Cambridge was treated to a fascinating talk last night by Joel Spolsky, thanks to the combined efforts of Red Gate Software, BLN, ideaSpace and the Cambridge Network. We’ve been lucky of late with the quality of speakers at such events, due in good part to an overlap in organisers with the Business of Software conference held annually in the USA.

Joel Spolsky is the CEO of Fog Creek Software, a company that makes tools for software project management. It’s a successful company and their web site attracts in the region of 10K unique visitors per month. The tools they sell are well-regarded, but if Fog Creek is famous for one thing it is their egalitarian approach to employee management which is often cited when people discuss the optimum software development process. It’s the antidote to the Cube culture that you’ll have read about in Dilbert cartoons.

But Joel is also associated with another venture, known as Stack Exchange, which “goes to market” under different brands such as Stack Overflow. The notion behind Stack Exchange is to set up a network of free, community-powered QA sites. It currently has over 80 such sites, and Stack Overflow is the most visited. It’s a venture people associate equally with his co-founder Jeff Atwood, who by the way is leaving it this month to pursue a better work/life balance for himself.

The Stack Overflow web site is a destination site for software programmers who have technical questions they need answering. It covers all the bases in terms of software platforms, programming languages, operating systems and frameworks; but for my money it’s the chief watering-hole for the .NET community. They’re the lucky people who get to write for the Microsoft environment.

I mentioned the web traffic to the Fog Creek site. The equivalent for stackoverflow.com is around 175K unique visitors per month, and that’s from the USA alone. It’s one of the top 100 sites by traffic ranking.

There’s an Entrepreneur / Startup story worth hearing right there. What is the strategy behind the creation of an adjacent business which becomes more successful (as measured by web traffic) than anything you do in your core business? And how do you gain from the success? The tactic with Stack Overflow seems to be promotion of their sister site, Careers 2.0.

Last night’s talk was nothing to do with those alluring questions. One for another day.

Instead, Joel talked about the way that they set about creating a culture around the community that makes up Stack Overflow. It’s a set of people that includes their employees but the majority are unpaid volunteers who give their knowledge to others.

If I were desperately trying to re-use something from my M.Sc. in Social Psychology, I could tell you that creating a Stack Overflow web site means solving the problem of creating a common social representation; which involves understanding how we think and communicate in social groups, and how practices and processes of knowledge exchange, self and group identity, ideology and social change can affect the outcome.

I’ll try again on that: it means how do you get a set of people equivalent in number to the population of a city to get along with each other? It has 1 million registered users and many more visitors who don’t need to log-in to use the site. It’s a community that has voted more than 8 million times. It has asked over 800,000 questions and left only 16% of them unanswered. It’s a community asking 1.1 questions and receiving 2.18 answers per minute.

I suppose if you really want to know “How did Stack Overflow succeed when it was starting out?” then you should seek the answer on Stack Overflow.

But it’s more entertaining to listen to one of the founders. Joel said that it was a process of engineering the community through cultural anthropology. It involved looking at earlier efforts in computing to address communication (from email to Usenet to forums) and noting that there are accidental features of cultural systems that produce unintended consequences.

You can steer it in one direction by allowing everyone to participate through voting up/down the questions and answers they read. You make the community moderate itself and ensure that moderators run for election. You award badges to reward behaviour. You (as in the community) create a set of rules as to what constitutes a bad question, which is then instantly marked as such (but not removed).

One interesting aside was that a successful QA site has to have a “No Fun” rule. The moment that someone takes the conversation off-topic with even the wittiest of asides is the moment when the relevance of the QA starts to diminish. Another was his operational definition of a question that is too broad: it takes one sentence to ask but a book to answer. This could be a handy rule for the parents of toddlers.

Invariably someone will get upset and quit. He told us about “rage quitting” which is when someone tries to leave and delete everything they have ever entered. In British terms, it’s taking the ball home with you. They have a limit on how many items you can delete per day specifically to curtail this.

There were questions in the QA wrap-up about Gamification and the SEO best practices that make a site like Stack Overflow successful.

Inevitably the talk ended with many questions unanswered. It’s a story that still has a way to go. Their audience is over-represented by males in the 18-34 age group (wasn’t Jimmy Wales of Wikipedia talking about that guy too?) and less than 1% of their visitors currently go on to visit the Careers 2.0 site. But that’s a problem most people would be happy to have.

This is a guest blog written by Paul Walsh, CEO of web-based document management software specialist, CogniDox. CogniDox is a solution for creating re-usable repositories of knowledge. It provides internal document management for product development teams and secures publishing to external Customer web portals with access permission rights.

View the discussion thread.blog comments powered by Disqus

Want the rest of this article? Please visit original source!