Copyright is owned by the Author of the thesis.  Permission is given for 
a copy to be downloaded by an individual for the purpose of research and 
private study only.  The thesis may not be reproduced elsewhere without 
the permission of the Author. 
 



DISSERTATION  1 
 

 

Google Trends as a Complementary Tool for New Car Sales Forecasting: A Cross-

Country Comparison along the Customer Journey.  

A case for India and South Korea 

 

Master of Business Studies 

In  

Marketing 

 

at Massey University Albany, New Zealand 

Mina Sayedi (Student ID 15369140) 

Supervisor: Vishnu Menon 

December 2020 

 

 

 

 

 

 



 DISSERTATION  2 
 

 

 

Abstract 

This dissertation aims to evaluate factors responsible for different consumption patterns and 

customer’s preferences for goods and services in India and South Korea. It further analyzes 

several model frameworks that reflect the different variants of conceptual ideas of different 

authors. It gives special importance to customer-oriented study with the aim of improving 

customer-oriented analysis. It also discusses the assessment of economic conditions of India and 

South Korea as purchasing power of customers is directly proportional to the state of the 

economy. It further discusses challenges involved in pricing of new car models. The focus is 

given to quantitative research to evaluate the correlation between the customer preference for a 

particular model of car and data of new car sales. This purpose is achieved using several linear 

regression models and cross-sectional studies. It has been observed that seasonality has little 

impact on car prices in India as per the data retrieved from Google Trend Index but in South 

Korea seasonality has significant impact on car prices. This dissertation also discusses the 

significance of Google Trend Index in both the countries. It further provides that use of studies 

used in this dissertation can serve as a valuable tool for predicting the demand for various car 

models in both the countries which would not be accurately provided by use of traditional 

approaches.  

 

  



 DISSERTATION  3 
 

Table of Contents 

List of Abbreviations ............................................................................................................................... 5 

List of Figures ......................................................................................................................................... 6 

1.0 Introduction ....................................................................................................................................... 7 

1.1 Background of the study ................................................................................................................ 7 

1.2 Problem statement ........................................................................................................................ 8 

1.3 Research objective and research questions .................................................................................. 10 

1.4 Specific Research questions ......................................................................................................... 11 

1.5 Significance of the study .............................................................................................................. 12 

Chapter 2: Literature Review ................................................................................................................. 14 

2.0 Introduction ................................................................................................................................. 14 

2.1 Systematic Literature Review ....................................................................................................... 16 

2.1.1 Production of new Car models ............................................................................................ 17 

2.1.2 Supply chain of new model cars .......................................................................................... 19 

2.1.3 Pricing ................................................................................................................................. 21 

2.2 Buying Decision Models ............................................................................................................... 23 

2.2.1 Five-Stage Model of the Buying Process ............................................................................. 23 

2.3 Cultural Dimensions in India and South Korea .............................................................................. 26 

2.3.1 Cultural Difference and Consumption Behavior ............................................................... 26 

2.3.2 Hofstede’s Cultural Dimensions ......................................................................................... 28 

2.3.3 Cultural Differences in India and South Korea ................................................................. 33 

2.4 Google Trends Predictions and Forecasting .................................................................................. 35 

2.4.1 Predictions and Forecasting ................................................................................................ 35 

2.4.2 Google Trends data as a tool of now-casting and forecasting ............................................ 38 

2.4.3 Shifts in the Forecasting Context ........................................................................................ 38 

2.4.4 Judgment Forecasting as a Source of Inaccuracies ............................................................ 39 

2.5 State of the economy ................................................................................................................... 45 

2.5.1 Relationship between GDP and new car sales in India and South Korea ......................... 45 

2.6 Theoretical Framework ............................................................................................................... 48 

2.6.1 The customer Journey using the AIDA Model ................................................................... 48 

Chapter 3: Methodology ........................................................................................................................ 53 

3.0 Introduction ................................................................................................................................. 53 

3.1 Google Trends Tool ...................................................................................................................... 53 



 DISSERTATION  4 
 

3.2 Research Design........................................................................................................................... 56 

3.3 Data measurement ...................................................................................................................... 57 

3.4. The Reliability and Validity of Google Trends Data ...................................................................... 58 

3.5. Advantages and Disadvantages of Google Trends in Comparison to a Survey .............................. 63 

Advantages of Google Trends in Survey ..................................................................................... 63 

Demerits of Google Trends in Survey ......................................................................................... 64 

3.6 Regression Analysis in the Context of Internet Data ..................................................................... 68 

Chapter 4: Results and Presentation ..................................................................................................... 72 

4.1 Data Analysis and presentation: State of the Economy and Car prices across Countries............. 73 

4.2 Seasonality impacts on Car prices ................................................................................................ 77 

Chapter 5: Conclusion, Discussion, and Future Recommendations ......................................................... 79 

5.1 Overview ..................................................................................................................................... 79 

5.2 Key Findings and Data Analyses ................................................................................................... 80 

5.3 Discussion .................................................................................................................................... 81 

5.4 Significance of the study .............................................................................................................. 83 

5.5 Limitation and Future Research ................................................................................................... 84 

References............................................................................................................................................. 86 

 

 

  



 DISSERTATION  5 
 

List of Abbreviations 

NCS    New Car Sales 

SPSS    Statistical Package for Social Scientists 

UN    United Nations 

SIAM    Society of Indian Automobile Association 

IMF    International Monetary Fund 

 

  



 DISSERTATION  6 
 

List of Figures 

Figure 1: Supply chain in automobile industry………………………………………………………………….21 

Figure 2: Kotler’s Five Stage Model……………………………………………………………………………..24 

Figure 3: Hoftede’s Cultural Dimension: India and South Korea…………………………………………….29 

Figure 4: Reflection model for Customers’ Intention to purchase products…………………………………47 

Figure 5: India’s economy and car sales………………………………………………………………………..49 

Figure 6: Relationship between Indian economy and commercial car sales…………………………………50 

Figure 7: AIDA model……………………………………………………………………………………………52 

Figure 8: Customer Journey……………………………………………………………………………………...54 

Fig 9: Google Trends Search Request for the Term “Toyota” comparison…………………………………..60 

 

List of Tables 

Table 1: Google Trends advantages and disadvantages……………………………………………………..70  

Table 2: The practicality and the simplicity of Google Trends………………………………………………...71  

Table 4.1.1: Model Summary (All variables)…………………………………………………………………77 

Table 4.1.2. Regression Coefficients Results…………………………………………………….....................78 

Table: 4.1.3: The accuracy rate of selected car models………………………………………………………79 

Table 4.1.4: Regression analysis with the absence of time lag ……………………………………………….81 

Table 4.2.1:  Seasonality and car prices in India………………………………………………………………82 

Table 4.2.2:  Seasonality and car prices for the case of South Korea………………………………………..83 

 

 

  



 DISSERTATION  7 
 

1.0 Introduction 

This chapter provides a brief background of the study, problem statement, research objective, 

research questions, hypothesis, and definitions of relevant concepts in the thesis.   

1.1 Background of the study 

The automotive industry is one of the essential industries in India and South Korea since these 

two countries are emerging economies (Vosen & Schmidt, 2011). Forecasting of sales within the 

automotive industry is essential because new cars in the present generation are built to either 

deliver or forecast. However, forecast usually leads to the effect of bullwhip since there are 

demand uncertainty and inaccuracies in forecasting (Karlgren et al., 2012). Although some cars 

are built for delivery, the issue of accurate forecasting is important for managers and decision-

makers to efficiently and effectively allocate and plan their resources.  

Google Trend is a tool employed by various companies in both developed and developing 

countries to collect updated and large sums of data on how consumers respond to new car sales 

(Carley et al., 2013).  After realizing how important this development is, many automotive 

companies have paid great attention to Google Search and other social media platforms to 

develop ideas and implement positive and significant decisions (Karlgren et al., 2012; Liu, 

2012). Typically, consumers spend much of their time searching for new and potential vehicle 

models. Kandaswami and Tiwar (2014) conducted a study and discovered that the average time 

spent by people on Google to search for cars of their choices was 10 hours a day. In Asian 

countries, including China, India, South Korea, and Japan, the number approached 70% 

(Kandaswami & Tiwar, 2014). Currently, the registered search engines like Google.com have a 

more extensive search platform for decision information on new car sales. Past authors have 



 DISSERTATION  8 
 

conducted research on the forecasting power of Google Trends on car sales. However, their 

results differ from time to time (Barreira et al., 2013; Geva et al., 2017). Geva et al. (2017) 

discovered that the forecast models for best car models sales is a combination of Google Trends’ 

data, forum mention and forum sentiment. In their research, they found that Google Trends and 

forum sentiment have the same predictive power and volume. According to information provided 

by Trading Economics global macro models, the use of Google Trends as a tool for forecasting 

new car sales across India and South Korea, the customer journey of AIDA model is equally 

essential (Gensler et al., 2017). AIDA describes the four procedures that the customer considers 

during the process of buying. First A means attention. Attention is always attained through 

publishing activities and sales promotion, that may be done, for instance, in Google adverts. 

Letter I means interest that is portrayed in the customer’s search behaviours it may be registered 

through Google Trends (Vosen & Schmidt, 2011). D means desire. The desire evolves in 

people’s sentiment that may be registered through positive and negative Google expressions. The 

last letter ‘A’ means action, which is the customer’s real purchasing power.  

1.2 Problem statement 

The process of decision making in companies is affected by the efficiency of forecasting 

methods. This is because it reduces the dependency on probability and acts as a scientific way to 

cope with external events (Waheed et al., 2014). The manufacturers of different cars are forced 

to be prepared for future and plan for the increased demand for new car models, especially 

within countries with a high developing rate. The demand efficiency across countries leads to 

several problems because approaches are needed to handle a wide range of data that affects the 

functioning of the automotive industry (Waheed et al., 2014). The automotive industry is 



 DISSERTATION  9 
 

composed of fast-changing needs of customers reflected by the dynamics of demand patterns that 

are observed as a threat to predict future requirements. Dharmani et al. (2015) discussed on the 

value of current statistical forecasting tools in enhancing the planning and decision making and 

having knowledge of the whole market. The value of suitable forecasts is encouraged by the 

Institute of Business Forecasting and Planning. This is because reduction in forecast error by 

only a point reduces the average saving rate. However many car manufacturers still base their 

forecast predictions on traditional and local tools that may not be able to manage their rising 

complexity. 

The decision to purchase a car is determined by extensive information search by the customer, 

although there are other factors that are relevant in the new car buying process, including the 

national culture. However, various sources, such as professional dealerships and personal 

contacts, are considered before the decision of buying is reached. Much attention is given to the 

presence of a time lag between the attention given to a car model on the internet and the purchase 

decision made by the decision that requires to be emphasized in the sales predictions. 

Researchers like Artola et al. (2015) explain that the common use of internet in recent years has 

changed the performance of traditional activities including the way how financial transactions 

are conducted together with the online buying of products. In addition, Ernst and Young (2015) 

explored that customers put much time in online searching for new car trends before buying a car 

in comparison to any other product (Goel et al., 2010). The decision of purchasing is always 

made in the store. Thence, the assumption that the interests of people are revealed in their 

behaviours online, and in the main words they send to search engines. Prediction accuracy has 

been improved through various fields of application by adding web data within a model. 

However, internet data, especially Google Trends, is rarely applied in forecasting sales that 



 DISSERTATION  10 
 

might arise out of issues of reliability and validity concerned with big data. Nevertheless, 

accessibility and volume of internet data act as a suitable solution to solve the increasing slow 

development in approaches of forecasting and help decision-makers to solve complex problems 

within the changing environment in the automotive sector.  

1.3 Research objective and research questions 

Comparing the search engine-based prediction of new car sales in India and South Korea has 

extraordinary potential. India produces about 3.8 million passenger cars annually, and South 

Korea about 3.7 million passenger cars per year (statista, 2020). This implies that these two 

countries produce approximately equivalent amounts of passenger vehicles annually, and both 

countries are considered to have different cultures according to the habits of consumers. In this 

paper, quantitative research is employed to assess the relationship between the interests of people 

for a specific car model and the data of new car sales. This is done using various linear 

regression models and analyses that are cross-sectional.  

The main goal of this thesis is to determine whether Google Trends tools are efficient forecasting 

for New Car Sales in India and South Korea. The Google Trends tool enables customized search 

query data that connect to a given time frame, country and relevant keywords. Although 

identification of the difference in internet searches across different countries is important in 

forecasting sales, less attention has been given to this issue. The paper additionally investigates 

the prophetic power of best pause in programme knowledge as some authors previously verified 

the Twitter data and different social media phenomenon. In this research, time lag refers to the 

time between product information research on the internet and the final decision made on 

purchases. The incorporation of time lag into the model helps in quantifying inaccurate 



 DISSERTATION  11 
 

prediction of the model in India and South Korea. The current high value of search engine data 

by determining new relationships and patterns significantly increases sales performance and 

forecasting on new car models. Besides, the practical value of free internet data (Google Trends 

tools) as a complementary tool to determine economic variables across countries is equally 

reviewed. This research study also intends to create search engine raw data awareness to other 

researchers and decision-maker and provide ways on how to handle issues concerned with 

internet data that are reliable and valid. Taking theoretical framework into consideration also 

improves the prediction value that enables in determining the difference in a time lag of the 

buying process that is related to the price of cars, the segment of vehicles and the respective 

model. This study emphasizes Google Trends as a complementary tool for forecasting new car 

sales and car purchasing decisions made by people in India and South Korea. This is intended to 

develop insights that allow positive response to demand volatility across countries. During this 

study, the specific research question and sub-questions below will be answered. 

1.4 Specific Research questions 

 What impact does the state of economy have on Google Trends’ ability to predict new car 

model sales? 

 What is the relationship between the state of the economy and car prices in India and 

South Korea? 

 Does the seasonality presence impact the ability of Google Trends to predict volume as 

well as prices of cars in both countries?  

 Does the average time lag length differ between high-priced and low-priced cars in both 

India and South Korea?  



 DISSERTATION  12 
 

 How does the accurate prediction of new car model sales differ in South Korea and 

India? 

1.5 Significance of the study 

This paper encourages the use of theoretical frameworks together with statistical models to 

predict and gain results which are of benefit. Analysis of the customer journey and the state of 

the country together with the Google Trends analysis relates the research streams. The literature 

of Google Trends, research stream of marketing, and the cross-country research promote 

possibilities of new research. This thesis improves the validity and accuracy of prediction 

models, including internet data and thus, it contributes to literature prediction. In addition, the 

raw and unprocessed search engine data is unsatisfactory, and so this report calls for the attention 

of researchers to make adjustments within the dataset to minimize the effect of random 

observations (Jerath et al., 2014). The study assesses the cross-country differences in the 

application of new tool Google Trends’ properties that extends the literature of Google Trends 

within India and South Korea for the very first time. This contribution paves the way for further 

research that may discover more about the differences found in cross-country comparison along 

the customer journey (Kinski, 2016). The study also provides information on Google Trends 

literature by investigating the search query predictions value on the level of products that result 

in proper applicability and forecasting. By executing more search queries, this research paper 

gives scholars who intend to conduct sales forecasting with advanced variables to reduce 

forecasting errors in their future studies. Furthermore, the determinations of a time lag between 

internet data generation and sales data occurrence encourage the use of engine data research and 

researchers on social media also become well versed with the phenomenon. The pros and cons of 



 DISSERTATION  13 
 

internet data within the field of academics and research are related to the previous survey that 

enables researchers to assess the analysis appropriately. The study further outlines some 

guidelines on how to enhance the reliability, validity and value of Google Trends data to improve 

the accuracy of economic variables prediction in the study.  

The study gives decision-makers within the automotive industry an evaluated tool to use when 

making cross-country forecast and thus reduces the usage of outdated and traditional approaches 

while forecasting. Also, the Google Trends tool can be applied as an innovative and 

complementary tool to increase the demand for new car models by adding an updated and 

efficient search engine data within the forecasting model. The patterns identified in the internet 

data are specified for some car models and countries and acts as a catalyst variable to justify 

changes in the planning capacity of a company (Dharmani et al., 2015). The concept of 

prediction depends on lead time that is required by the customer to search for information before 

making the final purchase decision. The use of search engine data to understand consumers’ 

interests brings about a competitive advantage with competitors until they recognize the 

relevance of internet value (Gensler et al., 2017). However, a comparison of search engine-based 

prediction for India and South Korea particularly helps decision-makers to assess and analyze 

which country and car model is less valuable and less important. The obtained results support the 

prediction of customer demand each day that lead to increased company savings and serve as a 

method of managing demand volatility (Gensler et al., 2017). Better forecasts for specific new 

car models support car producers to negotiate with their suppliers and customers. Suppliers tend 

to increase car prices depending on the customer’s accuracy in capacity planning (Dharmani et 

al., 2015). Car model classification into small size, medium-size and large luxury vehicles helps 

in determining the differences in the time of the car buying process. From the marketing 



 DISSERTATION  14 
 

perspective of car producers, the relationship with customer journey has played a role of 

increasing the understanding of their customers in India and South Korea. The consideration of 

customers’ interests is relevant during marketing activities to reach further customers, especially 

during most influential periods on decisions of buying. The emphasis of the economy’s state 

enables practitioners to be aware of operating in the dynamic purchasing decisions of cars as it 

adds a positive trend in marketing campaigns. The analysis of Google Trends is suitable for 

every car model portfolio and various countries through carrying out a few changes in the data 

request of Google Trends. In other words, this paper helps decision-makers plan and implement 

decisions past the automotive industry to determine the benefits and the limitations of huge 

amounts of data that is created on the internet.  

Chapter 2: Literature Review 

2.0 Introduction 

The chapter reviews the literature about the study variables. The study variables, in this case, are 

the production, sales, supply chain of new model cars and the overall automobile industry in 

selected countries of South Korea and India. Different studies from various researchers and 

publications or statements from Google Trends aid the development of reliable literature in this 

case. The literature review is a vital step in the development of research work. The effectiveness 

of the entire research project is deeply associated with literature studies. It guides the researcher 

towards its research objectives. Selection of the articles, journals and books are the key tools to 

develop a constructive research paper. According to Alvarado-Herrera et al. (2017), a review 

process can be categorized into five steps or three steps. However, most of the research studies 



 DISSERTATION  15 
 

prefer to apply five stage models for detailed coursework. The five-step methodology is 

interactive and comprehensive in nature. Wolfswinkel (1967) is the developer of the five-stage 

model. He believes that literature review is subject to the grounded theory method. The theory is 

comprised of five different steps First, find out relevant research papers; second, select pertinent 

literature; third, analyse the selected literature, and fourth, present findings. Finally, the fifth 

option is to enlighten the insight view of the study, including the conclusion. 

It has been observed that scarcity of the proper research paper is a considerable concern for any 

researcher. The literature work is restricted to relevant academic journals, articles or textbooks 

available in the last decade. The current paper is focused on discussing the first three aspects of 

the five-stage model. Meanwhile, the study includes the proposed review technique in support of 

the key literature study. For instance, the study does not include the data related to social media, 

including Facebook or Instagram, while considering the field review developed by the Google 

Trends Applications. This technique is followed to narrow down the scope of the research work 

as well as the volume of the literature papers. Therefore, the researcher needs to invest 

comparatively less time in the literature study along with that appropriate dataset is required to 

precede the research project: the database EBSCO data, Google Books, Scopus and Google 

Scholar. In the field of the research work, Google Scholar plays an essential role as it vividly 

mentions the author’s name along with the publication.  The detailed introduction of the research 

paper helps the researcher to find out the appropriate research material to leverage the research 

area. A set of keywords and their combinations are used to select the most useful research 

literature. Most predominant keywords are Google Trend Prediction, Google Trend Analysis, 

Customer Journey, Kotler’s Five Stage Model, Cultural Influences on Consumption, Purchasing 

decision method, Cultural differences, India, South Korea, Hofstede’s Cultural Dimension, 



 DISSERTATION  16 
 

Forecasting, Web Data Predictions and Predictive Analysis. With the help of these key search 

words, the researcher is expected to find out appropriate literature studies. This research process 

is an essential task to cite significant research articles for the current research paper. The third 

step is dedicated to select suitable articles as per the need of the study. The in-texting process is 

comprised of backward and forward (Bartels & Onwezen, 2014). A research paper is generally 

focused on the abstract, introduction and a summary portion of every paper. This provides a 

sufficient idea about the paper. In this way, the researcher can get an overall idea of the paper 

and hence, the analysis is expected to be relatively strong. The role of comparative analysis is 

significant to evaluate the significance of the research project. It explains how one research study 

is different from another study. An appropriate differentiation between the studies raises the 

importance of the research project. The current literature survey is developed on a number of 

scientific articles, books, internet resources and insight resources available on the online 

platform. 

2.1 Systematic Literature Review 

 Currently, the automobile industry around the globe is entailed with various risks and 

opportunities in the growing mature markets of new model cars. More so, profitable growth has 

become harder to acquire because of several challenges arising from the supply side to the 

consumers. Economies, especially of India and South Korea, have many competitors, much 

capacity, as well as overlap and more redundancy (Smith et al., 2013). However, the automobile 

industry in identified countries has been challenged with the problem of rising global price wars.  



 DISSERTATION  17 
 

2.1.1 Production of new Car models 

The big manufacturers of cars involve the production facility in various markets to be produced 

mainly for exports to other markets. Most important to note is that a big automotive industry in 

the country does not depend on a single factory which exports the new model cars to other 

economies (Lassen et al., 2014). Besides, new car models are not identical in several markets of 

countries. Many cars have a similar brand, technical, or design features, but differ according to 

the country. Cars are different in various economies due to the nature of customer demands and 

pricing strategies. For instance, Indians have higher incomes than South Koreans, which means 

that the needs of the customers as well as affordable cars (Smith et al., 2013). Also, in South 

Korea, customers consider more space in cars, which in turn helps the new car models to be 

successful (Lassen et al., 2014). On the other note, small cars are the most popular in India. This 

is because India has limited parking spaces due to its large population (Stein & Ramaseshan, 

2016).  Besides, it may not be possible for a high-volume market to initiate big cars to each and 

every part of the world. Therefore, different car manufacturers and suppliers should research the 

preferences of customers in this changing market. Effective research helps the car manufacturers 

to produce car models which are preferred by consumers. However, completely second-hand cars 

are mostly renovated and produced in various countries, though in small volumes. Such a 

situation is a barrier to exporting cars in several countries since they are demanded on smaller 

volumes (Waheed et al., 2014). Basing on large markets, the model of sales is high, which in turn 

requires several companies to own their own plants to produce and supply the cars.  

Due to the varying customer demand and sharp competition, the process of developing the 

product has been instrumental compared to product architecture (Wang & Park, 2015). However, 

the cycles for the products continue to shorten as many companies gain simultaneous 



 DISSERTATION  18 
 

engineering strategies. Also, some advancement in computer-aided engineering tools is acquired 

to do what physical prototypes do, as well as processes of testing. Different automakers in both 

South Korea and India are attributed to many responsibilities to design the products, as well as 

allowing the production base to change to strategic locations for reduced costs (Wang & Park, 

2015). Besides, due to inadequate standardization, several tiers during procurement of cars are 

not reduced. Some domestic matters are supposed to be taken into consideration especially when 

the model is replicated with the new physical development of the product. Therefore, such 

modifications are considered as legal, regulatory procedures and liabilities (Waheed et al. 2014). 

Moreover, there are several technical advancements on modules, especially the functional units 

which are self-contained with some standardized interfaces to act as blocks for building in 

different car models or any other product. However, the supply chain process in the automobile 

industry can be reshaped using modularization. This is because such designs of the components 

are shifted gradually to top supplier companies (Wang & Park, 2015). The Shifting of the design 

components helps the suppliers to increase efficiency, as well as reducing costs. However, the 

survey by an international motor vehicle program found out that there is an elusive cost saving in 

new car model production (Waheed et al. 2014). The architecture of automobile retarded the 

modularization probability of success leads to a visible module advantage. However, different 

factors affect the supply chain to operate towards modularization process, which includes the 

efforts of automakers to change investment risks to the producers. This, therefore, increases the 

usage of information and communication technology amongst vehicles, where the consumers are 

likely to get interests towards in built-to-order cars. Many researchers found out that customers 

gauge production from the ‘model’ and ‘make’ viewpoint only in reality, but the production of 

automobiles is considered as dependent on supplier layers to alter outputs for real assembly. 



 DISSERTATION  19 
 

Besides, several automobile companies focus on the activities of assembling only whereas some 

are entailed by the vertical chains. The automobile industry is entailed with the great planning 

horizon, as well as high costs which are involved with the new car models (Waheed et al. 2014). 

The level of economies of scale in the automobile industry is much involved with technological 

flexibility to continue producing several models in a similar platform. Most important to note is 

that the core technological issues are fundamental in increasing competency and energy 

efficiency in both external and internal combustion engines. Such energy efficiency and 

competency in internal combustion engines are essential in combating vehicle weights, 

incorporation of high-tech safety features, among others (Wijnhoven & Plant, 2017). In addition, 

the standardization efforts are meant to reduce fuelling costs as well as maintaining the costs of 

production at lower measurements.  

2.1.2 Supply chain of new model cars 

Most of the automobile companies currently have adopted global strategies in their operations. It 

is clear that the growth of the transplant in 1990s accumulated the establishment of many 

competitors in the entire world (Wijnhoven & Bloemen, 2014). By putting much emphasis on 

interchangeable and common platforms, an automobile is likely to alter the easy and faster 

deployment of new strategies to achieve low costs from the products. This is essential to tailor 

vehicles to penetrate the tastes and preferences of the populace in the selected countries 

(Wijnhoven & Bloemen, 2014). However, proper product differentiation should be assured in 

order to attain a high-efficiency level, as well as managing the equity of the brand (Stein & 

Ramaseshan, 2016). It is also clear that great automakers are currently operating worldwide. Due 

to the established investments, the industries for automobiles have tried to replicate the structures 

of the supply chain, seeking for the suppliers to be allocated in new places especially near the 



 DISSERTATION  20 
 

plants (Shafi & Madhavaiah, 2013). Most important to note is that the auto industry supply chain 

has been varying over the years. Furthermore, the players in original equipment manufacturers 

around the globe have increased their focus on assembling and designing the operations and 

servicing after-sale markets. Therefore, after-sale markets are preferred because they deal with 

small numbers of big suppliers (Shafi & Madhavaiah, 2013).  

Furthermore, the process of the supply chain has morphed in different sub-system integrators, 

commodity layers, and component makers. However, there is increased risk-sharing segregation 

which was termed as cost pressure by the car manufacturer (Wikström et al. 2016). It is clear that 

tier1 suppliers are engaged with the increasing risks from big automobile companies to shift the 

cost pressure to another supplier to focus on the sub-components production only. Figure 1 

clearly demonstrates the process of shifting the cost pressure to other companies. Generally, 

vehicle suppliers are categorized into some specific groups (Wikström et al. 2016). These groups 

include; global standardized manufacturers systems, systems integrator, components specialist, 

and suppliers of raw materials (Shafi & Madhavaiah, 2013). Many vehicle companies including 

Toyota, Renault and Volkswagen think that some mono-suppliers’ strategies, such as that of 

Ford, are bad but portraying large suppliers are considered better strategies. Ford makes sure that 

the suppliers are pushed to own the equipment and specific strategies to push risks involved with 

fluctuations of car models. Besides, suppliers must choose their schedules of amortization when 

issuing prices since the payback period for the investments is included in price (Shafi & 

Madhavaiah, 2013). Globally, companies like Renault and Volkswagen wish their suppliers to 

fully invest around their plants, as well as transferring information to indigenous players. Most 

important to note is that automobile companies advocate for quality standards, as well as 

affordable prices to develop the contracts between them and the suppliers. Generally, the value 



 DISSERTATION  21 
 

and length of the contract should be related to the targets of reducing prices that are supposed to 

be admitted by the suppliers (Shafi & Madhavaiah, 2013). On the other note, suppliers and 

assemblers of cars are also entitled to propose some methods which may yield similar results like 

the automobile companies.  

Figure 1: Supply chain in the automobile industry 

 

2.1.3 Pricing 

Pricing of new car models is difficult since it is affected by fixed costs, technological aspects, 

and economies of scale. Also, consumer demand and competition among different companies are 

fundamental in the process of pricing (Hu et al. 2016). Recently, most of the companies in the 

automobile industry focus on pricing strategies to determine whether the company is to survive 

for a long period in the business. Among the pricing strategies, the price reduction is considered 

as an alternative approach to gain popularity and survive. Automobile companies require a lot of 

decisions at each phase of production and selling. The production and selling phase starts from 

supply chain and production, as well as management factors to aid negotiations with the dealers. 



 DISSERTATION  22 
 

Most important to note is that price is considered as an important factor that aids the variability 

of sales for the goods and services in a significant way (Wu & Brynjolfsson, 2010). Several 

companies need effective policies to be practiced in an intelligent manner to yield better 

decisions (Somervuori & Ravaja, 2013). Besides, when prices are reduced, it does not 

necessarily increase profits but creates the popularity and reputation of the company. Therefore, 

the combination of quality preservation and product marketing decisions are essential when 

reducing prices to penetrate customers’ tastes and preferences (Somervuori & Ravaja, 2013). 

Reduction of car prices attracts disloyal clients who do not value the company, but rather through 

price reduction in the brands. The companies’ lifecycle helps them to gain many returns from the 

newly introduced car brands, which in turn reduces the costs of obtaining cars.  

However, different companies use strategies of pricing entitled to various segmentation of 

products bearing in mind the changing expectations of the customers. Furthermore, automobile 

companies use innovative methods basically to increase profits without exploiting esteemed 

customers (Risius et al. 2015). Most important to note is that prices are adjusted in relation to the 

purchase volume, qualities, loyalty, profitability, and potential for development factor. When 

fixed costs are high, companies need to improvise with many and practical strategies or models 

to make proper decisions about total output (Wu & Brynjolfsson, 2010). It is also clear that many 

brands are aided by the level of economies of scale to put down production costs. In case when 

the carmakers offer prices to their brands depending on the average cost, several deviations are 

expected to alter many outputs in the short run to influence the pricing strategies without 

considering factor cost changes. It is also believed that fixed costs are subsets of total costs, 

which means that they can change into variable costs of production (Risius et al. 2015).  



 DISSERTATION  23 
 

Among the car manufacturers, there is evidence that they face collusive pricings in their 

operations. However, the law about price disclosure in South Korea favoured several players to 

collaborate, which in turn led to collusion among themselves trying to fight for quality-adjusted 

prices. Moreover, big car companies were inhibited by the law in the struggle of providing price 

discounts. It is clear that the arguments provided by ability-motivation approaches or theories 

indicate that highly concentrated and stable environment markets with cooperative car producers 

create sustainability in the business. Research has shown that a market with current customers 

who are loyal, there is always motivation which helps them to aggressively create a base for new 

customers (Hu et al., 2017). Several firms believe that positive loyalty benefits from the 

customers enable them to survive in the business for a long period of time. Most important to 

note is that when the client’s loyalty in the market increases, several gains are likely to increase 

as the market shares are altered by aggressive competition to maintain the profit margin. In 

general, car firms are entitled to have the motivation to cooperate with prices (Scott & Varian, 

2013). The competition from foreign companies and eradication of local concentration has, 

therefore, changed the uniformity of prices. In summary, variation in pries helps the indigenous 

manufacturers in the specific countries to always coordinate with the counterpart firms during 

the process of setting prices for different car brands.  

2.2 Buying Decision Models 

2.2.1 Five-Stage Model of the Buying Process 

 Consumer behaviour is the key central idea of this model. Bartels and Onweze (2014) 

classify the buying decision into five different stages, including information search, evaluation of 

alternatives, purchasing strategy and purchasing behaviour. This model is known as Kotler’s 



 DISSERTATION  24 
 

Five Stage model. Every stage of the model is assigned to fulfil the different objectives of the 

research work. It has been observed that every stage of the model is interrelated to each other.  

 

 

 

 

 

 

According to Kotler (2012), research problem needs to be recognized at the beginning of the 

research project. The model is marked as a traditional decision-making model. The model leads 

the customers to reach their final goals to buy their products and services. The consumers realize 

the need for purchasing goods and services step by step. Most of the consumers follow this 

strategy to optimize their consumption goals. Meanwhile, the intensity of the choice and 

preferences guides the customers to identify the objectives behind product choice. In this regard, 

another research theory by Engel, Kollet and Blackwell (1968) can be considered as the 

fundamental step of the five-stage model in order to analyse consumer behaviour. The model is 

comprised of two comprehensive stages. The first stage is dedicated to analysing the impacts of 

the marketing tools on consumer choice. These marketing tools include television, online social 

platforms and newspaper. After collection of the data, consumers enter into the second stage in 

the decision-making process. This second step analyses the external impacts of the influencing 

factors (Dhandra & Park, 2016). The information provided by the manufacturers for the 

customers plays the influencing role underlying the consideration for purchase. Branding of the 

Problem 

Identification 

Search 
Information 

Find out 
alternatives 

Purchasi
ng 
strategy 

Purchasing 
behavior 

Figure 2: Kotler’s Five Stage Model 

Source: (as created by the author) 

 

 



 DISSERTATION  25 
 

product is the important parameter of the second stage. Branding reflects the standard of living of 

the consumers. 

 Meanwhile, ample information provides data on the alternative goods of the desired 

products. It has been observed that substitute goods play a significant role in determining the 

price of the desired product (Dhandra & Park, 2016). If many substitute products are available in 

the market, the consumer will get sufficient opportunity to transfer their choice of the products 

from one good to another good. For example, tea and coffee are substitute goods. In this case, the 

customer will prefer to buy coffee is the price of tea increases. A little change in the price of the 

goods brings a significant impact on the purchasing amount of the desired products. It is noticed 

that closed substitute goods are highly-priced elastic. 

In this regard, some companies offer an experimental opportunity to realize the practical 

utility of the product. This practical experience genuinely helps the consumers to acknowledge 

the multiple characteristics of the product and services. This inclusion of different sources of 

information leverages the buying power of the consumers and reduces the risk factors related to 

the products. Contextually, the availability of the large size information builds a strong brand 

image for the company (Dwivedi, 2015). This ensures the company’s strong reputation in the 

product market and hence, the customers feel confident about purchasing of the desired product. 

The initial set of information reveals the key important characteristics of the product. The next 

step is counted as the third stage of the Kotler’s model. The latest modification of this model 

depicts that the multiple stages of information gathering depend on how a consumer collects the 

data. It is a non-singular approach. In this case, the diversity in information is highly expected.  

Therefore, the customers need to wise while making their purchasing decision for the product. 

Both the negative and positive consequences need to be evaluated while finalizing the buying 



 DISSERTATION  26 
 

decision. The buyer should purchase the products when the benefits surpass the negative 

consequences related to the consumption of the products (Gupta & Sahu, 2015). The analysis 

needs to cover both the upside and downside aspects of the buying decision. Meanwhile, 

consumers often suffer while gathering information in an appropriate way. The post-purchase 

actions are greatly associated with pre-purchase actions. Keller notes that the customer’s level of 

involvement is highly appreciable to get a successful outcome from the purchasing decision of 

the customers. The most effective utilization of consumer expenditure is only possible when the 

good is consumed in a proper way. 

 Moreover, mention that buying decision gets influenced by the purchasing pattern of 

other consumers, including friends and family members. Their choices and preferences are 

influential factors. The consumer desires to buy products which are not urgently required to meet 

the daily. The linear integration of the information gathering and analysis process is highly 

necessary to develop a link between the research objectives and desired outcomes. This requires 

additional adjustments tom recognize the future assessment for the next purchasing activity. This 

dynamic transmission helps the customers to deduce an effective purchasing strategy for the 

products and services. 

2.3 Cultural Dimensions in India and South Korea 

2.3.1 Cultural Difference and Consumption Behavior 

The consumption pattern is significantly influenced by cultural values. Culture is the identifying 

parameter for every country. The people develop their choice and preference for the goods and 

services influenced by the culture. The people in India possess different taste and preferences for 

goods and services with respect to South Korea. Data reveals that the consumers in India expense 



 DISSERTATION  27 
 

more for spices than the South Korean customers. On the contrary, the consumption level of fast 

food is relatively higher in South Korea than India. Hur et al. (2015) urge that this difference in 

food consumption reflects the cultural influence on food habit. Meanwhile, it is observed that 

South Korean customers are more product conscious in respect of the Indian consumers. In 

South Korea, consumers spend considerable time to know the features and usefulness of the 

products. On the basis of limited availability of the information, the study evaluates the cultural 

impact on consumer behaviour through the Hofstede Model. According to the model developer, 

the cultural dimension is the influencing factor to determine the intensity of the difference in 

consumer behaviour. As South Korea is a technically advanced country, people have a growing 

tendency to buy electronic gadgets. This implies that product availability and consumer choice 

are intensely correlated across the two countries. As far as the model is concerned, the usage of 

technology can be mentioned to highlight the difference in the consumption pattern. South Korea 

is far way developed in the technology sector. On this account, the consumers of the country are 

accustomed to the advanced technology, which in turn increases the consumption of the 

electronic goods. 

In the context of India, the traditional culture is the dominant feature. For this reason, spending 

on fast food is relatively less than the other sample country from Asia. The adoption and 

innovation of technology are strongly dependent on the cultural pattern of the country. Several 

studies have verified that culture has a significant influencing factor in determining the lifestyle 

of the people. Data asserts that average consumers in South Korea are wealthier than India. The 

great expense on the luxurious goods reveals that South Korean consumers are more 

sophisticated than India.  



 DISSERTATION  28 
 

2.3.2 Hofstede’s Cultural Dimensions 

According to Hofstede, the model has five important dimensions from a cultural perspective. 

These five different aspects include individualism or collectivism (1), femininity/masculinity (2), 

uncertainty avoidance (3), power distance (4), long- or short-term coordination & (5) 

indulgence/restraint. The cumulative index number is assigned from 0 to 100. The study has been 

performed by the IBM employees. One hundred sixteen questionnaires are considered on the 

basis of seventy-two countries from 1967 to 1973. The model has an effective approach to the 

management and marketing research program. The model compares the mean value of the 

consumption pattern across the sample countries. 

Along with that, the model attains to correlate the different cultural dimensions to justify the 

objective of the model. Despite the Hofstede model, several models have been developed to 

analyze the relative disparity in the cultural behaviour in respect of the individuals and 

organizations (Valaei et al. 2016). The next section briefly illustrates the significance of the five 

different models.  

 



 DISSERTATION  29 
 

 

Figure 3: Hofstede's Cultural Dimension: India and South Korea 

Source: (Hofstede-insights.com, 2020) 

Considering the Figure 4, Hofstede’s Cultural Dimensions in India and South Korea implies that 

there is a significant difference for uncertainty avoidance, long-term orientation, masculinity and 

individualism. The indices for power distance and indulgence are almost the same for both 

countries.  

Power distance: This is regarded as a first dimension which explains that unequal power 

distribution makes some people powerful and the rest of the people gets controlled by those 

powerful persons. It is observed that powerful organization exhibits a large social gap between 

employees and employers in India as compared to South Korea. The organizational hierarchy 

triggers a great challenge to the productivity for India. The dominating attitude of the higher 

authority on the lower authority reflects the country’s score of 77.  The subordinates working 

under a boss felt a great distance in the workplace. This distance is an expected feature of a 

standard organization. This helps to maintain a cultural difference with respect to the 

designation. This is expected to provide a successful work environment for the employees. 

According to Yadav & Rahman (2017), most of the younger employees own relatively less 



 DISSERTATION  30 
 

power as compared to experienced employees. Purchasing tendency of high pricing car of the 

company owners clearly reflects the power distance among the society. 

Individualism: Individualism stands for the moral stance, ideology and social outlook. It can be 

defined as the right of the individual to follow its self-realization. It is an evolving concept for 

the self-realization power of individuals (Howes et al., 2016). This leads to the betterment of the 

individuals and further for society. In this aspect, India scores 48, and it is considerably higher 

than South Korea’s score of 18. This further verifies that the disparity is driven by the cultural 

choice of the inhabitants (Sreen et al. 2018). According to the model, the characteristics of the 

consumers are influenced by cultural practice. Along with that, rational thinking is another 

parameter which plays an effective role to justify the consumption pattern provided with the 

dimensional aspects of the cultural activity. 

Collectivism: It is the alternative concept of the individualism. In terms of the collectivism, the 

integration among the people is the essential criterion. The people are cohesive in a group. The 

decision is taken through a collective method. Zhao et al. (2016) urge that individual perspective 

is not an important factor to assure the development of the society. The dependency on the other 

people strengthens the bonding among the people. The loyalty exchange is the prime 

characteristic of the collectivism theory. It develops an unquestioning loyalty between the 

customers and sellers. Therefore, it prefers high dependency than self-supporting consumption 

pattern. The self-interested group is not the fundamental concept of societal development (Fumo 

& Biswas, 2015). Considering the fact, both South Korea is strongly emphasizing on the 

collectivism mechanism as it scores lower in the individualism index. For example, the consumer 

choice in South Korea is majorly controlled by the central government and therefore, the 

consumer choice gets compromised. 



 DISSERTATION  31 
 

Masculinity:  This term referred to as the dominant values of society. The acquisition of wealth 

without considering the utility of others shows the impact of the masculinity. The dominating 

attitude is the reflection of the masculinity (Billing et al., 2014). For example, men are passionate 

about buying two-wheelers instead of beauty products. On the contrary, the female prefers to 

purchase cosmetics even if costs extravagant prices.  Here, India exhibits greater gender disparity 

on the consumer choice and preferences for the goods and services as its score is 56 in terms of 

the masculinity index. In contrast, the influence of masculinity index is lower for South Korea. 

According to Dwivedi (2015), the country having high femininity score exhibits a better quality 

of life. People care for each other when the country is dominated by the feminism concept. The 

considerable difference between India and South Korea in terms of the masculinity results in the 

wide gap between the power concentrations. According to economists, masculinity indices score 

high due to the overwhelming wealth inequality. In the context of India, inequality in wage and 

wealth distribution is a significant concern for the economy (Le at al., 2016). On this account, it 

can be mentioned that the majority of the Indian families own a family car.  

Uncertainty prevention: Second highest difference gets reflected by the uncertainty avoidance. 

In this case, South Korea has again scored a comparatively high level with 85, whereas India has 

scored 40. This indicates that the customers in South Korea are risk-averse than Indian 

customers. The people in South Korea are not willing to take risks while purchasing products 

from the commodity market. Uncertainty avoidance countries exhibit less intention to purchase 

online products provided with a 5% level of significance. According to Dhandra and Park 

(2016), online purchasing activities are a comparatively, less attractive option for risk-averse 

customers. This dimension highlights the willingness of the people to avoid the uncertainty 

related to the product. This uncertainty reflects the ambiguous nature of the consumers towards 



 DISSERTATION  32 
 

their purchasing decision for the products. In the context of India, most of the consumers exhibit 

ambiguous nature while making the purchasing decision as its score point is 40 in comparison 

with 85 of South Korea (Hoffmann et al., 2016). Here, the low per capita income of the Indian 

people can be one of the reasons behind this uncertain nature.  

Long-term orientation: The score of long-term orientation has scored 100 for South Korea with 

respect to 51 in India. This difference implies that individualism is about two times than India. 

This score of indices implies that the people of South Korea are concerned about future 

consumption than Indian consumers. Economic resources are used in a more sustainable way 

when the country owns high long-term orientation score. This emphasizes the future perspective 

rather than on the current moment. This idea assists the sustainable development of the society 

and economy. The factors of the production are consumed in such a way that future generation 

will not be deprived of consuming all those products. For example, growing to emphasize the 

consumption of non - renewable products are aimed at maintaining the ecological balance in the 

environment. Henceforth, the consumption pattern needs to be changed to achieve a long-term 

goal. On the other, short-term orientation is subject to the personal stability. This aspect is 

developed from the perspective. Singh & Venugopal (2015) point out that most of the Indian 

consumers are reported to have less saving habits compared to South Korea. Therefore, it can be 

assured that India has less available resources to continue the investment process for the future. 

  



 DISSERTATION  33 
 

2.3.3 Cultural Differences in India and South Korea 

 There are several numbers of approaches which can be incorporated to analyze the role of 

the cultural values on the consumption style. The seven-dimensional approaches of Trompenaars 

examine the cross-cultural disparity across the countries. This measurement is acknowledged as 

Tightness – Looseness method, which is used to determine the strengths of social rules (Treviño 

et al., 2019). It further examines the tolerance power of consumers. This latest concept not only 

is based on the previous five-dimensional model also is concerned about politics and history. If 

tightness is high in a country, the social norms are relatively rigid. This model has been 

developed comparing the cultural differences across thirty-three countries (Montoya & Horton, 

2014). Contextually, the South Korean government is popular for its strict rules and regulation. 

From this point of view, the consumption pattern of South Korea is highly influenced by 

government intervention. On the contrary, India is a democratic country and consumers can 

freely express their choice and preferences for the goods and services (Gupta & Sahu, 2015). 

This cross-cultural dimension can be related to daily life-style. In case of strong life situation, 

people face limited options to express their choices due to income constraint or limited access to 

the products. On the contrary, weak-day life offers limited behavioural options to the customers 

in terms of comprehensive access to the products and services. This implies that individuals 

belong to the tight countries lead strong everyday life-style due to the strong intervention of the 

government. 

 However, several numbers of studies have found several loopholes of the Hofstede 

model. According to McSweeney, the small sample size is selected for the analysis of this 

dimensional model. Moreover, the numbers of questionnaires are not sufficient for model 

development. The concerned IBM survey was taken for two times in 1968 and 1973 for more 



 DISSERTATION  34 
 

than one thousand respondents (Jiang, 2017). The survey was conducted in Pakistan over a 

hundred employees of IBM. However, this survey raises questions on the validity and reliability 

of the model. The author also explains that every employee selected for the sample survey is the 

representative of Pakistan, and all of them are working under the same organization (Nyadzayo 

& Khajehzadeh, 2016). Therefore, focusing on a certain company is expected to influence the 

respondents. Apart from that, all the respondents are employs. The respondents can be 

unemployed, poor and elite class people (Beugelsdijk et al., 2015). They all are representatives 

of the entire nation. Considering the fact, the people from different section will contribute their 

significant values on the change in consumption pattern. Most importantly, people belong to a 

single country share the same values and preferences. In order to find out the degree of cultural 

influence, the relationship between national culture and national states needs to be examined. 

This important factor has not been emphasized during the development of the Hofstede model. 

Nonetheless, the deduced correlation model on the consumption pattern and culture was studied 

a long time ago. This study needs to be revised. The cultural difference depends on the change of 

time period. Therefore, the time frame is a vital aspect to observe the cultural changes with 

respect to the time period. The cultural difference needs to be treated as a quantified object to 

analyze the consequences of the changes in the consumption style (Dessart et al., 2016). In the 

case of Indian and South Korea, people have been experiencing robust changes in the 

consumption pattern. The growing consumption of the fast-food clearly denotes to the shift in the 

food intake pattern. Fast food is basically dominating the culture of the Western countries. Our 

sample Asian countries are traditionally fond of cooked foods. This shift in the taste of foods 

highlights the importance of the time frame. 

 



 DISSERTATION  35 
 

2.4 Google Trends Predictions and Forecasting 

2.4.1 Predictions and Forecasting 

Trend analysis is a popular statistical model to predict future values. This provides a broad idea 

which products should be bought or sold from the future perspective. Prediction provides 

possible logics in order to finalize the purchasing decision. As far as the study is concerned, the 

prediction is always not able to construct accurate decision regarding the purchasing habit. 

However, the appropriate application of the theoretical framework results in the effective 

outcome of the predictions. In general, the outcome of the prediction seems like an approximate 

outcome. It provides an expected outcome of the experimental work. The predicted value may 

exceed the original outcome due to the impact of the selected variables. The application of the 

statistical tool is crucial to deduce the desired outcome with respect to objectives. Meanwhile, 

the data mining system is another important aspect of future observations. In terms of Çifci et al. 

(2016), a predictive model is the best applicable model for the empirical data. A new observation 

can be added to the original data while computing the predictive model. They have vividly 

illustrated the difference between explanations and predictions. The explanations highlight the 

importance of the task, and the latter one analyses the outcome of the model. The empirical 

analysis evaluates the significance of the predictive model. The predictive model is used to test 

the hypothesis in order to support the theoretical work. It illustrates the constructive framework 

as per the requirement of the study. The vast explanation of prediction is required to justify the 

impact of predictive modelling. It focuses on the future basis of the sample observation. 

However, there is a fine difference between the prediction and hypothesis. The hypothesis is 

taken before initiating the research work, whereas the prediction refers to the outcome after the 

calculation is done. The understanding of prediction depends on the theoretical approach towards 



 DISSERTATION  36 
 

the model framework. It is important to understand the role of the predictive analysis as it 

reduces the uncertainty related to research work. 

Fritzsch et al. (2019) surveyed about whether Google trends are used to forecast and improve the 

sales and value of a certain product. The combination of standard time series and data about 

search query are fundamental in forecasting the sales of the product. Fritzsch et al. (2019) 

included the company’s search volume, and brand-related keywords, which are provided by the 

Google trends. The keywords in Google trends help in forecasting the sales of the product since 

they act as new predictors. The study used time-series data from January 2015 to 2016 

December, which helped them to get evidence about the essence of Google trends towards the 

performance of conventional models.  

Mavragan et al. (2018) assessed the tools, methods, as well as statistical models used in research 

about Google trends.  In the information overload era, the evolution of big data analytics is 

essential to solve and manage the data available to be interpreted. The study discussed the usage 

of web-based data about public health on the behaviour for individuals. Therefore, Google 

Trends is a common tool used for gathering information about fluctuating issues such as sales, 

public health issues, among others. Web-based behaviour previously was used in analyzing and 

monitoring the actual behaviour of the subject under investigation. Therefore, the stakeholders 

by using web-based behaviour also aid in predicting, assessing, as well as preventing the 

obstacles of production.  Mavragan et al. (2018) used a systematic review to report critical 

analysis of the methods and statistical models about Google trends, which were carried out to 

intervene with health-related issues from 2006 to 2016. The methods used in their study followed 

the guidelines of Meta-analyses by selecting the individual term “Google trends” from the 

database of PubMed.  



 DISSERTATION  37 
 

Forecasting is the cumulative research work on the available data set to predict the future 

performance of the selected variables. The estimated probabilistic value offers to illustrate the 

possibility of getting the desired outcome. The prediction analysis has an important role in 

measuring the future growth of the business, population and other important aspects of the 

economy. For instance, a prediction on the cycle sales can be like this. The sales growth can be 

hiked by 50%, whereas there is a 10% for which car sales growth can be declined for a certain 

period. Several numbers of factors play as a catalyst role in this predictive model. On this 

account, it is considered as one of the complex data structures under the forecast model. 

Referring to the time period, the forecast model can be classified into two categories, such as 

short, medium and long term forecast. The short term forecast is done for one to three months, 

the medium-term prediction is made for five to ten years, and finally, the long-term forecast is 

developed for five to ten years. According to the experts, the forecasting approach is relatively 

complex than prediction. The accuracy of the sample variables is essential to attain an accurate 

outcome. A proper justification of the forecast is derived from the work of the forecast model.  

The insight view of forecasting model assures that theoretical has the great potentiality to equip 

with the advanced predictive analysis (Zhang & Oczkowski, 2016). As far as the literature 

survey is concerned, both forecast and predictions have a different approach for new 

observations. Moreover, it distinguishes between out-of-sample incidents and forecasts for future 

prediction with the help of the supportive theory. Meanwhile, it can find out the impacts of new 

additional variables into the model. Predictions provide an outcome for the short-term instead of 

applying the probabilistic approach (Xu & Perron, 2014). The forecast technique underlying the 

current study incorporates the Google Trends Data. This data acts as an independent variable for 



 DISSERTATION  38 
 

the study. In general, both terms are used in an interchangeable way without considering the 

theoretical difference. 

Mavragan et al. (2018) used 109 published papers by focusing on their correlations, seasonality, 

and visualizations features which are used to determine the category of issues under 

investigation. The study examined the performance of Google trends by using time series 

analysis, the definition of the key terms. Mavragan et al. (2018) observed that 23% of the studies 

used data from Google Trends to assess seasonality, whereas 40% of the studies used Google 

trends to get modelling and correlations. Online queries monitoring and assessments provide 

critical insight into the researchers’ behaviours on how to make decisions about a specific matter 

under investigation.  

2.4.2 Google Trends data as a tool of now-casting and forecasting 

Internet data application is categorized into now-casting and forecasting measurements, which go 

beyond what other approaches, can perform. Now-casting, in this case, refers to the nature of 

prediction basing on current occurrences (Pieper, 2016). Now-casting helps the researchers to 

predict models which are based on short time period usually three weeks (Carrière‐Swallow & 

Labbé, 2013). Now-casting helps the decision-makers to timely use time series always weekly 

and on a quarterly basis (Labbe & Swallow, 2013). The published data sometimes provide rough 

estimation, which, therefore, yield unreliable decision making (Mitchell, 2011).  On the other 

note, data composition can be distinguished by using periods since some data do not appear at a 

specific period of time.  

2.4.3 Shifts in the Forecasting Context 

 Over the last two decades, the introduction of the internet made commendable changes in 

the forecasting context. Globalization and comprehensive forecasting model have also led the 



 DISSERTATION  39 
 

forecasting system to the advanced level with the help of a software program. However, 

deficiency in technical knowledge is often acknowledged as drawbacks of the forecasting 

context. The inefficient training system, coupled with a knowledge gap misinterprets the 

outcome of the forecasting model (Zhang, Wei & Zhou, 2017). Altogether, it reduces the 

accuracy of the forecasting result. According to an article, forecasting analysis is unable to 

improve the sales prediction on a car on the basis of data over a decade. This inappropriate 

outcome calls for considerable improvement in the forecasting system (Mestre et al., 2015). On 

this account, new statistical tools are invented to obtain more accurate outcome related to the 

sample objective. In nowadays, the overwhelming presence of the internet raises the need for 

predictive analysis of online products. This growing concept is known as “online product buzz”. 

The model expresses the interest of consumers for online products and services. This sort of 

forecasting analysis gets structured on the basis of the review, blogs and online resources related 

to products available on the search engines. It has been observed that the accuracy of the 

predictive model has been improved by a significant level considering the online resources. All 

the online data helps in developing an advanced sales forecasting through online interviews or 

reviews written by the customers on the online platform (Huang & Kitani, 2014). It allows the 

customers to give reviews after the usage of the goods. Henceforth, online reviews are expected 

to give a significant contribution to enhancing the predictive study.  

2.4.4 Judgment Forecasting as a Source of Inaccuracies 

 Several research papers have examined that sluggish improvement in the forecasting 

analysis results in the development of the complexities in the forecasting models. As a 

consequence of that, the desired outcome is not achieved as per the expected level. Lack of 

accuracy in the statistical tool misinterprets the statistical properties, which are the fundamentals 



 DISSERTATION  40 
 

of the model. Judgment is a fundamental step to judge human characteristics. The forecast will 

be appropriate if the experts do not get the available data as per the requirement of the model. 

Moreover, judgment is often reported to surpass the pragmatic outcome of the statistical analysis. 

Currently, more than 80% of commercial organizations are dependent on the forecast tool 

(Hendry & Mizon, 2014). Thereafter, the derived outcome gets adjusted in accordance with the 

objectives of the model.  The degree of adjustment is subject to the knowledge of the market 

experts and the requirements of the company. Biasness and misinterpretations are the most 

common consequences of the inaccurate forecast model.  

 Optimism bias is regarded as the psychological tendency to influence the judicial process 

towards a definite direction. The insight view of business reflects the large involvement of the 

personal perspectives, such as personal benefits, clients’ objectives and surrounding situations of 

the society and environment. The effect of optimism biases can be controlled through the 

selection of the accurate empirical database. Further, an effective comparison with a similar type 

of projects can help in fetching the desired outcome (de Carvalho et al., 2015). The expert finds 

out that political interference and organizational pressure misinterprets the survey outcomes, 

which are acknowledged as a part of strategic misrepresentation of the dataset. A survey 

conducted on Scandinavia explains that traffic forecasting outcomes get misinterpreted due to the 

presence of the incentive (Howes et al., 2016). There should be sufficient logic behind the 

initiation of the incentive policy in order to expand the capacity of the organization. The 

negotiation of the funding often misrepresents the outcome of the forecasting methods. 

Contextually, the reward policy can be introduced to encourage the analysts to calculate an 

accurate forecast. The managers are generally bribed to leather the original outcomes. The less 

number of inaccurate sources the high chance of transparent results.  Considering the facts, the 



 DISSERTATION  41 
 

analysts always prefer to use transparent statistical tool to investigate the role of the likelihood 

functions and outliers (Lana et al., 2018). A lot of research works have been formed to highlight 

the role of the forecast methodologies to justify the adjustments of the forecast outcomes. It has 

been found that 99% of the forecast model uses statistical tools to verify the predictive outcome. 

The common approach used in the forecast model is the combination of the statistical inference 

and judgmental adjustments. Along with that, the respondent outcomes are incorporate as to 

cover all the influencing factors into the discussion. On average, it has been noted that a 10% 

error in the forecasting method corresponds to the human error. The use of Google trend 

application needs human interaction. The data extraction and interpretation are made by a 

human. On this account, a considerable portion of mistakes is expected in terms of human errors. 

These potential errors must be ignored while interoperating the forecasting analysis. Therefore, it 

is mandatory action needs to be taken to tackle the computation error related to the human 

research work (Brugler et al., 2015). Meanwhile, data extraction can be modified and controlled 

to avert the misinterpretation of the forecast module. 

Prediction of economic variables and private consumption 

 A paper observes that Google develops the plan for a holiday trip on the basis of the 

travel destinations and the model of transport available for the customers. In the same way, the 

housing sales and motor vehicle shop are analyzed on the basis of the searching trends 

(Vaijayanthi & Shreenivasan, 2017). The paper explains that there is a strong correlation 

between the search frequency for several countries from India to South Korea and the data on the 

actual visitors come for the tours. This available data set helps the experts to set the trend for the 

tourists mitigating the uncertainty along with affordable expenditure (Fatma et al., 2016). The 

sales data reveals that car selling is highly influenced by online research work. Buyers do 



 DISSERTATION  42 
 

intensive research work before purchasing a car. The paper has constructed a regression model to 

find out a relationship between the cars and light truck sales of the basis of twenty-seven 

predetermined car makers (Peng et al., 2014). The available sales data act as a dependent 

variable in this tidy. The paper has the Google prediction to analyze the sale trend of the 

automobile market and retail sector. Trend data accumulate sales data on a monthly basis. The 

Google trend analysis makes the researchers acknowledged with the market trend. In this way, 

the management can structure the business plan for the future. It has been observed that the 

regression model under Google trend can improve the trend performance of the selected 

variables up to 20% with respect to those models which are not using the Google data (Dale et 

al., 2014). Google trend analysis measures the prediction using the mean absolute errors. Two 

different types of comparisons get developed, excluding the Google data as well as including the 

Google data. 

 Another study reveals the simplicity and accuracy of the web search data are useful to 

predict the house sales data. This will help to determine the market price of the South Korea 

housing sector (Youn& Cho, 2016). The report states that around 35% of the absolute mean error 

can be expected without using the Google data, while 5% of the absolute mean error can be 

expected using the Google data (Andreano et al. 2017). Further, Google trend analysis is 

reported to generate useful data for the emerging market condition.  A set of research papers 

have pointed out that the merger of Google data with other available data resources develops 

multivariate data analysis (Lassoued & Hobbs, 2015). This predictive model helps in developing 

a long-term forecast up to 24 months which is ahead for the first time. In this context, a number 

of studies have criticized that the time horizon of the previous model is less than a few months. 

This analysis is developed on the basis of the data provided by the Federal Motor Transport 



 DISSERTATION  43 
 

Authority in Germany. The data is comprised of both foreign and local car producers (Jindal et 

al., 2019). The data includes seasonally adjusted GDP data, unemployment rate and petrol price. 

The report concludes that macroeconomic variables are able to outperform on the basis of long-

term. In the short-term, they have been failed to perform beyond the expectation. Nonetheless, 

parsimonious hardly considers the complexity of the long-term forecast with the help of Google 

data (Puška et al., 2018). 

The relationship between unemployment and Google keywords is studied by other scholars in 

India. The study is limited to four predetermined sets of search keywords, including 

unemployment office, Monster. According to the scholars, there exists a strong association 

between the unemployment level and Google queries (Shen et al., 2018). Henceforth, many 

studies attain to predict the unemployment situation on the basis of key search words used on the 

internet. The similar exercise is done for the unemployment level among the French youth. It is 

concluded that the intensity of the job search gets doubled at the beginning of the first ten weeks.  

Google Trends Predictions in Comparison to Traditional Survey 

The analysis of Google Trends prediction as compared to that of a traditional survey in 

research is not a new phenomenon. Several scholars and researchers have touched on the aspect 

in previous times, and it is still a topic of wide concern in the world of academia and research. 

With the traditional methods of data collection, there is always the need for a massive input of 

resources. These resources come in terms of time, logistics, materials and human resources hence 

making the whole process very demanding, expensive and labour-intensive (Mellon, 2013). 

These logistical implications are among the reasons that posed the urge for researchers to look 

for an alternative means of data collection. 



 DISSERTATION  44 
 

The internet has, in the recent past, proved to be a meaningful resource in collecting such 

data with the implication of a very minimal resource allocation to the process. Google trends are 

such a platform which proved to be a meaningful resource in the aspect of sourcing internet 

derived information and gathering data needed by scholars and researchers. It allows users to 

interact with internet-based data which is potentially useful and easily accessible upon demand. 

The process of purchasing a specific product or brand is influenced by several factors, especially 

price, packaging, design, consumer knowledge and marketing strategy such as celebrity 

endorsements (Sun et al. 2015). Most developed countries like South Korea have got mature 

markets for retail products. Indians and Koreans are always keen to understand the brand, which 

is prevailing in the market (Mavragan et al. 2018). Do people pose different questions about the 

prevailing brands such as how will the product fulfil their wants? Research has shown that 

consumers’ decision making on a specific product is based on the current information for both 

external and internal environment. The internal information about the product is gathered from 

advertisements, whereas external information emanates from market places or peers.  

Decision making of purchasing a product is influenced by different brand selection procedures 

for specific products. Most important to note is that brand selection is based on the nature of the 

cohesiveness of the product. Therefore, car companies have to focus on the packaging of their 

new cars to affect the buying decisions of individuals (Zhang et al. 2017). Some of the areas on 

which the automobile companies need to improve are; consumer perception, customer 

information, marketing strategies (e.g., celebrity endorsement), and design.  

 

 



 DISSERTATION  45 
 

2.5 State of the economy 

2.5.1 Relationship between GDP and new car sales in India and South Korea 

There exists a positive and direct correlation between the gross domestic products (GDP) of a 

country and its automotive industry. In this paper, GDP includes consumption, investment, and 

government expenditure within a specified time frame. According to Saberi (2018), the 

automotive industry is a knowledge-intensive and capital-intensive industry that is very 

important in the social-economic development of a country. In the current era, the industry is 

booming across the world, and so many countries have adopted the business of car production. 

However, the alignment of forces within the market of the automotive industry is dynamic (Geva 

et al. 2017). All economies growing at a faster rate, including India and South Korea, there is a 

need for efficient transportation means and faster mobility (Wikström et al. 2016). This forces 

the production and sale of more vehicles that, in turn, attract investment. In this situation, the 

automotive industry gains from the cycle. So (2020) explains that the manufacturers of several 

car brands like General Motors, Honda, Toyota and Ford experienced increased sales growth in 

Asia when countries in the West were hit by an economic crisis. He further explains that during 

2019, the Gross Domestic Product (GDP) of South Korea amounted to 1642.38 billion US 

dollars. He derived this information from the World Bank data and projections from trading 

economies. South Korea’s GDP, therefore, contributes a portion of 1.37 per cent of the World’s 

total GDP. Currently, the automobile industry of South Korea is ranked the fifth largest producer 

for passenger cars across the world. Relating the this industry with the country’s GDP, the 

automobile industry contributes 13 percent of the manufacturing plant, 12 percent of value-

added, and as an industrial area, it generates over 12 percent employment opportunities for South 

Koreans (Barreira et al. 2013). The automotive sector is the central industry that influences all 



 DISSERTATION  46 
 

sectors under industrialization from materials like glass, steel, nonferrous metals to advertising, 

transportation, construction and financial services. For the past decades, the automotive industry 

of South Korea has gone through a revolution and grown, making it one of the economic growth 

indicators of the country.  

For the case of India, the figures released in the dismissal April automobile sales by the 

Society of Indian Automobile Manufacturers (SIAM) indicate that whole vehicle production 

ecosystem in India could be a core growth indicator for the nation. SIAM explains that the 

production of vehicles within India reduced tremendously from the FY 2019 to FY 2020. They 

claim that the problem might have been caused by the presence of a global pandemic (Corona 

virus). This reduction in car production affected all the car sales segments (NetApplications.com, 

2016). In the past, sales of cars in India have a weak relationship with the GDP growth due to 

factors like consumer sentiment.  

Figure 5: India’s economy and car sales  

 

Source: Goyal, 2019 



 DISSERTATION  47 
 

The figure above indicates the relationship between GDP growth and car sales growth and two-

wheel sales growth. It shows that recently in 2019, the trend of car sales with GDP growth is 

directly proportional to the GDP growth compared to the two-wheel sales growth.  

Figure 6: Relationship between Indian economy and commercial car sales 

 

Source: Goyal, 2019 

Figure 2 above portrays the impact of GDP growth on commercial vehicle sales in India 

from FY 2006 to FY 2019. Goyal (2019) discusses that the sales of commercial vehicles have a 

strong relationship with GDP growth. Goyal (2019) explains that although the annual car sales 

growth has been showing a positive trend, many car producing companies have been seen 

fighting for space on roads and this greatly reduces the growth of car sale.  

 



 DISSERTATION  48 
 

2.6 Theoretical Framework 

2.6.1 The customer Journey using the AIDA Model 

AIDA model is developed to illustrate the influencing impact of the demographic transition of 

the consumers. The full form of AIDA is Attention (A), Interest (I), Desire (D) and Action (A). 

Continuing with the very initial stage, the product needs to be attracted by the customer, as stated 

by Lee & Hoffman (2015). Then the customer should focus on the interest of the products 

(Wijnhoven, & Plant, 2017). This makes the products attractive for the consumers. The awareness 

for the products gets evolved among the consumers. In this way, the consumers attain to focus on 

the consequences following the consumption of the products. Considerable awareness for the 

product results in the improvement of the purchasing amount for that particular product. 

Customer awareness develops when customers devote considerable time to collect product 

information (Pashootanizadeh & Khalilian, 2018). The simple approach of the model makes it 

more popular as compared to the five-stage model. The model explicitly illustrates in which way 

a company can improvise its product features in accordance with the growing consumer demand. 

The hierarchy-effects are the principle strategy to make both supplier and consumer awareness of 

the current market condition (Gnann et al., 2015). Apart from considering the direct involvement 

of the consumers, the theory also considers the purchasing activity of the consumers. The model 

emphasizes on the practical experience of the consumers. The complete framework of the model 

gets accomplished only after analyzing the reaction of the consumers. 

 

Attention Interest Desire Action



 DISSERTATION  49 
 

Figure 7: AIDA model 

Source: (as created by the author) 

 

Figure 2 exhibits a linear structure of four different stages of the AIDA model. The concept of 

the customer journey is conventional, combining with a number of alternative theories and 

marketing activities. It narrows down the broad concept of the buying decision towards a definite 

objective (Hassan et al., 2015). The model follows a customer-driven approach, and on this 

account, the theory is often called a customer journey model, as mentioned by Puzach & Hass 

(2014).  

The AIDA model is the purchasing model that reveals the entire decision process of a 

customer in the form of information search, decisions and processing. The AIDA model contains 

four processes that include attention, interest, desire, action respectively that customers follow 

from the point of searching for a product to the point of deciding to buy it. Google Trends 

usually represent the stage of interest that real customers have for a product or service. Research 

from different scholars shows that search activities exhibit the intention of buying and also 

predict the behaviour of a consumer and also the sales of the products (Wijnhoven & Plant, 

2017). However, some researches indicate that Google Trends are not suitable for showing the 

desire for a product since they do not portray negative or positive sentiments (Liu, 2012). Basing 

on the AIDA model, interest strongly indicates the intention of a customer to buy more than 

attention (Wijnhoven & Plant, 2017). On the other hand, desire has a stronger intention to buy 

compared to interest. Thus, there is always a stronger correlation between desire and sales 

compared to that of interest and attention. Furthermore, minimum time lags are expected to 

prevail between desire peak moments and the sales peak moments than for attention and interest 



 DISSERTATION  50 
 

(Wijnhoven & Plant, 2017). Nevertheless, Google search has a strong influence on decisions 

made after the search results. This is very common when conditions are so indecisive at the stage 

of searching (Epstein & Robertson, 2015). Since high-priced purchases need more research, 

Google Trends tool can be a suitable forecasting tool for the volume of new sales. (Yang et al. 

2015).  

 Nonetheless, the impulse buying behaviour of the consumer is highlighted by Hawkins 

Stern in 1962. The impulsive behaviour of the consumer makes him concerned about the product 

features when he comes to know about the product for the first time. Stern categorizes impulsive 

buying nature into four angles. First one is an impulsive purchase, the second one is reminding 

impulsive buying nature, the third one is suggestive impulsive buying nature, and the final one is 

planned impulsive buying nature. Impulsive buying nature provides ample opportunities for 

marketers to understand consumer behaviour. 

 

Figure 8: Customer Journey 

Source: (as created by the author) 

Active  
Evaluat

ion

Purchas
e 

moment

Experien
ce after 

purchase

Intial 
conside
ration



 DISSERTATION  51 
 

Brand consideration is the main stage of this Customer Journey model. A single product can be 

purchased by a different number of manufacturing companies. Initially, there are a potential 

number of customers for around a hundred numbers of car companies. The customers need to 

phase out car brands as per the requirements. This will narrow down the choices of car 

companies, and this will further develop a constructive comparative analysis among the sorted 

number of car brands. Henceforth, the evaluation phase is much dependent on the customer and 

choice of preferences. According to Rosenbaum et al. (2017), the evaluation process is driven by 

online research work. This active search process is developed through a number of selected 

brands. The next stage defines the moment of the purchase. This focuses on the attitude of the 

customers while they initiate their purchasing process. This process encourages customers to 

accelerate the purchasing activity. This is an integral part of the post-purchase activity. In this 

way, a strong brand loyalty gets formed among the customers. This reflects on the proactive 

recommendation for the purchasing decision. The proactive suggestion is to account for the 

important aspect of brand loyalty for the customers. It highlights the importance of brand loyalty 

in case of easing the development process of the customers.  The cumulative actions of this 

research process help the customers to examine the presence of alternative brands (Diţoiu & 

Cǎruntu, 2014). In the view of modification, a considerable amount of information is required to 

justify the changes in the model. Lemon and Verhoef (2016) identify that the importance of the 

customer-driven purchasing activity is the key significant part of generating public reviews about 

the available commodities. Meanwhile, it supports the product evaluation strategy. It outlines the 

availability of substitute goods for the desired product. From the customer’s point of view, it is 

important to focus on the nature of the substitute goods and their purpose of services. The 

circular motion highlights the interdependency of the products. 



 DISSERTATION  52 
 

Nymand-Andersen and Pantelidis (2018) conducted research on Google econometrics on new 

car sales forecasting and the requirements of big data quality. In their words, the authors said that 

“Big Data” is turning into a useful issue in the daily lives as it’s a digital source of information 

that helps to cover a large population (Nymand-Andersen & Pantelidis, 2018). The information 

sources may help in the formulation of new datasets leading to proper insights into the use of 

digital behaviour and the present actions of economic agents. This, in turn, may help the 

promotion of early indicators of both financial and economic trend in activities. The authors 

based their paper on investigating the importance of Google search data in nowcasting of car 

sales within the euro area (Nymand-Andersen & Pantelidis, 2018). This acted as a major 

macroeconomic indicator and took the quality requirements for employing the new data sources 

as a tool for policymaking and sound decisions. The results revealed that although Google data 

may be capable of being predictive for now-casting car sales within the euro area, there was a 

need to make improvements in the sources of data requires so as to go further in experimental 

statistics (Nymand-Andersen & Pantelidis, 2018). The authors concluded that quality 

requirements could be achieved and the resulting improvements in knowledge and theory of 

interpreting big data help to re-shape people’s thoughts in both socio-economic issues and 

behaviour.  

Wijnhoven and Plant (2017) explored the importance of sentiment analysis and Google Trends 

data while forecasting car sales. They explain that previous authors had conducted on how the 

two techniques may be used in forecasting sales but did not base on goods like cars and so they 

covered that gap by forecasting sales for cars. In their study, the researchers employed over 

500,000 social media posts for the eleven car models in the market of Dutch. The analysis was 

conducted with models of linear regression. In addition, the study relates the outcomes from the 



 DISSERTATION  53 
 

analysis with Google Trends predictive power (Wijnhoven & Plant, 2017). The findings show 

that sentiments of social media have less predictive power on car sales as well as Google Trends 

data and social mention volume significantly have more predictive power while suited in the 

prediction model. They conclude by providing a prediction model having time lags developed 

with a decision tree regress