Using System Dynamics to Understand Disruption: A ...cfp.mit.edu/events/10Apr/Chintan Vaishnav-CFP...
Transcript of Using System Dynamics to Understand Disruption: A ...cfp.mit.edu/events/10Apr/Chintan Vaishnav-CFP...
1
Using System Dynamics to Understand Disruption: A General Model for Technology and Industry Disruption
Chintan Vaishnav MIT
© 2010 Chintan Vaishnav, MIT
Steps in the System Dynamics Modeling Process
Problem Articulation What is the problem? Key Variables Time Horizon
Formulating Dynamic Hypotheses Formulation of a Simulation Model
Causal Loops, Stocks and Flows Estimation
Testing and Validation Consistency with the purpose and boundaries Comparison with Expected Behavior Robustness of the Model Sensitivities and other tests
Policy Evaluations
2
© 2010 Chintan Vaishnav, MIT 3
Problem/Research Question
How can we improve decision making amidst technology and industry disruption?
© 2010 Chintan Vaishnav, MIT 4
Model Formulation, Testing, and Validation
© 2010 Chintan Vaishnav, MIT 5
Model: Philosophy and Principles
Regulatory
Policy
Dynamics
Technology
Dynamics
Corporate Strategy
Dynamics
Industry
Structure
Dynamics Customer
Preference
Dynamics
Telecommunications System
Technology
Dynamics
Regulatory
Policy
Dynamics
Corporate
Strategy
Dynamics
Industry
Structure
Dynamics
Customer
Preference
Dynamics
Subsystems Subsystem-level behavior
Analysis
Emergent (system-level) behavior
Synthesis
Innovation and competition emerge from the interaction of five gears…
Ref: Gear Model, Charles Fine
Model formulations rest upon: 1. Theories of adoption, tech strategy, and innovation 2. Unstructured interviews with stakeholders
© 2010 Chintan Vaishnav, MIT 6
Example: Model based on Theory
Firm Price Primary Performance (Quality)
Ancillary Performance (Innovation)
Incumbent (e.g. AT&T)
High High Low
Entrant (e.g. Skype) Low Low High
Christensen’s Conditions for Disruptive Technology (Christensen 1997)
Customer
Preference
Dynamics
Assumptions: • 2 Firms – Incumbent, Entrant • Each firm represents a typical firm in their industry
© 2010 Chintan Vaishnav, MIT 7
Example: Model based on Unstructured Interviews
Corporate Strategy
Dynamics
Customer
Preference
Dynamics
Assumption: • Features of each service are separable into (identifiable as) primary performance (quality) and ancillary performance (innovation)
© 2010 Chintan Vaishnav, MIT 8
Example: Model based on Unstructured Interviews (Contd.) (making Firm’s Strategy endogenous)
Corporate Strategy
Dynamics
Customer
Preference
Dynamics
Assumption: Both firms endowed with equal total attention (resources)
Price
Product Attractiveness
Adopters +
+ -
Firm's Market Share +
Maximum Ancillary
Performance
Time to Acquire Ancillary
Performance
Resources to Ancillary
Performance
+ - Ancillary
Performance
Primary Performance
-
+
Resources to Primary
Performance -
+
+ +
Time to Acquire Primary
Performance Maximum Primary
Performance + -
R1 Network Effect
innovation
quality
B4 Effect of Competitive
Pressure on Innovation
R5 Effect of Competitive Pressure on Quality
Effect of Firm's Expected Market Share on Resources
to Ancillary Performance
“The only strategy was that of a monopolist. Incumbent A did not care what other features you want!” Director, CTO Organization, Incumbent A
-
© 2010 Chintan Vaishnav, MIT 9
Example: Model based on Unstructured Interviews (Contd.) (making Firm’s Strategy endogenous)
Corporate Strategy
Dynamics
Customer
Preference
Dynamics
“Incumbent cares about ancillary performance only with: the entry of the non-traditional competitor, and the growth of its market share.” Director, CTO Organization, Incumbent A
Assumption: Both firms endowed with equal total attention (resources)
Price
Product Attractiveness
Adopters +
+ -
Firm's Market Share +
Maximum Ancillary
Performance
Time to Acquire Ancillary
Performance
Resources to Ancillary
Performance
+ - Ancillary
Performance
Primary Performance
-
+
Resources to Primary
Performance -
+
+ +
Time to Acquire Primary
Performance Maximum Primary
Performance + -
R1 Network Effect
innovation
quality
B4 Effect of Competitive
Pressure on Innovation
R5 Effect of Competitive Pressure on Quality
Effect of Firm's Expected Market Share on Resources
to Ancillary Performance
Competitor's Market Share
+ Effect of Competitor's
Market Share on Resources to Ancillary Performance
-
© 2010 Chintan Vaishnav, MIT 10
Example: Model based on Unstructured Interviews (Contd.)
“First [when the entrant enters] the question is whether this is a price game or a performance game. Then, you realize that the future is ancillary.” Chief Strategist and Architect, Incumbent B
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Competitor’s Expected Market Share
Firm’s Attention to Innovation (Ancillary Performance)
As competitor gains market share
more risk aversion
Effect of Competitor’s Market Share on Resources to Innovation
© 2010 Chintan Vaishnav, MIT 11
Model Testing and Analysis
Model Verification, Validation
1. Calibration with real-world trends 2. Expert Opinion: Regulators (FCC, MIAC-Japan), Industry Architects
and Strategy Experts (Motorola, BT, Nokia, Cisco, Comcast, Verizon), Academics (Primary Sources)
Testing 1. Sensitivity of each exogenous parameter (including those that were
made endogenous later) 2. Analysis of a unit model to understand structural forces and incentives 3. Analysis of calibrated model to understand timing and magnitude of
the forces 4. Industry Structure Scenarios
Integrated Incumbent Remains Dominant Niche Entrant, modular in technology and industry structure, displaces the
Incumbent Erstwhile Entrant (new Incumbent in the modular structure) remains
dominant with a new modular entrant present
© 2010 Chintan Vaishnav, MIT 12
Model Validation
100
75
50
25
0
1935 1943 1951 1959 1967 1975 1983 1991 1999 2007 Time (Year)
% Households
PSTN Penetration : Model
PSTN Penetration : Data Source: FCC NECA & USAC Data (2008)
Through calibration with data, agreement with shared mental models of stakeholders, and expert opinion
PSTN Calibration
© 2010 Chintan Vaishnav, MIT 13
Results
© 2010 Chintan Vaishnav, MIT
Base Case Behavior Passive Base Case - Incumbent does not respond to threat Active Base Case - Incumbent responds to threat
Adopters 200 M
150 M
100 M
50 M
0 3 3 3
3
3 3 3 3 3 3
1
1 1
1
1 1 1 1 1 1
0 24 48 72 96 120 144 168 192 216 240 Time (Month)
Adopters[Incumbent] : Passive Base Case 1 1 1 1 1 Adopters[Entrants] : Passive Base Case 3 3 3 3 3
4 4 4
4
4 4
4 4 4
2
2 2 2
2 2
2 2 2 2
Adopters[Incumbent] : Active Base Case 2 2 2 2 2 Adopters[Entrants] : Active Base Case 4 4 4 4 4 4
© 2010 Chintan Vaishnav, MIT 15
Market Uncertainty: Network Effect Phase Plot
0
10
20
30
40
50
60
70
80
90
100
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.4 2.8
Strength of Network Effect Sensitivity of Attractiveness to Compatibility (εn)
Max. Market Share
(Entrant)
No Network Effect (Services Interoperate)
Network Effect Helps Entrants Network Effect Helps Incumbents
Winner Takes All (WTA) outcome less probable Winner Takes All (WTA) outcome more probable
© 2010 Chintan Vaishnav, MIT 16
Market Uncertainty: Switching Cost Phase Plot
0
2
4
6
8
10
12
14
4% 3% 2% 1% Same -1% -2% -3% -4%
% of Incumbent's Consumers Considering Switching Relative to % of Entrant's Consumers
Time of Disruption (Years)
Time at which Entrant’s Market Share Crosses Incumbent’s Market Share
γ1 < γ2 γ1 > γ2
© 2010 Chintan Vaishnav, MIT 17
Market Uncertainty: Consumer Choice Phase Plot
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3
Sensitivity of Attractiveness of Attribute under Consideration
Max. Market Share
(Entrant)
Sensitivity of Attractiveness of Any Other Attribute
Consumers Value the Attribute as much as
Other Attributes
Consumers Value the Attribute Twice
as much as the Other Attributes Consumers Value the
Attribute Thrice As much as
the Other Attributes
Consumers Do Not Value the Attribute
εn
εi
εq
εp
© 2010 Chintan Vaishnav, MIT 18
Limits to Technology and Industry Disruption
Technology Disruption
No Technology Disruption
Industry Disruption
No Industry Disruption
• Weak Network Effect • Consumer highly price sensitive and willing to risk adopting innovative service with low quality and compatibility
Quadrant Not Studied • General double helix dynamics without technology disruption
• Strong Network Effect • Consumer value quality and compatibility over innovation and low price
• Incumbents can affect switching behavior heavily • Incumbents innovate while maintaining quality • Entrants struggle to offer quality due to lack of functional control or market power
© 2010 Chintan Vaishnav, MIT 19
THANK YOU