Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Social Network Effects Bertil Hatt EconomiX, France Telecom R&D
S´eminaire Draft – Nanterre October 10, 2006
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Introduction Increasing return without lock-in Some properties and our intuition Findings in quantitative sociology Local preferences Structural concerns Layers networks Consequence for a symmetric service Coordination favors a group-adoption Consequence for an asymmetric service Publication monopoly Publication competition Conclusion and discussion Results and developments
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Introduction Increasing return without lock-in Some properties and our intuition Findings in quantitative sociology Local preferences Structural concerns Layers networks Consequence for a symmetric service Coordination favors a group-adoption Consequence for an asymmetric service Publication monopoly Publication competition Conclusion and discussion Results and developments
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Introduction Increasing return without lock-in Some properties and our intuition Findings in quantitative sociology Local preferences Structural concerns Layers networks Consequence for a symmetric service Coordination favors a group-adoption Consequence for an asymmetric service Publication monopoly Publication competition Conclusion and discussion Results and developments
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Introduction Increasing return without lock-in Some properties and our intuition Findings in quantitative sociology Local preferences Structural concerns Layers networks Consequence for a symmetric service Coordination favors a group-adoption Consequence for an asymmetric service Publication monopoly Publication competition Conclusion and discussion Results and developments
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Introduction Increasing return without lock-in Some properties and our intuition Findings in quantitative sociology Local preferences Structural concerns Layers networks Consequence for a symmetric service Coordination favors a group-adoption Consequence for an asymmetric service Publication monopoly Publication competition Conclusion and discussion Results and developments
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Increasing return doesn’t always lead to lock-in
I
Increasing returns lead to lock-in (Arthur) I I
I
Bounded preference Successive entry
MacOS, Azerty, languages, etc. Our approach: relations matter I
Local, unevent, structured
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Increasing return doesn’t always lead to lock-in
I
Increasing returns lead to lock-in (Arthur) I I
I
Bounded preference Successive entry
MacOS, Azerty, languages, etc. Our approach: relations matter I
Local, unevent, structured
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Increasing return doesn’t always lead to lock-in
I
Increasing returns lead to lock-in (Arthur) I I
I
Bounded preference Successive entry
MacOS, Azerty, languages, etc. Our approach: relations matter I
Local, unevent, structured
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Some properties and our intuition
I
Consider relatives
I
Local stability
I
Some influential users
I
More efficient & universal
I
Social structure
I
Rapid diffusion
I
Foci: family, work, etc.
I
Lingered by connotations?
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Some properties and our intuition
I
Consider relatives
I
Local stability
I
Some influential users
I
More efficient & universal
I
Social structure
I
Rapid diffusion
I
Foci: family, work, etc.
I
Lingered by connotations?
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Some properties and our intuition
I
Consider relatives
I
Local stability
I
Some influential users
I
More efficient & universal
I
Social structure
I
Rapid diffusion
I
Foci: family, work, etc.
I
Lingered by connotations?
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Outline Cases of increasing return without lock-in Some properties and our intuition
Some properties and our intuition
I
Consider relatives
I
Local stability
I
Some influential users
I
More efficient & universal
I
Social structure
I
Rapid diffusion
I
Foci: family, work, etc.
I
Lingered by connotations?
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Social network properties
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Homophile and clustered (Rogers & Bhowmik)
Homophily Preference for similar people Heterophily is limited: I Similarities help understanding, but I Difference is interesting Endophily Preference for one’s own group Clustering Significantly more ties within groups
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Homophile and clustered (Rogers & Bhowmik)
Homophily Preference for similar people Heterophily is limited: I Similarities help understanding, but I Difference is interesting Endophily Preference for one’s own group Clustering Significantly more ties within groups
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Homophile and clustered (Rogers & Bhowmik)
Homophily Preference for similar people Heterophily is limited: I Similarities help understanding, but I Difference is interesting Endophily Preference for one’s own group Clustering Significantly more ties within groups
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Social network structure is efficient (Barab´asi, Newman)
Intermediaries belong to several groups Hubs (high-degree) are better connected Small-world network: most “paths” are shorts Power-law degree distribution: numerous hubs Scale-free network: clusters exists are any size Degree-assortivity: hubs know more hubs
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Social network structure is efficient (Barab´asi, Newman)
Intermediaries belong to several groups Hubs (high-degree) are better connected Small-world network: most “paths” are shorts Power-law degree distribution: numerous hubs Scale-free network: clusters exists are any size Degree-assortivity: hubs know more hubs
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Social network structure is efficient (Barab´asi, Newman)
Intermediaries belong to several groups Hubs (high-degree) are better connected Small-world network: most “paths” are shorts Power-law degree distribution: numerous hubs Scale-free network: clusters exists are any size Degree-assortivity: hubs know more hubs
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Incompatible foci draw different networks (Feld)
I
Foci can be incompatible
I
Technology allows coexistence
I
Institutional structures needed
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Incompatible foci draw different networks (Feld)
I
Foci can be incompatible
I
Technology allows coexistence
I
Institutional structures needed
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Local preferences Structural concerns Layers networks
Incompatible foci draw different networks (Feld)
I
Foci can be incompatible
I
Technology allows coexistence
I
Institutional structures needed
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
Consequences for communication services
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
Coordination favors a group-adoption
I
The cost-quality ratio leads to a contact threshold
I
Groups large enough should subscribe
I
Explicit belonging helps coordination
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
Coordination favors a group-adoption
I
The cost-quality ratio leads to a contact threshold
I
Groups large enough should subscribe
I
Explicit belonging helps coordination
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
Coordination favors a group-adoption
I
The cost-quality ratio leads to a contact threshold
I
Groups large enough should subscribe
I
Explicit belonging helps coordination
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
A meta-network approach
I
Group- (rather than individual-) adoption
I
Coherent by focus
I
Structure yet to be studied
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
A meta-network approach
I
Group- (rather than individual-) adoption
I
Coherent by focus
I
Structure yet to be studied
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
A meta-network approach
I
Group- (rather than individual-) adoption
I
Coherent by focus
I
Structure yet to be studied
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
A meta-network approach
I
Group- (rather than individual-) adoption
I
Coherent by focus
I
Structure yet to be studied
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Groups simplify coordinated adoption Group adoption pattern
A meta-network approach
I
Group- (rather than individual-) adoption
I
Coherent by focus
I
Structure yet to be studied
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Publication monopoly Publication competition
Skewed demand leads to uncommon results
With exogenous quality, price: I Is proportional to cost I Decrease with skewness Endogenous quality increases with I Network skewness (usually) I Easier innovation I Cost of service
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Publication monopoly Publication competition
Skewed demand leads to uncommon results
With exogenous quality, price: I Is proportional to cost I Decrease with skewness Endogenous quality increases with I Network skewness (usually) I Easier innovation I Cost of service
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Publication monopoly Publication competition
Competition between two publication services
I
Discriminating mixed equilibrium
I
Incentive for lower price
I
Innovation as a lock-in
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Publication monopoly Publication competition
Competition between two publication services
I
Discriminating mixed equilibrium
I
Incentive for lower price
I
Innovation as a lock-in
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Publication monopoly Publication competition
Competition between two publication services
I
Discriminating mixed equilibrium
I
Incentive for lower price
I
Innovation as a lock-in
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Results Developments Discussion
Conclusion
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Results Developments Discussion
A social angle change network effects
I
Local stability
I
Still path-dependent
I
Frontiers are costly but rational
I
More specific services
I
Faster replacement
I
Competition shift to quality
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Results Developments Discussion
A social angle change network effects
I
Local stability
I
Still path-dependent
I
Frontiers are costly but rational
I
More specific services
I
Faster replacement
I
Competition shift to quality
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Results Developments Discussion
A social angle change network effects
I
Local stability
I
Still path-dependent
I
Frontiers are costly but rational
I
More specific services
I
Faster replacement
I
Competition shift to quality
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Results Developments Discussion
Developments from this paper Available work in progress: I Power-law foci combine into a log-normal mode More research to be done: I Endogenous structure I I
I
More precise subscription model I I
I
Social ties with schedule constraint Knowledge of publication service Multihoming at a cost Situated rationality model
Groups & foci I I
Adapt over-lapping algorithm Link group-structures to foci
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Results Developments Discussion
Developments from this paper Available work in progress: I Power-law foci combine into a log-normal mode More research to be done: I Endogenous structure I I
I
More precise subscription model I I
I
Social ties with schedule constraint Knowledge of publication service Multihoming at a cost Situated rationality model
Groups & foci I I
Adapt over-lapping algorithm Link group-structures to foci
Introduction Findings in sociology Symmetric service Asymmetric service Conclusion and discussion
Results Developments Discussion
Discussion & questions Presentation outline Introduction Increasing return without lock-in Some properties and our intuition Findings in quantitative sociology Local preferences Structural concerns Layers networks Consequence for a symmetric service Coordination favors a group-adoption Consequence for an asymmetric service Publication monopoly Publication competition Conclusion and discussion Results and developments
Paper outline 1. Introduction 2. Basic properties of complex, scale-free and social networks 3. Adoption can be sigmoidal without local preference 4. General model of our research 5. Clustering and niche stability 6. Homophile ties favor specific services 7. Asymmetric service and assortivity 8. Focus-organized world 9. Cross-properties strengthen these effects 10. A dynamic approach would understand longer-term stability
Criticism and changes Section 7 is completely re-done: I
ζ(·) simulated values were false;
I
still problems with dd aλ : does the innovation difficulty increase quality? p
∗
I I
a∗ = γ−λ+1 φ(γ) with φ(·) ' 1: sign inclear; need a relevant innovation difficulty λ.
Coherence of the paper: I
No connection between symmetric & asymmetric case
I
Difficult to grasp the overall goal
Possible restructuring Split and detail the two cases: I
Two distinct papers on symmetric/asymmetric case
I
Include more discussion
I
Consider multi-homing with a cost
Dynamic integration of the two cases as a last paper Symmetric Asymmetric Combination Single service Multihoming Symmetric Asymmetric Combination Single service Multihoming I
Possible restructuring Split and detail the two cases: I
Two distinct papers on symmetric/asymmetric case
I
Include more discussion
I
Consider multi-homing with a cost
Dynamic integration of the two cases as a last paper Symmetric Asymmetric Combination Single service Multihoming Symmetric Asymmetric Combination Single service Multihoming I
