The model of products competition in mobile devices market Yana - - PowerPoint PPT Presentation

Yana Salikhova, PhD

• Assoc. Prof. of Marketing Department in St. Petersburg State University of

Economics Victoria Grigoreva, PhD

• Assoc. Prof. of Management Department in St. Petersburg National

Research University Higher School of Economics Ivan Svetunkov, PhD

• Assoc. Prof. of Centre for Marketing Analytics and Forecasting in Lancaster

University, UK

The model of products competition in mobile devices market

Work is performed with financial support

• f the Russian Fund for Fundamental Researches,

grant № 16-02-00172 «The development of multi-level competition theory, its methods and techniques».

Outline

Relevance Theoretical background Research method Findings and discussion

3

The relevance of the topic

• The complexity of market structure

platform ecosystem developers

mobile device producers intermediary and sellers

Customer

The relevance of the topic

Traits of US mobile device market

(Kenney, Pon 2011)

The role of the network carriers Carriers not only operate the network, but also sell handsets and provide content carriers are making decisions about how data is consumed on their networks and what services can be used

The relevance of the topic

• The anti-trust lawsuits

bundling its browser to its operating system (Winkler, 2014)

Microsoft

Google service as main on the device ((Edelman, 2015).

Google

investigating Google’s tactics in mobile (Kendall and Barr, 2015)

Google

Theoretical background

Structural

• examines whether it is the

concentration in the market that causes the firms to mutually collude to enhance market power, or there are some other factors responsible for it (Rezitis, 2010) Non-structu ral

• observes that when the size of the

firm increases, its share in market also increases and provides an

• pportunity for that firm to earn

higher profits (Claessens and Laeven, 2004).

Research method

Sample selection

• Top 6 US

tablet and smartphone vendors Data collection

• Euromonitor

Internationa l database,

• IDC

Data analysis

• segment

approach

• Model of

shares dynamics

Segment approach to market concentration

• heterogeneity of commodity markets , a set of

separate parts (segments) reflecting the demand characteristics of different types of consumers;

• product differentiation and methods of its

marketing;

• the intensity of competition on one segment

may differ from the intensity of competition in another segment and the entire market

Evaluation of market structure

• the degree of market concentration (CR, HHI,

Holl-Taydman, an index of entropy and Jeanie's coefficient and etc.)

• the degree of market share distribution by

segment - Ratio of shares distribution

Market concentration

US tablet and smartphone vendors, Market share, First Quarter 2017 Brand Tablet Smartphone Market share Apple 24,7 43,6 24,1 Samsung 15,1 28,5 17,5 Motorola (Lenovo) 7 5 4,1 Huawei 6 0,4 HTC 3,2 1,6 LG 9,6 9,4 CR4=55,1 K=0,09

Sources: comScore Mobilens, IDC Quarterly Personal Computing Device Tracker, 2017, Euromonitor International database, 2017

Market types in the concentration degree

Market share distribution by segment Market concentration Low concentrated Moderately concentrated Highly concentrated Uniform distribution Oligopoly Inhomogeneous

• ligopoly

Segment monopoly Nonuniform distribution Oligopoly with with features

• f segment

monopoly Inhomogeneous

• ligopoly with

features of segment monopoly Monopoly with segment competition

Research development

Modelling dynamics of ecosystems

• f products in mobile devices

market

The Idea

• 1. Model the dynamics of shares of companies in the

market of mobile devices;

• 2. Take possible connections of different types of

products (e.g. iPhone and iPad);

• 3. Get estimates of overall shares of companies in the

mobile devices market;

• 4. Forecast the shares dynamics for several years

ahead.

What to do

• 1. Calculate shares of sales of each product type in each

category.

• 2. Transform the data by dividing by number of categories, so all
• f that adds up to one.
• 3. Form vectors of shares in mobile devices market.
• 4. Choose the model.
• 5. Model estimation and validation.
• 6. Error measure.
• 7. Analyse the model.
• 8. Produce forecasts.

1. Calculate shares of sales of each product type in each category. let’s assume that we have the following shares in 2016:

Apple Samsung Lenovo LG Overall Smartphones 42% 23% 15% 20% 100% Tablets 37% 30% 27% 6% 100% Overall 79% 53% 42% 26% 200%

• 2. Transform the data by dividing by number of categories, so all of that

adds up to one.

Apple Samsung Lenovo LG Overall Smartphones 21% 12% 8% 10% 50% Tablets 19% 15% 14% 3% 50% Overall 40% 27% 22% 13% 100%

• 3. Form vectors of shares in mobile devices market.

Name of product Shares, 2016 Smartphones, Apple 21% Smartphones, Samsung 12% Smartphones, Lenovo 8% Smartphones, LG 10% Tablets, Apple 19% Tablets, Samsung 15% Tablets, Lenovo 14% Tablets, LG 3%

These values will be encoded as Yt, where t is the time index.

• 4. Choose the model.

where A1 is the square matrix of parameters, which has size of n×n (in our examplen=8) and Et is the vector of errors of the model, assumed to be distributed normally. This way matrix A1 will contain parameters that characterise the relation between different shares today and the same shares last year. For example, if there is a connection between “Smartphones, Apple” in 2016 and “Tablets, Apple” in 2015, this will be captured by the corresponding parameter inA1.

5.Each of the elements of Yt should be positive; We propose a simple modification that addresses this issue: where A∘B is the Hadamar product (element wise multiplication of vectors) and Et is assumed to have multivariate log-normal distribution. In this model A1Yt−1 is a vector, each element of which is multiplied by each element of the vector Et. Now Yt is always positive.

• 6. The sum of elements of Yt should be equal to one.

For example, the previous model can be modified the following way: where 1n'is the vector of ones, which means that in the denominator we sum up all the values produced by the model. The number of these past shares is denoted as p: An alternative approach of reconciliation is to use modified VAR model, where

• nly firs tn−1shares are taken into account and the last is formed as “one

minus sum of firs tn−1 shares”.

Last steps

• 7. Model estimation and validation.
• 8. Analyse the model.
• 9. Produce forecasts.

Thank you for the attention!

Yana Salikhova, PhD

• Assoc. Prof. of Marketing Department in St. Petersburg State University of

Economics Victoria Grigoreva, PhD

• Assoc. Prof. of Management Department in St. Petersburg National

Research University Higher School of Economics Ivan Svetunkov, PhD

• Assoc. Prof. of Centre for Marketing Analytics and Forecasting in Lancaster

University, UK

The model of products competition in mobile devices market