Home
Work
in progress
News
Workshops
References (project)
Current References (general)
Project
Work
in Progress You will find
here the working papers prepared for a special issue of "La
Revue d'Economie Politique" sous le titre "Marchés
en ligne et communautés d'agents" (Virtual
communities and online markets). You can access
and download the papers (french version) and browse the abstracts
(english version). "Biens
informationnels et communautés médiatées" : Michel Gensollen
(ENST) Information
goods and online communities Communautés
d'expérience Demand Self-Organization : Learning through Infomediation "Communautés
d'expérience et concurrence entre sites de biens culturels"
: Marc Bourreau et Michel Gensollen (ENST) Competition between Online Suppliers of Cultural
Goods "Transition
vers un marché électronique et communautés
médiatées" : Eric Darmon et Dominique Torre
(LATAPSES) From Search to Electronic markets: The Role
of Virtual Communities "Communautés
de consommateurs et marché infomédié"
: Alain Raybaut et Stéphane Ngo-Maï (LATAPSES) Communities of consumers and "infomediated"
market Communautés
épistémiques d'innovation "Communautés
épistémiques: réseaux cognitifs et interdépendance entre
les partenaires" : Bernard Conein (EHESS) Epistemic
communities and cognitive networks: cooperation and distributed
cognition "La
régulation de la connaissance : arbitrage sur la taille et
gestion aux frontières dans la communauté Debian"
: Nicolas Auray (EHESS) The regulation
of knowledge: size and scope control for the Debian community "Existence,
stabilité et efficacité des communautés épistémiques"
: Nicolas Curien, Gilbert Laffond, Jean Lainé, François Moreau (CNAM)
Existence,
stability and efficiency of epistemic commmunities Communautés
épistémiques de production "Relations
inter-entreprises et communautés médiatées: une
analyse préliminaire" : Richard Arena (LATAPSES) Inter-Firm
Relations and Virtual Communities: A Preliminary Analysis Knowledge and
social interactions in on-line communities
Digitalization and virtualization of information goods gradually
moves the value of these goods from their physical layer towards
the social process which develop and prepare a demand that suits
them best. In the other hand, the ICT (Information and Communication
Technologies) are currently challenging the top-down information
model, from producers to final consumers, through the mass-media
system and the retail network. On Internet, new structures of
interaction between consumers (virtual communities) are already
shaping effective ways to initiate and develop the demand for
information goods. This paper presents three types of virtual
communities which jointly play a significant role to develop a
bottom-up model : (a) files sharing communities (like Napster),
that challenge the top-down editorial model; (b) experience sharing
communities (like Amazon.com), that help consumers form an ex
ante estimation of experience goods characteristics ; theses communities
trigger the value migration from information to meta-information;
(c) knowledge sharing communities (like free software user list),
that help consumers understand and adapt complex goods to their
needs ; these epistemic communities could lead information goods
designers to better understand the consumers' usage constraints.
These virtual communities are original interaction structures:
they are characterized by very weak ties or even no interpersonal
relations; they center around an information corpus, built and
used by anybody for his or her own sake. The survival of these
communities does not depend mainly on the altruism of the participants
but on the structure of the corpus (the file sharing utility,
the reviews, the forum or the user list, etc.). As a regulation
model for final markets, virtual communities could lead to a wider
variety and a greater adaptability of end-products.
In this paper we analyse a process of demand learning
based on "infomediation", i.e. relying upon information
exchanges across consumers within Web communities. The characteristics
of the supplied good are subject to horizontal differentiation and
change over time, so that the buyers' information on available varieties
is imperfect at each period. We look for conditions under which infomediation
proves able to decentralize efficiently demand segmentation, i.e.
leads to welfare maximizing. We compare the output of self-organized
infomediation to the ideal benchmark of segmentation under perfect
information and we discuss the influence of various parameters, such
as : the number of infomediation communities ; the respective weights
of advices retrieved on the Web and of external factors in purchase
decisions ; the degree of informational externalities across communities.
We study competition between two online suppliers
of cultural goods (like books, CDs, or DVDs), when the suppliers
can introduce software tools which allow consumers to search for
and evaluate products. Using a sequential entry game, we show that
one supplier offers only "star" goods (i.e., the goods
whose quality is known by all consumers), whereas its competitor
provides only "non star" goods (i.e., the goods whose
quality is uncertain). We study the incentives of the suppliers
to introduce three types of tools - samples, search tools, and a
customer review system - and their effect on the market equilibrium.
Keywords: Cultural goods; Internet; E-commerce; Search.
In this paper, we study the impact of virtual
communities together with the transition from traditional (search)
to electronic markets. For that sake, we analyse the dynamics
of an economy composed of buyer-seller agents trading on these
two markets. On the search market, the uncertainty deals with
the occurrence of an exchange; on the electronic market, consumers
can imperfectly appreciate the qualities of the goods they face
to. We first describe the properties of this economy, as agents
are not able to learn on the electronic market; we point out
the crucial role of the stickiness of the search market and
of the exogenous expertise of agents. We then introduce virtual
communities. We show that these communities are able to promote
the use of the electronic channel. Their welfare impact may
be found globally positive although some categories of agents
may be worst off. We turn at last to the organisational aspects
of such communities. As predicted by intuition, the size of
the community has a global positive impact. Besides, according
to their composition (both the heterogeneity's degree of the
individual expertise and of individual tastes), some types of
communities may be found superior in a changing context.
e-commerce - search - learning - virtual communities - coordination
- Internet economics
This paper examines a market where a
final good producer, some communities of consumers and an infomediary
interact. We distinguish, on empirical basis, two types of distributions
for the size of communities, a log-normal law and a power law,
and are interested in the viability conditions of the market.
The emergence of either distribution is related to the weight
of collective beliefs of the consumers. The infomediary collects
information upon communities and sells it to the producer. His
control variable bears on the number of hyperlinks between communities
(connectivity). We compare the viability conditions and show the
existence of critical ranges for the equilibrium connectivity
associated with optimal behaviors in the two distributions settings.
The goal of this article is to define the mode of
cooperation specific to epistemic communities and to agents involved
in the logic of innovation based upon learning and exploration.
The notions of epistemic community and cognitive network are used
to characterize a species of collective action and mode of cooperation
involved in use of the internet. However, I argue that the terms
'community' and 'cognitive network' should be restricted in their
use. Close examination needs to precede labels. I also argue for
a full recognition of the cognitive properties involved in knowledge
exchange. Some modes of cooperation are efficient means for knowledge
growth. Internet based technology and Open Source architectures
provide a powerfull scaffolding for cooperation because they are
both social tools and knowledge facilitators.
The epistemic communities are confronted with a recurrent
problem : they must regulate their scope, so as to arbitrate between
the club effect and the congestion effect. This compromise integrate
actually a complex plurality of regulation choices. What should
be the scale of the openness for the new entrants ? How to proportionnate,
for each developper, the time allowed to select the applicants and
the time allowed to send an advice to the community ? How shall
the member of the community share out his effort between assistance
for newbies and cooperation amongst experts ? The hypothesis of
the paper is that the caracteristics of the regulation depends on
the cognitive content of the flow of ideas. It is with a sociological
analysis of a complex community that the article tries to answer
to those questions.
The
model presented in this paper separates two stages of labour within
an epistemic group, such as an open-source community : a first stage
is devoted to project design and it generates increasing returns
to scale with respect to the community size, due to a club externality
specific to creative activities ; then, a second stage is devoted
to implementing the epistemic system and it exhibits decreasing
returns to scale, incurred when integrating individual inputs into
a consistent collective output. We show that, depending on exogeneous
factors (club externality, integration level of output, project
resistance to implementing, users' preferences between system capacity
or interoperability), an efficient community may be either degenerated
and restricted to a single agent, or it may extend to the whole
background population, or its size may be bounded and greater than
unity. Now, the optimum cannot be easily decentralized through individual
behaviour, because of a stability/efficiency dilemma which can better
be solved through an authority controlling entries in the community,
rather than by adopting an allocation rule of common gains, a result
consistent with empirical evidence from the open source world.
The purpose of the contribution is the investigation
of the analytical relations which could allow the extension
of the scope of the concept of virtual community from consumers/internet
-users to firms/internet users. The temptative answer to this
preoccupation is given thanks to three steps. The first part
of the paper is supported by the existing literature related
to the emergence and working processes of virtual communities
formed by individual agents within a knowledge-based economy.
This literature permits to point out the basic logical criteria
by which communities can be defined and characterised in real
economies. The second part of the paper utilises the literature
focused on inter-firm relations to try to find in it tools which
could permit to assimilate firm communities to a specific form
of these relations. The last part of the paper checks if it
is possible to extend the basic logical criteria pointed out
in the first part to the case of firm virtual communities. The
main result of the paper is that such communities might exist
and that they correspond to some precise forms of inter-firm
cooperation.
How could we understand community usages of internauts
? In order to respond to this question we go back to two literature
areas : Marketing and Knowledge Management, as they have started to
analyse the cognitive impact of communities since the eighties (first
part). Taking back some results of these works, we offer an analytical
model of the on-line behaviour in communities (second part). Our factual
study, based on a group of leader users, leads us to identify several
results (third part). These ones deal with the structural impacts of
technical and social networks ; the hybrid aspect of on-line community
practices ; the learning dynamics linked with electronic communities
; the identification of the key role plaid by the "frontier-members"
; and, ultimately, the properties of two "ideal-types" of
on-line communities.