As a primer to tomorrow's luncheon on "The Dynamics of Collaborative Innovation: Exploring the tension between knowledge novelty and reuse" with Karim Lakhaniand Ned Gulley, Berkman intern Zack McCune sat down with our guests to discuss collaboration in the age of computing "2.0", distributed innovation, and motivations for participation.
What does collaborative innovation mean in terms of software development?
NG: People working together to make software. It rarely happens any other way.
KL: The myth of the lone genius programmer wrestling the compilers demons in a death match and ultimately triumphing, is just that, a myth. Beyond toy programs, most software development is a collaborative endeavor. This should be viewed as the base case for software development.
Do you think that the contest environment (in which everyone is motivated to create the best software solution) would apply in real life?
NG: Yes. The contest environment was purposefully modeled on real life, albeit a souped-up overheated version of real life. But I see this kind of thing happening all the time in real life.
KL: Most real world settings have an orientation towards performance, for example, how do we judge good code from bad code. The contest environment creates a context where participants know that they are going to get explicit and comparative data on how their submissions ranked in relationship to others. This promotes effort, I people try harder, and results in both learning and innovation.
What are the motivation factors for programmers? Did anyone have ownership issues?
NG: Everybody has ownership issues, but for most people these are easily by the returned value on the collaboration. Some people had rather severe ownership issues and complained loudly. But you either get over that or you stop playing. I know for a fact that some people don't play simply because of this ("why should I bother contributing when someone is going to steal my code?"). On a "real" software project, people like that generally cause more trouble than they're worth.
KL: My studies of programmer motivations to participate in open source software development http://mitpress.mit.edu/books/chapters/0262062461chap1.pdf have shown that both intrinsic motivations (having fun, learning, enjoyment, community belonging) and extrinsic motivations (job, money, career, reputation) are important to programmers - and that there is no one major driver for participation. So programmers participate for diverse reasons that resonate with their particular circumstances and personalities. I don't believe that there is one common driver for participation.
What kind of problems (or not) would such a system face when it comes to compensation?
NG: If monetary compensation is perceived by the participants as surpassing the reputation-based rewards, it would severely warp the way the game is set up now. That doesn't mean a rewards-based system wouldn't work, but the contest would have to be re-worked significantly to support that. The more money is involved, the more the concept of "fairness" is invoked. Fairness is simultaneously considered very important and at the same time it is essentially mealy and unknowable. Firms manage "fair" compensation by letting senior people make somewhat arbitrary decisions. Open source projects dodge the money question, making the problem of fairness more distributed and less nettlesome.
KL: The underlying assumption is that people need direct monetary compensation for participation. But the last 10 years of the Internet (for example the hundreds of thousands of reviews on Amazon and Yelp) - and most of history has shown is that indirect rewards are equally as important and if not more. Two other elements also need to be considered: 1)Differing "cost functions" and 2) Modularity and granularity of work.
The diversity of motivations to participate also creates differential cost functions for participation. Many people simply enjoy the task of programming and get significant fun and joy out of participation. So how do we think about compensation - when work = fun? Typically people pay for fun activities - going to movies, buying sudoku books etc. So maybe we should be charging people to participate!!! More seriously, at the individual level, we all face different cost curves - i.e. willingness to spend time on an activity depends on how much we value that activity. Many of these distributed innovation settings attract people who actually value the activity for its own sake and not necessarily consider time spent to be very expensive.
Tied to lower costs for participation is the notion that the work itself comes in smaller chunks (Carliss Baldwin, HBS and Yochai Benkler, HLS have written about the importance of modular and granular work components and tasks in open source). So that the level of effort exerted over any quantifiable piece of work is actually quite small. This also mitigates in many cases the question of compensation.
What do you think collaboration in the age of computing "2.0" says about individualism? Can collaborative spaces actually produce new types of individuals, or do these systems obscure a traditional concept of individuality in favor of new community-oriented individuals?
NG: Individuals are the people forming these 2.0 communities! Communities don't drop out of the sky. I would say that 2.0-style collaboration empowers the vast majority (the masses) tremendously, people, as Clay Shirky tells us, that would otherwise be spending their time watching sitcoms. At the same time, it enables people with ideas and dedication to be at the center of the hub, motivating and benefiting from the efforts of the peripheral contributors. It's all about leverage. The coming strength is that of formulating problems well, of orchestrating the swarm. There are plenty of bees, but not enough flowers.
KL: Ned is right on here. Communities are formed from individuals. We have always had communities. One of the strengths of the American Culture has been its voluntary associations. So in that sense there is nothing new under the sun. As Ned notes - via Clay Shirky - is that the cost of organizing via communities has dropped significantly over the last 10 years.
As a follow-up, do you find that individuals feel less responsible for their contributions to collaborative ventures than they might for personal work? And might this incentivize (or perhaps uninhibited) individuals to work harder?
NG: Work at a firm is often masked, as when you work for a year on software that is never released. That's horribly unmotivating. Collaborative publicly accessible work is completely unmasked. If you're making a difference, you know it. And that's the thing we all want to know, do I matter? So it's fantastically motivating to people who can see they are making a difference. You get a huge boost to productivity for those people. But there are still plenty of people on the fringes who play a little and can see they're not moving the needle. But they're able to put themselves into perspective rather than being told by a manager that they're no good. We often see people who lurk during this contest roar to life in the next one.
KL: Countervailing this free rider problem is transparency. These collaborative ventures provide very significant accounting of the work - so that people cannot claim credit for things they did not do. The transparency can be quite scary - as there is a public record of all the work that has been done. Some individuals do not like that - others thrive in these conditions.
Are community, collaborative situations more likely to be innovative than individuals working alone? Is there really a 'maverick in the masses' so to speak?
NG: The mass is made up of mavericks. An important point about this contest is that there is no social smoothing: no voting or popularity contests. Those things can drive away the brilliant lunatics. They happen in plenty of 2.0 applications, but they aren't a feature here. So if you're repugnant and eccentric but you've got a good idea, we'll take it.
KL: There are is some empirical evidence and conjecture that average groups can outperform super smart individuals. I am working with Ned and other colleagues (Kevin Boudreau) to actually figure out some experiments to see if that really is the case. Stay tuned. My hunch is that this will depend on the nature of the problem to be solved and the relevant diversity of individuals in groups and the knowledge of the experts.
Is it appropriate to be collaborating in studying collaboration? How has your partnership been ideal (or less than ideal) for critically engaging the way innovation can come from individuals working together?
NG: I'm running out of time to answer all these questions, but I would like to say that it's been great working with Karim. Our work has certainly been collaborative (as only seems fitting) and we have relied on our complementary backgrounds to fill out the ideas presented. I've learned a lot about research, and Karim has been extraordinarily generous in opening doors and making introductions for me.
KL: Ditto for Ned! Actually another collaborator of mine, Bob Wolf (article above on motivations) - pointed me to Ned's work - via another collaborator of his! So this collaboration was borne out of my other collaborations. I have learned tremendously from Ned and feel fortunate to have him as an intellectual colleague. Plus he is just ultra-cool.
How is collaborative innovation related to "crowd sourcing"? Is the collaborative innovation always at the service of its administrative power base?
NG: No. That topic is the basis for my current research. Crowdsourcing derives power from the crowd, but it can also provide power to the crowd. I think this will be a big shift in viewpoint over the next few years.
KL: Many people think that crowd sourcing is some magic pixie dust that they can sprinkle on a community and viola - all this free labor and innovation happens. Nothing is further from the truth. Engaging in collaborative innovation with outsiders is a mixed blessing. For example, a firm, working with just its staff has full freedom in the marketplace to take whatever actions it wants to maximize profit. If now this same firm has a significant "crowd" working with it and helping to create its products - then the crowd will expect a say in its decision making process as well. Hence the crowd is both liberating and constraining at the same time. See this article in Inc Magazine about Threadless.com and how they work with the crowd to get a sense of benefits and costs of working with the crowd. http://www.inc.com/magazine/20080601/the-customer-is-the-company.html
Addressing Karim's Business School foundation, what might your findings mean to business and entrepreneurship?
KL: I think most companies are just now becoming savvy to Joy's Law: "No matter who you are most of smartest people work for someone else." Many executives and entrepreneurs are now realizing that knowledge and innovation are widely distributed in the world and that the biggest challenge facing most firms is how to access this knowledge. Collaborative and distributed innovation systems show one way of accessing this knowledge. It is quite clear that open source software had forever changed the landscape of software development - and legitimized distributed and collaborative innovation. A similar realization has now happened in the content business with the rise of Wikipedia, YouTube etc. The key question is where else? I am currently advising a firm Local-Motors.com that wants to take these insights and apply them to automotive manufacturing! There are many experiments going on and but the trend is clear - more and more industries and start-ups are embracing distributed innovation.
How is collaboration in programming tied to the spirit of free and open source software? Did you discover that a certain quality of open source ideals were present in the participants?
NG: Yes.
KL: Ditto!