The Wharton School, University of Pennsylvania

Center for Knowledge Governance, Copenhagen Business School

Road Map

• Collective Open Source Innovation (COSI)
  • Prevalent and has significant economic impact
  • Created by coordinated collectives
  • Products and services available to anyone
  • Poses a puzzle for theory and practice - Why? When?
• Research sites and methods
• Framework:
  • Non-rival Goods & Identifiability
  • Generalized Exchange & Coordination Institutions
• Implications & future research

COSI has Important Economic Impact

• Open source software has become a viable alternative to commercial products
  (Guth 2003; Lohr 2003)
• File sharing is extremely common, and arguably causes substantial revenue loss
  (Madden & Lenhart 2003; Napster court case)
• Viruses and other attacks cause wide-ranging disruptions to businesses and individuals
  (Guth & Machalaba 2003; Thompson 2004)

COSI is Created by Collectives

• Operate in coordination
• Accomplish innovative goals
• Products or services have economic value and impact
• No formal hierarchy or organization
• Little social information & interaction
• Strong orientation to accomplish goals rather than preserve relationships

Openly Available

• The products and services are freely available to anyone
• No attempt to limit access, although technologically feasible
• No legal protection, except to protect open access

A Puzzle to Theory

• Why COSI is sustainable
  How it avoids the "tragedy of the commons"?
• Why COSI is effective
  How it undermines commercial products?
• Can we predict when COSI will emerge?

Research Site & Method

• Three Usenet groups that are a clearinghouse for requests for digital music
• Users post requests and files
• Universal access, non-moderated
• Content analysis of 2,000 messages
• Semi-structured interviews

Advantages of Research Site

• Most interaction is observable
  • Little private communication
• Interaction is archived
  • Asynchronous, easy to obtain and analyze
• Goods offered are generic
  • Rules out learning benefits
    (von Hippel & von Krogh 2003}
• Identities are cloaked
  • Rules out reputational effects
    (Lerner & Tirole 2002)

Findings in Brief

• Individuals send requests for files
• Benefactor posts files in response or voluntarily
• No payment or direct exchange
• Accessible to anyone
• Free riding is common and acknowledged
• Strict adherence to established rules
• Little social "off-topic" interaction

Framework

• Non-rival good
• Perception of fairness

=> Sustainability

• Based on generalized exchange
• Coordinated through institutions

=> Efficiency

Non-Rival Good

• When one's consumption of the good doesn't interfere with another's consumption of the same good.
• Rival goods: food, clothes, housing
• Non-rival goods: radio, road, safety

Non-Rival Goods Sustain COSI

• Proposition #1:
• COSI is more likely to be sustainable when the product or service is a non-rival good.

ADD

• Low cost of replication
• Positive chance of reciprocation
• Explains why not limit access, but share freely
• But doesn't explain why people don't feel violated fairness

Perception of Fairness

• People don't want to be treated unfairly
  (Kahneman, Knetsch and Thaler 1986)
• Willing to suffer cost to punish a free-rider
  (Fehr and Gächter 2000)
• Will withdraw if sense exploitation

Violations Don't Cause Withdrawal

• Why violations don't lead to withdrawal of contributions?
• People are less willing to punish an unidentified offender
• More willing to assist an identifiable beneficiary
  (Small and Loewenstein 2003)

(Non) Identifiability Sustain COSI

• Proposition #2:
• COSI is more sustainable when the beneficiaries are identifiable, while defectors are non-identifiable.

Generalized Exchange

• Reception constitutes obligation to reciprocate to any other member
  (Ekeh 1974:48)
• Neither immediate reciprocity nor obligation to a specific benefactor
• Previously documented
  e.g. pacific islanders, CEOs, immigrant communities, lawyers
  (Malinowski 1920; Westphal & Zajac 1997; Portes & Sensenberger 1993; Lazega 2001)

Direct vs. Generalized Exchange

Gen. Exchange Facilitates COSI

• Allows contributions from unacquainted others
• Lets neophytes ("newbies") participate
• Cuts on cost of bargaining and transacting,
• Especially fit for knowledge and intangibles

Gen. Exchange Facilitates COSI

• Proposition #3:
• Collective open innovation is more efficient when based on generalized exchange rather than direct exchange.

Coordination Institutions

• Collective develops routines and defines acceptable patterns of action
• Formalizes and communicates them
• Designates and mans roles
  • Designated forum to meet and produce (e.g. mailing list)
  • centralized collection of knowledge (e.g. archive)
  • Handbook of procedures (e.g. FAQ)
  • Enforcer of violations (e.g. FAQman)

Coordination Instit. Facilitate COSI

• Proposition #4:
• Collective open innovation is more efficient when coordinated through institutions.

Framework

• Non-rival good
• Perception of fairness

=> Sustainability

• Based on generalized exchange
• Coordinated through institutions

=> Efficiency

Conclusions

• COSI is important in and beyond software
• Framework accounts for individual motivation and collective coordination
• Fits more cases than competing explanations
• Offers testable propositions
• Questions remain:
  • How applicable off-line? Why some projects are more successful?

ADD

• Sport enthusiasts - off-line COSI
• Wireless networks
• On-line forums
• "COSI: The Emergence of a New Economic Actor"
• Is there an important distinction between open and free software?