We build high quality, affordable websites

Can Stack Overflow Grow Beyond Programming?

[Reminder / Disclosure: I work at Answers.com, which is in some ways a competitor]

Congrats to Joel Spolsky on getting funding from an all star team of investment bankers. Having this combination of great entrepreneurs and great VCs certainly increases this companies’ chances. But the company’s current plan seems to based on two (long shot?) assumptions both coming in:

Stack Exchange

Stack Exchange

  1. Stack Overflow’s success in building great content and community will be replicated to many other areas, even though so far their efforts here have failed.
  2. The expert content and community will one day bring significant revenue, even though so far their efforts here have failed.

Stack Exchange Sites

“Over the course of the 6 months we have found a lot of successful sites have gotten created, my favorite is mathoverflow.net for PhD level mathematicians. …  It’s a really awesome site, get’s a lot of traffic … It’s an incredible resource for mathematicians that basically erupted on the StackExchange platform in about 2 months, that it went from 0 to pretty much all the mathematicians on the internet are on there right now.”

– Joel Spolsky, Mixergy: Why Didn’t Stack Exchange Work?

OK, getting all the mathematicians on the internet is pretty cool. Here are the numbers from Alexa:

Stack Exchange Sites

Stack Exchange Sites

MathOveflow.net is that blue line there at the bottom, sometimes peaking above the level of 100,000th top site in the world.

Assuming Alexa’s numbers are roughly accurate here (I’d use Quantcast, but MathOverflow.net doesn’t rank high enough to get a listing there right now) then at least one (and possibly both) of the following are true:

  1. Spolsky’s a little off in his estimate that they have all the mathematicians on the internet.
  2. “All the mathematicians on the internet” is a very, very small market.

Not very encouraging that this is the one site he points to as their biggest success in all the interviews and posts I saw from him on this subject.

ServerFault and SuperUser have had more success, but they’re almost spinoff sites to Stack Overflow, going after neighboring markets that overflowed from SO’s natural user base.

Key Factors Behind StackOverflow’s Success with Programmers

“The hardest thing about making a new Q&A site is not the programming—it’s the community. You need a large audience of great developers so you have the critical mass it takes to get started. Without critical mass, questions go unanswered and the site becomes a ghost town. I thought the combination of my audience (#15 on Bloglines) and Jeff’s (#89) would bring enough great developers into the site to reach critical mass on day one. So Jeff and I decided to go in together on this.” — Joel Spolsky, Stack Overflow Launches

A number of key factors helped Stack succeed with programmers:

  1. Spolsky and Jeff Atwood had very large tribes of programmers who followed them here. Generally speaking, these community members:
    1. Were committed enough to the art of programming that they read the industry’s top blogs
    2. Had been reading Spolsky and Atwood’s great posts, and were now getting a chance to both give back and show off their knowledge to their mentors and their community
    3. Already felt like a community, and already acknowledged these two site owners as their leaders.
  2. Programmers generally:
    1. Do their job by typing on their computers, usually while on the internet.
    2. Are comfortable learning new sites and programs, and using the Web to share information. Many programmers I know prefer to get their information from virtual friends on the internet than from the guy in the next office.
  3. Programming questions are generally of little interest to anybody who isn’t a professional programmer. Therefore programmers generally:
    1. Don’t make a living by charging for their advice (like most lawyers, for example)
    2. Don’t get hounded for professional advice while out with friends (doctor, I’ve been having this pain …)
    3. Can run a Q&A site where a very high percentage of the people are their peers, asking and answering each others’ questions. By contrast, Spolsky discussed building a site for professional gardeners. It would be very hard to stop that site from being filled with people like me asking stupid questions about my plants.

Take another example Spolsky suggested, Home Improvement:

  • Instead of Spolsky and Atwood you could start with Bob Villa and Tim the Tool Man Taylor (the Buzz Light Year of Home Improvement). Or whoever the equivalent bloggers are in that industry (if such people exist). You may need to give them significant equity in this new site so that they’d bring their large online communities in, and you’d need them to hang around and mingle for a while .
  • Most home improvement professionals aren’t typing on the computer while doing their job. They’d generally need to answer these questions at night from home.
  • Most questions would come from non-professionals. And eventually, probably most answers as well. Experts would more often visit the site looking for clients than expecting to spend quality time with their professional peers.

If you tried to do an SEO Q&A site (as has been suggested) you’d at least be starting with people who are online. Then you could try to get people like Rand Fishkin, Aaron Wall, Danny Sullivan, and Michael Gray to bring their communities on to your Q&A site. Right, because that’s why they built these communities, so they could give them to you. And then they’d have to hang around and help lead your community. And even if you could pull that off, the SEO industry is filled with consultants who get paid to share their information, and who build web sites for a living. Good luck getting their readers to volunteer their time to building your enterprise.

Last example I’ll give is VCs. Fred Wilson, Chris Dixon and the other Stack Overflow angels already have an interest in the company. Maybe they’ll commit themselves to making AngelOverflow work. But I’m guessing that VCs’ audiences fit the pattern of most audiences on the web. Most of Fred Wilson’s readers are probably NOT fellow Angels, but rather entrepreneurs who find Angels and their information fascinating and potentially useful. The programming world is a fairly rare exception, because only programmers care about most programming questions. On AngelOverflow, I’d be asking the questions, and I’d love to see how long Chris Dixon sticks around to answer my questions. It’s not a symmetric community. It’s an expert / celebrity and his audience / fans.

Happy to hear what subject areas you think are going to work. I haven’t heard any that I think have a good shot, other than the awesome MathOverflow site that “gets lot of traffic” and has “pretty much all the mathematicians on the internet.”

Stack Exchange Revenue

Well, it’s 1999 again. Stack is going to accumulate eyeballs and assume there’s going to be a way to monetize them.

Within a few days of Stack’s announcement that the paid Stack Exchange 1.0 failed and they were going to pivot to a free model:

  • Ask.com’s parent IAC announced that they were shutting down FiLife, the free Financial Q&A site they had been running with Dow Jones.

It was interesting that at the same time top angels took the leap of faith that Stack Exchange would be able to turn eyeballs into money.

Thanks to Jeff and Joel for documenting their monetization attempts, and to William Shields for gathering these in his excellent post about Stack’s failed efforts at monetization:

My favorite is Jeff’s comment “well, we have earned $1.16 so far. I think I will go buy a twix, and we can share it.”

But, we’re going to party like it’s 1999. I hope this one ends better.

What Comes First, Quality or Quantity?

Six months ago Joel had a soul searching moment when he asked if slow growth = slow death, discussing Ingres’ battle with Oracle.

“Executives at Ingres meant well. According to [Geoffrey] Moore, they felt that the company “simply cannot grow any faster than 50 percent and still adequately serve our customers. No one can. Look at Oracle. They are promising anything and everything and shipping little or nothing. Everybody knows it. Their customers hate them. They are going to hit the wall.” Of course, Oracle overcame those concerns and eclipsed its rival. And this got me worried. Were we Ingres?” — Joel Spolsky, Does Slow Growth Equal Slow Death?

This story may be part of what drove Spolsky to raise venture capital to grow Stack Overflow more quickly. And it may have changed some policies regarding his other company, Fog Creek. But he remains smugly confident that he’ll conquer the Q&A market one niche at a time, while focusing on high quality at moderate quantity. He mocks sites that are taking the Oracle approach of trying to grow very quickly and then scampering to get the quality to catch up to the quantity.

Other Issues

Stack is also betting that verticals will be more successful than horizontals in the Q&A market. That’s what most of the smart people thought about search too. But building these individual communities almost from scratch is going to be a tremendous challenge.

Stack is betting that their model of democratically choosing which sites to build next will result in the creation of the sites that will succeed in terms of community and revenue.

Closing Thought

Joel is leading a great team of entrepreneurs, angels, and software developers. If anybody can make this model work, it’s them. But IMO that’s a very big if.

Am I missing the boat here? Please let me know your thoughts.

  • Jonathan

    Excellent analysis and amusing as always.

  • http://managinggreatness.com Gil Reich


  • Gil Reich

    Hey Jeff, thanx for the info and the insights. You’ve been so forthcoming with details about failed monetization efforts, and so vague regarding monetization successes (unless I’ve missed some posts, which happens) that the only hard facts we have to work with are those that indicate troubles.

    Regarding AdSense, it works very well for some sites. To take Joel’s example of a gardening site, even if the content is largely provided by professionals, if the site develops significant traffic from hobbyists coming from search engines, I suspect you’ll want to try AdSense again, which will often hook up these hobbyists with relevant opportunities. You can always not show these ads to the site’s members / contributors (that’s what we do at Answers.com).

    Anyway, good luck Jeff.

  • http://codinghorror.myopenid.com/ Jeff Atwood

    A few clarifications:

    - what wasn't working was AdSense or anything like it (eg, auto-picking Amazon book ads). AdSense is such a broken ecosystem, I don't know how it works for ANYONE, honestly.

    - what DOES work is old-school “guy on the telephone talking to people and building relationships” ad, job, and careers sales. Essentially what we are repudiating here is the idea that software algorithms can be your sales force. Sales, at least for our market, is a high touch business involving real human beings talking to other real human beings. (you'll note that Joel has gone on the record saying many, many times that Fog Creek, though successful, has historically failed utterly at hiring salespeople — this is a mistake we are NOT repeating with the new company. Quite the opposite in fact, it's a primary focus of the NY team.)

    - yes this does imply that sites that have “professionals” are easier to support beyond ads alone. Not exactly earth shattering news to you, I hope, but we'll have a mixture of some sites that fit this mold and some that don't.

    - we were (and are) actually making quite a bit of money through jobs, careers, and ads. I can't disclose specifics, you understand, but we were comfortably profitable with real livable salaries at staffing levels we had prior to the VC deal. And that includes outsourcing ad sales at a hefty premium (which will change rapidly).

    - in all honesty I was never particularly fond of the SE 1.0 model — “pay us $129 a month and whatever happens, happens!”. Joel's simple venn diagram here http://blog.stackexchange.com/post/518474918/st… captured the myriad problems with that approach, so I won't belabor it.

    - The new SE 2.0 model is much more akin to what I am a fan of, and what we were already doing: curating sites that progressively take baby steps farther and farther out from our core audience, sort of how Twitter started with “techies” and after 5 years got fairly mainstream. We plan to democratize this process so everyone has a say in, and tangible stake in, the sites that get created. We'll run them in service of and in harmony with the community and the greater internet.

  • http://managinggreatness.com Gil Reich

    Hey Jeff, thanx for the clarifications. I'm glad you're making some money through those things. You were so forthcoming with specifics regarding your monetization efforts that failed, any chance you can give some specifics on the ones that are succeeding? Could you also clarify what “tangible stake” means? Does that mean revenue sharing?

    Regarding AdSense, it works quite well in some cases. I expect for example if you do that site for gardeners that Joel mentioned, and you get a lot of visitors, that AdSense would work quite nicely. Anyway, good luck!

  • Pingback: Stack Overflow 2010: What’s Working, What’s Not

Seen Speaking at:


SMX East

SMX Advanced

SMX Israel

SMX London

SES Toronto