This morning’s keynote presentation is by Siobhan O’Mahony, who worked as a consultant for Price Waterhouse and EDS before joining the faculty of the Harvard Business School. She began by mentioning a theme I heard last week in the ITechLaw conference in San Francisco: there is no one open source model. Choosing the right model enables success, and choosing the wrong one can lead to failure.

An interesting statistic: Red Hat software spends about $40 million on R&D, and is able to tap into a $1 billion global R&D effort on Linux. While that’s an obvious financial reward, Siobhan made it more general: she argues that it moves the IT function from efficiency enhancing to value creating.

There are three meanings of “open source”: a type of license; a type of software development methodology; and a type of business model. The latter one is interesting: it either “embraces” or merely uses the open-source license and/or the development methodology. Some companies (and governments and transnational organizations like UN and EU) are starting their own open source software projects. Other organizations are experimenting with elements of the OS model in new contexts. Still others adopt just the legal elements, while others adapt the “community” element.

Siobhan says the open source movement has matured to the point where there’s a new collaboration model: community managed projects, which, among other things, provide some limited rights to non-profit foundations who participate in the effort. In addition, more traditional firms (like IBM) sometimes participate, in order to donate resources, donate code, hire/support individual contributors, research market/customer needs, supply complementary software, and hardware resources. Underlying all of this is the Free Software Foundation, and the Open Source Initiative.

There was an interesting discussion of “sponsor founded” open-source projects, such as Mozilla and OpenOffice. These are founded by an incorporated sponsor, often spun off from a proprietary project. There are actually several examples: Apple’s Darwin, PHP, and several others. There are both similarities and differences with the “organic” model of open source. Sponsored open-source projects are more likely to have sponsor-controlled ownership rights; and key governance decisions may be more controlled and less open. As an example of the sponsor-founded open-source project, Siobhan reviewed the Eclipse project — which was founded in November 2001 with a $40 million investment by IBM.

Siobhan then moved on to discuss the open-source process as a mechanism to foster collaboration. This can be either “inbound” (within a company, to foster idea generation and improvements to existing ideas), or outbound (to foster second, generational, spin-off or complementary products to help move ideas from marginal to mainstream). One example is Eli Lilly’s “Innocentive” program, which matches 30 companies who post problems in chemistry and biology (these are “seekers”) with 90,000 scientists who are registered as “solvers.” The seekers are anonymous: you don’t know which companies have the problems. Innocentive helps post the challenges, and offers rewards of up to $100,000 to those who solve the problems. (I haven’t figured out who provides the money, but hopefully that’s a small detail!)

Another example is Procter & Gamble’s “Connect and Develop” program, which plans to have 50% of its innovations sourced externally by 2008; they do this by tapping into an army of 8,000 researchers (including alumni and retired employees), 600 partners, and 5 business units. Their R&D productivity is up 60%, and 35% of their innovations now come from outside the company. The percentage of sales invested in R&D is down from 4.8% to 3.4%.

Still other organizations are adopting the legal elements of open source. An example is PatentCommons.org, which consists of 500 software patents released by IBM. Other corporate contributions come from Nokia, Computer Associates, Sun, HP, Novell, and RedHat. To illustrate that all of this involves a larger universe than just software-related initiatives, Siobhan provided several examples of open-source projects in the field of biology and biotech. Part of this presentation included some startling statistics: 20% of the human genome is owned by patents, and 70% of biotech patents are owned by a mere six companies. She also reviewed the “Creative Commons” initiative: you can reserve certain rights: attribution, noncommercial use, no derivative works, “share alike,” etc. — all of which can be applied to music, art, photography, etc.

Siobhan reviewed Wikipedia as another example: 19% of the global Internet audience visited Wikipedia sometime in March. It contains 1.1 million articles in English, and had 131,949,000 unique visitors in March, 2006. Nature magazine compared 42 science articles and found 162 minor errors in the Wikipedia entries, and 123 minor errors in Encyclopedia Britannica.

Another interesting statistic: 87% of 512 companies surveyed in Aug-Sep 2005 use open source in some capacity; 74% use Linux and 70% use Firefox.

Siobhan finished her presentation with a summary of when it makes sense to use an open-source approach, and when it makes sense to pursue other, more traditional approaches to development, innovation, and collaboration. She had a wonderful example of a real experiment: the open-source cookie, in which a food company tried to create a new cookie with healthy attributes, using an XP, “managed R&D,” and “open source” approach. Interestingly, the XP approach did worst, but it was also interesting to see that the managed R&D approach did best.