Hysteresis in consumer Markets with focus on the Mobile communications Market

1
Hysteresis in consumer Markets with focus on
the Mobile communications Market

• Conor Twomey
• B.Sc. Financial Mathematics and Actuarial
  Sciences (Honours)

2
Overview

• Introduction
• Hysteresis
• Irish mobile phone market
• Motivation for this project
• Case Studies
• Simple switching in payment methods
• Switching between companies
• Effect of a special offer
• Conclusion
• Observations
• Questions/Discussion

3
Introduction

• Hysteresis in consumer Markets with focus on the
  Mobile communications Market

4
Introduction

• Hysteresis
• A system that exhibits rate-independent memory
• Hysteresis transducer
• Thresholds a and ß

5
Introduction

• Hysteresis
• Parallel connection of relay
• Preisach function µ(a,ß)
• Problem of Identification
• Preisach Model
• Applications

6
Introduction

• Irish mobile phone market
• 1986 Eircell begin operations
• 1995 Technology breakthrough
• - a growing necessity and demand
• 1997 - Introduction of Pay-as-you-go
• - Esat Digifone awarded license
• 2001 Meteor awarded license
• 2003 Introduction of full number portability
• 2005 Full Market Penetration
• Today Highly competitive market

7
Introduction

• Irish mobile phone market
• Market Statistics
• Call Minutes 1,449,462,000
• SMS Minutes 1,048,112,000

8
Introduction

• Motivation for this project
• Recent changes in marketing strategies
• Increase in people switching services
• Identification of thresholds to optimize profits
  and market share for companies

9
Case Studies

• Hysteresis in consumer Markets with focus on the
  Mobile communications Market

10
Case Study 1

• Simple Switching model in payment methods
• Introduction
• Theoretical Framework
• Fitting the data
• Summary

11
Case Study 1

• Introduction
• Customer choosing between two payment methods
• Bill pay (B)
• Pay-as-you-go (P)
• Non-monetary transaction cost
• P to B (ap,b)
• E.g. financial commitment to operator, effort in
  switching etc.
• B to P (ab,p)
• E.g. sacrificing handset subsidy, restricted
  services etc.

12
Case Study 1

• Theoretical Framework
• Switching
• ub(t) - up(t) gt ab,p will cause switch B to P
• up(t) - ub(t) gt ap,b will cause switch P to B
• Input
• x(t) ub(t) - up(t)
• z(0)z0
• Output
• z0 if -ab,p lt x(tk) lt ap,b for all tk
  0,t
• P if t in 12 months, t st x(t) ab,p
  and for all tk in t,t x(tk)ltab,p
• B if t in 0,t st x(t) ap,b and for
  all tk in t 12 months, t x(tk)lt -ap,b

13
Case Study 1

• Fitting the data
• Deciding on input
• Xi (t) Expected savings per month at time of
  switch
• Average monthly spending
• Distribution of observed thresholds
• Assumptions
• Independence
• ab,p ap,b gt 0

14
Case Study 1

• Fitting the data
• Geometric Interpretation

15
Case Study 1

• Fitting the data
• Preisach
• µ(A,B) P(A).P(B) Preisach
  function
• Using Mathematical software
• 0 and 1
• Market Data
• 0.75 are P and 0.25 are B
• 0.25

16
Case Study 1

• Summary
• Good first estimate
• Room for complexity
• More input variables
• More individual data
• Results weakened by assumptions

17
Case Study 2

• Switching between companies
• Introduction
• Theoretical Framework
• Geometric Interpretation
• Conclusion

18
Case Study 2

• Introduction
• Customer choosing between three companies
• Vodafone(V)
• 02 (O)
• Meteor(M)
• Transaction costs
• Cost for switch from company i to company j aij
  
• Utility associated
• Utility for company i Ui

19
Case Study 2

• Theoretical Framework
• Transaction cost matrix
• Input
• x(t) ui(t) - uj(t)
• y(t0)
• Output

20
Case Study 2

• Geometric Interpretation

21
Case Study 2

• Conclusion
• Data constraints
• Identification of input variables
• Accumulation of data
• Mathematical complications
• Complex calculations
• Two-dimensional input

22
Case Study 3

• Effect of a special offer
• Introduction
• Theoretical Framework
• Fitting the data
• Summary

23
Case Study 3

• Introduction
• Customer choosing between two types of company
• Traditional marketing strategy(VO)
• Predatory marketing strategy(M)
• Non-monetary transaction cost
• VO to M (avo,m)
• E.g. Moving to less well established company,
  learning new rates and rules that apply, losing
  network coverage etc.
• M to VO (am,vo)
• E.g. sacrificing special offer, calls to you cost
  more etc.

24
Case Study 3

• Theoretical Framework
• Switching
• um(t) - uvo(t) gt am,vo will cause switch M to VO
• uvo(t) - um(t) gt avo,m will cause switch VO to M
• Input
• x(t) um(t) - uvo(t)
• z(0)z0
• Output
• z0 if -am,vo lt x(tk) lt avo,m for all tk
  0,t
• VO if t in 0, t st x(t) am,vo and for
  all tk in t,t x(tk) lt am,vo
• M if t in 0,t st x(t) avo,m and for
  all tk in t, t
• x(tk) lt -avo,m

25
Case Study 3

• Fitting the data
• Deciding on input
• Xi (t) Expected savings per month at time of
  switch
• Average monthly spending
• Distribution of observed thresholds
• Assumptions
• Independence
• avo,m am,vo gt 0

26
Case Study 3

• Fitting the data
• Geometric Interpretation

27
Case Study 3

• Fitting the data
• Preisach
• µ(A,B) P(A).P(B) Preisach
  function
• Using Mathematical software
• 0 and 1
• Market Data
• 0.797 are VO and 0.203 are M
• 0.203

28
Case Study 3

• Summary
• Observed results not corresponding to market data
• Observed population
• Accuracy of data
• Results weakened by assumptions
• Room for complexity
• More input variables
• Consider timing
• More individual data
• Flow of customers

29
Conclusion

• Hysteresis in consumer Markets with focus on the
  Mobile communications Market

30
Conclusion

• Observations
• Acquiring Data
• Companies not willing to share
• Thresholds are theoretical if in both directions
• Subjective nature of more complex input
• Interpretation of input
• Rate of growth
• Somewhat comparable to a population model
• Effect of time and increase in demand of
  consumers

31
Conclusion

• Questions/Discussion
• 3 state relay in 3 company problem
• Tests on the validity of distributions
• Use predator prey type modelling in market share
  for effect of special offer
• Put a price a more accurate price on non-monetary
  transactions costs

32
Most common question

• What phone operator should I be on ?

33
Most common question

• METEOR!