a far wider variety of disciplines than does 'programmer'; and the relative breadth and depth of knowledge varies widely between the <10 user office admin to the 3rd line network engineer at AT&T (for example).

From my observation at work (more on the 'enterprise' side of the scale) most infrastructure departments tend to be pretty siloed, particularly as the size of the organisation increases. You have your network geeks, your storage geeks, your database geeks, your *NIX geeks... the list goes on. Each of these groups probably frequent a different set of forums and blogs, with little overlap. You may have better luck trying to engage each of these communities separately? There are several excellent 'jack of all trades' on serverfault already; is it worth now trying to court the specialists?

Answer 16

Try reaching out to the IT Pro Evangilsits at Microsoft (there's only 3 of them) and see if they have any suggestions as to IT Pro bloggers you can talk to. I've got contact into for the west coast guy if you don't have you local IT Pro guys info (your local MSDN Evangilist can hook you up).

Answer 17

I've been thinking about this as well. I'm more in tune with the Microsoft side of things and was thinking that for SF enlisting MVP's would be a good start. There are a number of well-known MVP's already participating on SF (indeed, even one of the mods) who provide a wealth of great, advanced knowledge. Would be great to get more, and get them to help evangelize.

Answer 18

I think DLux got it pretty much right. The exchange administrators are relatively easy crowd to please, you'll know SF is good when the network engineers accept it. As they are relatively comfortable with Nanog, vendor-nsp mailing lists, irc and other relics, I guess it'll take at least few years. Even if SF starts going to the right direction today.

However, as the good and interesting questions are scarce and the crowd really doesn't appreciate them or the good answers, it's pretty hard to imagine how SF wouldn't self-select itself into uselessness.

Some people have suggested advertising, and I would have to say it wouldn't help. To make it work and gain visitors who would actually return, SF would have to be a valuable resource that just needs to be found. The value is not there at the moment. When someone has a SF-related problem, Google usually finds some obscure blog or years-old mailing list post, but seldom anything from SF.

Big part of this can be attributed to the fact that old stuff runs the world. It's not uncommon to have to work with decade old hardware and even older software. With programming, the pressure is towards using the most productive tools. With systems and network engineering, the pressure is towards causing as little capex as possible. SO crowd has questions about new stuff, the intended SF crowd has questions about antique. Because of this, it will take way longer for SF to find its way to the first results of sysadmins' google searches.

When it comes to finding the "Joel Spolsky of sysadmin bloggers", the fact is that there is none. Actually, Joel isn't really a "programming blogger", I'd say he's read more because of his business orientation than technical content. Sticking to technical stuff limits the readership. To have in-depth knowledge of some platform or technology usually means you don't have broad knowledge about related stuff and optionally you work for the vendor. And the field is even more fragmented than programming. On the other hand, if it was possible to attract the active niche gurus to SF, they could provide really great answers to hard questions.

As a quick fix and easy improvement, there really should be links to SF and SU on the "ask a question" page. It should probably be in the sidebar, with brief, clear description.

Answer 19

One major difference to keep in mind is that while all three fields are changing rapidly, only the programming field is changing due to its own users. Further, software can be rapidly prototyped in the comfort of one's own home with no special skills or products. If you want to try out a new RAID product, you have to buy the product.

Or, in other words, the consumers of programming expertise are also the producers of that expertise. New languages every day. New methods, techniques, etc.

The world of servers and users are supplied by companies that may or may not listen to users.

It's a cathedral and bazaar situation, in a small sense. The company that produces the product is the expert, but when one can't get help from them it's hard to find other users that also have the same products, have experiences the same issue, and have the expertise to help.

It's also due partly to how small the problem space is. Programming is large, but in comparison a single server product from a single company can generate as many questions and problems as a small programming language. Multiply this by 500 products and again by 50 companies, then add interoperability problems and you find the problem space dwarfs programming. Programming is a huge space to begin with, but you can almost always find an expert that can understand and help solve your problem because the cost of entry is so low in terms of both time and money.

I believe it's endemic to the fields. Given that new hardware replaces old over the years (whereas many programming problems are timeless in nature) the existing questions will cease to be useful at nearly the same rate that questions are entered. This means that even time won't result in a problem corpus large enough to sustain the sites.

About the only thing you can add to the serverfault site that might help is vast, deep, timeless, general server principles. Suck in the newbies and keep them around to solve the later problems they encounter that aren't general purpose knowledge. Some of this content may or may not already exist, but the key is that whenever someone does ANY server information search on google, serverfault pops up in the first page with general principles. A Server University, of sorts.

I'm not sure the same can be done for superuser - it may only ever be the 'sidecar' to Stackoverflow where programmers ask and answer each other's non-programming questions.

-Adam

Answer 20

Apart from just building a fantastic website and community (which you already seem to have done), what about advertising (responsibly, of course) on other websites? I'm not too sure what websites you would want to advertise on. Reddit maybe?

As for Server Fault, what if you tried to form some kind of 'partnership' (or agreement) with some hosting companies to provide a place for technical resources and other content that may be beneficial for their customers. Of course, the downside is that you do not want Server Fault to be seen at the support website for a particular web hosting company. What Slicehost did with Slicehost Articles is very good, and it could be used as a source of inspiration as to what to do.

Answer 21

Many of the questions here seem to be focused on advertising.

I think one thing you could to on this front is improve flair so I can actually link to my trilogy profiles from other social networking sites. The badges and flair are a nice idea, but right now they are pretty much locked up on the trilogy sites unless a user is running their own web server.

Right now I cannot use flair in most places because wherever you post it either has to allow iframes or javascript. Both are blocked by pretty much every social network and hosted blog on the Internet. See flair as image.

I am aware there are scripts available that I could download and use, but I don't think this is something a user should have to do. I am thinking image-based flair should be supported officially. Or come up with some official way I display my flair and link to the trilogy sites from popular sites like facebook, myspace, blogger, wordpress.com, linkedin, plaxo, etc.

I think you should be looking for other ways to tie in to other social networking sites. Perhaps make it more obvious about how a person can have their activities on trilogy sites show up on their wall.

Answer 22

I wonder how much overlap there may be with the /. crowd. StackOverflow has gotten some press there, but I don't see any articles in my RSS feed about ServerFault or SuperUser.

Perhaps advertising over there, or coming up with something to grab their attention enough to have a post about either of them.

On the other hand, I could see a rather strong argument that the level of user that would likely be found there would be... uhmm... given to the wrong sorts of behaviour.

Are there other communities with better sound/noise ratios that you could advertise to?

Answer 23

I guess we have to find where are they now and what do they want.

As it happened with programmers? You knew we were spread for in a number of websites dealing with problem of finding a good answer that were both accurate and up to date. ( Across msn, the hypen site, and many other forums )

I think this process was not performed for sysadms and powerusers, what are their frustrations, where are they getting ( or trying ) to get answers right now.

EDIT

heh heh , yeeap I agree with the rest. Once you find out what blogs are they reading, Advertise there, I got here too by codinghorror in first place

Answer 24

90% of your traffic comes from Google, right? What I think you need to do is figure out which topics people are Googling to find your site, and then do a push for more of those kinds of answers, similarly to what the R (statistics package) folks are doing right now on SO. If you can develop content on under-served topics, it should certainly increase your Google traffic.

The follow-up question would be, thus increasing this kind of hit build community, or just page loads? I'm not sure.

Answer 25

As bad as it may sound, I think you just have to keep doing what you're doing.

I think the potential audience size for Server Fault is larger than that for Stack Overflow; the potential audience size for Super User is practically every human who uses the web. As a result, there is a lot more "noise" already out there on the web for both of these groups -- tons of scattered pages with "tips" for computer use, or half-finished notes written by admins to themselves; put up on the web because, well, why not?

It will take a while for the "signal" from these sites to start to attract more attention.

If and when Super User reaches critical mass, though, ...

Answer 26

Ads on tomshardware and similar sites would probably help somewhat. The fora on computer help are fairly extensive, I'm afraid.

You could also do Dead Tree advertising and try to bring in maw & paw who have questions and don't know much.

Perhaps you could hire some techs from local repair shops to post their FAQs & answers on SU.

Answer 27

Have you given thought to how you might poach some of the IT audience from Experts Exchange? They do have a fairly large community, after all. Do these people even know ServerFault exists?

Look at things like how they advertise, who they connect to. Who are the "rockstars" there? Maybe even join for a couple months (it's free for "experts") and troll the IT support areas to see where they hang out, what ads Experts Exchange shows them, etc.

I'm not recommending approaching their members directly. That would be sleazy. But it might be useful to get to know some of them just to get ideas on how to reach their peers.

Answer 28

Well, lets summarize: we need advertising to "reach out to the power user and sysadmin communities".

What about a "referrals" system? I reference some people, and if they register because of my reference, I earn X reputation points (and possible badges) - and so on the person that registered. There are a lot of wesites that do this kind of thing. I thought about it after I used Dropbox and remembered they have this approach.

This way you don't have to "pay" for advertising, you're using the users that think referring someone it's worth. Of course there would be a lot of statistic rules need to avoid people gaming the system, but I don't think it would "harm" the system at all.

You don't pay these people with advertising money, you pay with visibility in the community itself with reputation points.

What do you think of this idea? Would this "cannibalize" the reputation system? And, most importantly: would it be effective?

Answer 29

... If no one else has said it, You can always send me more stickers for the events I go to!

I value greatly the ones I received for being a high reputation user, but would never give those out as I only have a few... Whenever I go to any event, I am always telling people about Super User and Server Fault, however freebies that can be given away always speak louder than recommendations!

(Being serious... or, if I can have the design, I will see if I can get any printed cheap)... At the end of the day, I am on SuperUser.com all the time... refreshing when I am bored etc... It is annoying to see no new questions or the same ones cycle all day! It is amazing to see Stack Overflow have ~10 new questions/answers a minute and would love to see something similar on SuperUser!

Answer 30

My suggestion (partly stolen from Podcast 76).

Let's get good questions.

Let's (the user community) find a lot of the questions people are searching for on the internet, that are spread out across the internet on forums and the site that will not be named, rephrase them so that they're good, answerable questions, and then post them to the appropriate site.

Suggestions for high value questions:

• SuperUser
• Firefox customizations
• Windows performance improvements
• The meaning of windows error messages
• Overclocking and performance enhancements
• ServerFault
• Virtualization
• Multiple server management
• Database Management
• Backups

Let's do it.