The impact of COO and other attributes on consumers’ smart phone purchase

(1)

The Impact of COO and Other Attributes on Consumers’ Smart Phone Purchase

Thesis of Lei Wang (10681965)

University: University of Amsterdam Business School

Study program: Executive Program Management Studies – Marketing

Academic year: 2014-2015

Thesis supervisor: Dr. E. Peelen

Version: Final

(2)

Statement of Originality

This document is written by Student Lei Wang who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Acknowledgements

This master thesis is written as a closure for my study of Executive Program Management Studies - Marketing track at University of Amsterdam Business School. The past two years has been a great journey where I enjoyed being a student again after left university more than six years ago in 2009. I had the privilege to attend lectures that were taught by outstanding professors/lecturers who not only brought academic knowledge but also their rich industry experience into classes. It was also a great pleasure to meet and work with fellow students who are from different backgrounds but share the same ambition and motivation.

However, being a full time employee during the whole period, there were challenges as well. Especially it was not easy to find a balance between work, study and life at times. Thanks to all the support I have received, there was not a single moment that I had doubt in finishing the study. Therefore, I would like to thank a number of people who have helped me greatly during this journey.

First of all, I would like to thank my thesis supervisor Dr. Ed Peelen, who is not only a good teacher but also a very helpful and supportive supervisor. I really appreciate all the advice, guidance and patience he offered during the thesis period.

Secondly, I would like to thank my previous manager Chris Keizer, my current manager Stefan van der Maesen and my department director Stephen May from TNT. They have provided me the possibility of combining a high level study and my work. Without their flexibility this would have not been possible.

Furthermore, I would like to thank my partner Sebastiaan Limbach who has been extremely supportive and understanding in the past two years. I also cannot be more grateful to my mother, who has offered her help unconditionally in various ways.

Last but not least, I would like to thank all friends, colleagues and other respondents who participated in the interviews and/or questionnaires to make this research possible.

With best regards, Lei Wang

(4)

Abstract

The global smart phone market has witnessed extraordinary growth in recent years. It is not clear however, if consumers behave differently comparing to when they purchase other products. This research is interested in finding the attributes that influence consumers’ smart phone purchase decision and investigating if Country of Origin impacts consumers’ purchase behavior for smart phones as for some other products.

There are two parts of this research: a pilot study that is qualitative and exploratory, aiming at finding out consumers’ general opinion about what they consider as most important attributes or cues while purchasing a smart phone; and the main study is consisted of a questionnaire which is quantitative. A conceptual model is developed where Screen Size, Camera quality, Brand and Country of manufacture serve as independent variables; Price is dependent on the four independent variables while it also impacts Smart phone purchase decision; the dependent variable is Smart phone purchase decision; last but not least, Consumers’ view of smart phone being utilitarian or hedonic is proposed to be a moderator between CoM and purchase decision. The technique of conjoint analysis is used to design the questionnaire and the analyses are done using multiple regression analysis in both SPSS and R.

Brand is found to be the most important factor when consumers are evaluating which smart phone to purchase, while Screen Size, Camera resolution and Country of Manufacture are found not statistically significant when being tested jointly. CoM is however found statistically significant when being tested alone. Unlike the outcomes of in some of the literatures about more traditional products, the proposed moderator of consumers’ view of a smart phone being utilitarian or hedonic does not show statistically significant impact.

This study selects smart phone as the research subject as there are very few researches focusing on consumers’ purchase behavior when it comes to smart phones, although it is a relatively new product and has become one of the most commonly owned electronics and being purchased by more and more consumers. This paper brings the high tech product into attention and by combining both qualitative and quantitative methods, provides some insight in what consumers find important when it comes to purchase a smart phone and contributes in both academic and managerial perspective.

(6)

2

1. Introduction

“Smart phone vendors shipped a total of 375.2 million units during the fourth quarter of 2014 (4Q14), resulting in 28.2 percent growth when compared to the 292.7 million units shipped in the same period in 2013 and 11.9 percent sequential growth above the 335.3 million units shipped in the third quarter of 2014.” -- eWeek, 02 February 2015

The global smart phone market has witnessed extraordinary growth in recent years, with shipments rising by 40 percent in 2013 to exceed the 1 billion unit threshold and $266 billion in value.1 In 2014, 375.2 million units of smart phones were shipped in the fourth quarter alone. This extraordinary figure also indicates a much globalized production locations of smart phones. Take Apple’s iPhone as an example, although being designed in the USA, the components to make the phones are sourced from all over the world before gathering in China, where the final products are being assembled and shipped for sale in the global markets.2

The growth of the high amount of production and shipping is driven by the high demand from consumers. The number of smart phone users worldwide will surpass 2 billion in 2016, according to new figures from eMarketer—after nearly getting there in 2015. In 2015, there will be over 1.91 billion smart phone users across the globe, a figure that will increase another 12.6% to near 2.16 billion in 20163.

Advances in information and communication technologies are constantly changing the way people use and experience technology, which is ever more pervasive in consumers’ life (Petruzzellis, 2010). The presence of these technologies, such as smart phone has impacted consumers purchase behavior as well. For example, according to a research done from January to February 2014, commissioned by Google, which explores the growing importance of online platforms in the consumer journey, 32% of consumers in UK make a monthly purchase on their smart phones, while the numbers are 8% in France, 15% in Germany and 19% in Sweden. More than half (57%) of smart phone owners have also used their device to search for information while out shopping. With most shoppers using their smart phones to

1_{http://www.lenovo.com/transactions/pdf/CCS-Insight-Smart phone-Market-Analysis-Full-Report-07-2014.pdf} 2

http://comparecamp.com/how-where-iphone-is-made-comparison-of-apples-manufacturing-process/

(7)

3

compare prices (63%), look for discount vouchers (42%) and looking for product information or other options on a different retailer’s website (34%)4

.

While it is clear that smart phones are helping smart consumers to make better choices, many people may ask: what make consumers choose their smart phones? Also, given the much globalized production of smart phones, does the Country of Origin have an impact on consumers’ purchase decision?

(8)

4

2. Literature review

2.1 Country of origin

In today’s globalized market with fierce competition, many multinational companies have moved or outsourced their production to other locations, usually in low-cost developing countries. Also, there are increasing number of companies worldwide adopting a strategy of using foreign brand names – that is, “spelling or pronouncing a brand name in a foreign language” (Leclerc, Schmitt, and Dube 1994, cited by Melnyk et al, 2012). In both situations, while the branding implies specific country of origin (COO) in an effort to build or enhance perceptions of certain product attributes, the “made in” label (actual COO) reveals that the product was manufactured in a developing country.

In the past decades, the effect of a product’s COO on consumers’ perception, evaluations and intentions has been one of the most widely studied phenomena in the international business, marketing and consumer behaviour literatures (Lerman, 2011). Researchers have defined or conceptualized COO effects differently. Samiee (1994) considers COO effect as any influence or bias that consumers may hold resulting from the COO of a product. Nagashima (1970) defines it as the picture, the reputation and the stereotype that businessmen and consumers attach to products of a specific country. In the study of Lerman (2011), COO effect was referred to as the extent to which the place of manufacture influences consumer evaluations and related decisions. Roth and Romeo (1992) states a country’s image arises from a series of dimensions that qualify a nation in terms of its product profile. Such dimensions include innovative approach (superior, cutting-edge technology); design (style, elegance, balance); prestige (exclusiveness, status of the national brands); and workmanship (reliability, durability, quality of national manufacturers). Kotler and Gertner (2002) claim that the strong associations between the country image and product quality in relation to product/brand evaluations necessitate the identification of how global consumers perceive the redefined concept of COO. They also make the distinction within COO, between country of design (CoD), and as the country of manufacture/assembly (CoM/A). In this study, the definition of COO is referred to CoM/A rather than CoD.

(9)

5

According to Miller (2011), approximately one-quarter of consumers make purchase decisions on the basis of COO information (cited by Melnyk et al, 2012). Although manufacturing in less developed countries can assist corporations in enhancing their cost advantages, corporations also face the risk of potential loss due to negative COO effect. There is enough evidence showing where a product is made, can have an impact on consumer product evaluation and purchase decision (Bilkey and Nes, 1982; Okechuku, 1994). Many studies have also concluded that consumers typically view products made in developing countries less favourably (Cordell, 1992).

2.2 Hedonic versus Utilitarian Products

Melnyk et al (2012) define “hedonic products” as products that are associated with sensory, experiential, and enjoyment-related attributes and are consumed and evaluated primarily on the basis of benefits related to enjoyment, taste, aesthetics, and symbolic meaning. They define “utilitarian products” as products associated with functional, practical, and tangible attributes that are consumed and evaluated primarily on the basis of functional, instrumental, and practical benefits.

According to several prior research (Chittui et al, 2008; Gill 2008, cited by Melnyk et al, 2012)) consumers activate different sets of goals depending on the type of product they are considering; they tend to have functionality-related goals when they consume utilitarian products and pleasure-related goals when they consume hedonic products. These goal related differences imply that consumers’ product evaluation criteria and the information-processing procedures they use to evaluate products may differ systematically between hedonic and utilitarian products.

Bilkey and Nes (1982) illustrate that the cognitive approach sees a product as a cluster of cues: product-intrinsic cues which include cues such as taste, design, material and performance; and product-extrinsic cues, such as price, brand name, store reputation, warranty and COO. According to Godey at al. (2012), consumers generally rely more on intrinsic attributes when forming their opinions. However, in certain circumstances, consumers prefer extrinsic attributes, finding them more credible and reliable than their own assessment (Srinivasan et al., 2004). The use of extrinsic attributes can also relate to

(10)

6

situational factors, especially when status or self-image affects the purchase of a product (Piron, 2000; Quester and Smart, 1998).

Melnyk et al (2012) suggest that consumers employ different processing strategies for hedonic versus utilitarian products. In turn, the processing strategy will determine how much attention they pay to individual product attributes (e.g., price, shape) versus cues (e.g., COO) and the type of product information they consider diagnostic in their product purchases. The result of Melnyk et al (2012) implies asymmetric effects for hedonic versus utilitarian products. They find that consumers process information about utilitarian products by using an attribute-based cognitive elaboration strategy with which they are less likely to pay attention to cues in general (including COO cues). On the opposite side, hedonic products are processed more holistically, using heuristic and cues; therefore, the COO information is considered more important.

(11)

7

3. Research questions and research design

3.1 Research questions

This study is interested in finding out what product attributes are considered most important by consumers while purchasing a smart phone. Furthermore, whether Country of Origin has an impact on consumers’ decision making is also studied.

Research questions:

How do consumers decide which smart phone to purchase?

- Which attributes of a smart phone are considered most important by consumers? - In what way do consumers COO evaluations influence smart phone purchase

decisions (or intentions)?

- Does consumer’s view of smart phone being utilitarian or hedonic moderate the relationship between COO and consumers’ smart phone purchase decision?

3.2 Research design

There are two parts of this research: a pilot study that is qualitative and exploratory, aiming at finding out consumers’ general opinion about what they consider as most important attributes or cues while purchasing a smart phone; and the main study is consisted of a questionnaire which is quantitative.

3.2.1 Pilot study: Interview

Pilot study: a qualitative research was conducted. Interviews were held to gather real consumer insights of what made them purchased their current smart phone (if applicable). The idea is to generate the most considered product attributes that impact real consumers’ smart phone purchase decision. The technique used to approach the sample is Convenience Sampling. 9 smart phone users were randomly approached at the head office of TNT Express where the researcher works. 3 of the participants were female and 6 were male, their age range from 29 to 42. Due to the international working environment, multiple nationalities were involved and the interviews were conducted in English, which is the working language of all participants.

The interview consisted of three parts: 1) at the beginning, a small opening line where participants were ensured confidentiality, 2) an introduction was given to all participants

(12)

8

explaining the reason and content of the interview, 3) last but not least, 8 questions were asked during the interview:

1. How long have you been using smart phones? 2. How many smart phones have you had so far? 3. Which smart phone are you currently using? 4. How long have you had your current smart phone?

5. Are you happy with your current smart phone? Why (not)? 6. What made you purchase your current smart phone?

7. What will be the criteria for you to consider to buy your next smart phone? 8. Do you see a smart phone as a functional product or more as a hedonic product? Question number 6 and 7 are the most important ones which are intended to collect the attributes of smart phones that are important for consumers, but the other questions help participants to warm up and get into the topic.

Below is a summary of all the answers to question 6 and 7 collected and the number of times that they have been mentioned by the participants:

Table 1

Total mentioned times

Screen size 8 Price 7 Camera quality 6 Brand 5 Memory size 4 Operating speed 4 Ease of use 3

Size of the phone 3

Operating system 3

Screen resolution 3

Battery life 3

Design of the phone 2

Network speed 1

Colour 1

Noticeably, Screen Size was mentioned most, 8 out of the 9 participants clearly indicated that the screen size of the smart phone was one of the decisive attributes to them. Not surprisingly,

(13)

9

Price came strong as the second most mentioned reason for participants to consider while purchasing their smart phones. However, unlike the other independent attributes, participants mentioned that the price they were willing to pay was largely dependent on how “good” the smart phone was. So Price itself is dependent on the other attributes. Two thirds also mentioned that they needed a phone that is able to take good pictures. These attributes were followed by “Brand”, being mentioned 5 times. Other attributes such as Memory size, Operating Speed etc. were also mentioned, but by less than half of the participants.

3.2.2 Phase 2: Quantitative Questionnaire

Phase 2: the most mentioned attributes that are collected from the phase 1 qualitative interviews, together with COO are then used to conduct a questionnaire to test whether and to what degree they impact consumers’ purchase behaviour when it comes to purchasing a smart phone. The technique of Conjoint Analysis is used to conduct the questionnaire.

Data is collected by means of an online survey. Invitation is sent via both emails and social media means such as Facebook and LinkedIn. The sample consisted of a mix group of nationalities, both Dutch and non-Dutch. The questionnaire is composed in English since the research is aimed at a more international consumer base. The English language used in the questionnaire is rather basic and participants are expected to be able to understand the questionnaire without difficulties. At least 100 completed questionnaires are expected while the goal is to have 150 completed.

(14)

10

4. Proposed Model and Hypothesized Relationships

4.1 Conceptual model

As mentioned in the interview part, Screen Size, Price, Camera quality and Brand are ranked as the most important attributes when it comes to purchasing a smart phone for the participants. Participants also state that the price they are willing to pay depends on the level of the attributes the smart phone processes. They are willing to pay more for a phone with bigger screen, better camera comparing to a phone with smaller screen and less camera quality. So the conclusion can be drawn that price is not an independent variable, rather, it is dependent on the other attributes. Price is thus not included in the hypotheses, as it is designed to be a dependent variable that is decided by the other factors, as a result being a reflection of the smart phone’s screen size, camera quality, brand and country of manufacture. The inclusion of Price will make the model over complicated and creates too many variables. So for the main study, Price is considered as redundant information and not included in the model calculation itself. As it is still interesting to see if and how much Price itself impacts the purchase decision, it is being analyzed separately in later section.

From the literature review, the statement of “consumers process information about utilitarian products by using an attribute-based cognitive elaboration strategy with which they are less likely to pay attention to cues in general (including COO cues). On the opposite side, hedonic products are processed more holistically, using heuristic and cues; therefore, the COO information is considered more important” is considered. The assumption here is that: if a consumer perceived a smart phone as a utilitarian product, then the COO information will not influence his/her purchase decision; if the consumer considers a smart phone as a hedonic product, then the COO information will impact his/her purchase decision. And this is a potential moderator for the relation between CoM and purchase decision.

Based on the outcome of previous mentioned phase 1 qualitative interviews and literature review, the conceptual model is formed, where Screen Size, Camera quality, Brand and Country of manufacture serve as independent variables; Price is dependent on the four independent variables while it also impacts Smart phone purchase decision; the dependent variable is Smart phone purchase decision; last but not least, Consumers’ view of smart phone being utilitarian or hedonic is a moderator between CoM and purchase decision.

(15)

11 Figure 1 Conceptual Model

4.2 Proposed Hypotheses

The first tree hypotheses are the results from phase one interviews, where real life consumers mention that they find Screen size, camera quality and Brand important when they make decision about which smart phone to purchase. Thus Screen size, Camera quality and Brand are assumed to have an impact on consumers’ purchase decision:

H1. Screen size of the smart phone impacts consumers’ purchase decision. H2. Camera quality of the smart phone impacts consumers’ purchase decision. H3. Brand of the smart phone impacts consumers’ purchase decision.

The hypotheses related to Country of origin are derived from literature reviews where a connection is identified for some products. Furthermore, a possible moderating effect is proposed based on the findings of Melnyk et al (2012): that consumers process information

Country of Manufacture

Smart phone purchase decision

Consumer’s view of smart phone being utilitarian or hedonic Brand Price Camera quality H1 Screen size H2 H5 H3 H4

(16)

12

about utilitarian products by using an attribute-based cognitive elaboration strategy with which they are less likely to pay attention to cues in general (including COO cues). On the opposite side, hedonic products are processed more holistically, using heuristic and cues; therefore, the COO information is considered more important. So when consumers perceive smart phones as utilitarian products, they are less likely to pay attention to COO information; while when consumers perceive smart phones as hedonic products, the information about the smart phones are processed more holistically and the COO information is considered more important and is thus more likely to have an impact on consumers’ purchase decision.

H4a. Country of origin of the smart phone impact consumers’ purchase decision when consumers consider a smart phone as a hedonic product.

H4b. Country of origin of the smart phone does not impact consumers’ purchase decision when consumers consider a smart phone as a utilitarian product.

H5. Consumer’s view of a smart phone being utilitarian or hedonic moderates the relationship of COO and consumers smart phone purchase decision.

(17)

13

5. Conjoint analysis and measures

5.1 Conjoint analysis

Conjoint Analysis is a quantitative technique used in product management and marketing analysis for assessing how people value different features that characterize an individual product or service. Instead of directly asking the consumers’ opinion about different features of a product, conjoint analysis asks respondents to evaluate different profiles of a product. Conjoint analysis is a main-effects analysis meant to estimate the joint effect of a set of independent variables measuring the attributes of a product or service on a dependent variable. In another word, it measures the preferences of consumers (Bodog and Gyula, 2012). Conjoint analysis originated in mathematical psychology and was developed by marketing professor Paul Green at the University of Pennsylvania. In 1988 professor Srinivasan developed a linear programming (LINMAP) procedure for ranking ordered data (Bodog and Gyula, 2012). In the late 1970s, Green and Srinivasan (1978) summarized two basic methods when it comes to data collection procedures in conjoint analysis:

 The two-factor-at-a-time procedure, and

 The full-profile approach

The two-factor-at-a-time procedure, which is also known as the “trade-off procedure”, considers attributes on a two-at-a-time basis. The respondent is asked to rank the various combinations of each pair of factor levels from most preferred to least preferred. The limitation of this method includes: by decomposing the overall set of factors to two-at-a-time combinations, some of the realism is sacrificed; by breaking down the factors, say six factors, each at four levels, the respondent could be asked to fill out 15 tables, each with 16 cells. The total number of required judgments is very large; the procedure appears to be more suited to verbal descriptions of factor combinations, rather than pictorial or other kinds of iconic representations.

The full-profile approach, also referred to as the concept evaluation task, utilizes the complete set of factors. Among its advantages is its ability to measure overall preference judgments directly using behaviourally oriented constructs such as intentions to buy, likelihood of trial, chances of switching to a new brand, and so on. Such measures are particularly useful in the context of introducing new products/services. The elicitation of such constructs from respondents requires that each option being described on all of the attributes,

(18)

14

that is, the full-profile approach (Green and Srinivasan, 1990). Green and Srinivasan (1990) also state that the full-profile method of conjoint analysis works very well when there are only a few (say, six or fewer) attributes. The major limitation of this approach is the possibility of information overload and the resulting temptation on the part of the respondent to simplify the experimental task by ignoring variations in the less important factors or by simplifying the factor levels themselves (Green and Srinivasan, 1978).

Later, more approaches were introduced, such as Self-explication approaches, Hybrid methods and Adaptive conjoint analysis. Green and Srinivasan (1990) compare all the advantages and limitations and conclude that for situation where the number of attributes can be kept down to (say) six or fewer factors, the full-profile conjoint analysis is most recommended. Given there are 4 attributes identified for smart phones in this study, the full-profile approach will be used.

5.2 Attribute levels and orthogonal design

As mentioned before, Screen Size, Camera Quality, Brand and Country of Manufacture are the attributes of smart phones that this research is interested in testing. As for Screen Size, the most common 4-inch (iPhone 5s) and 5-inch (Samsung Galaxy S4) are selected. Camera resolution is taken as an indication for the camera quality, and the common offering in the market are 8 and 16 megapixels, which are chosen as the two different attribute levels.

Based on real smart phone market situation, the most popular brands Apple and Samsung are selected (25% and 47% of market share in the Netherlands according to Telecompaper). ‘Huawei’, which is a Chinese brand, is also becoming recognized by some consumers in the Netherlands thanks to its active efforts in expanding internationally (according to Telecompaper 5% marketshare in the Netherlands). It is also chosen due to its Chinese background. As these brands are from the U.S, Korea and China respectively, these three countries are used as the possible country of manufacture (CoM) locations. The table below shows the attributes and their levels:

Table 2 Smart phone attributes and their levels

Number of options

Brand Apple Samsung Huawei 3

(19)

15 Camera resolution 8 megapixels 16 megapixels 2

CoM U.S Korea China 3

Total 36

Using Orthogonal Design function in SPSS, the 36 combinations are generated as below: Table 3 Initial Orthogonal Design result from SPSS

Card ID Brand

Screen

size Camera COO 1 Apple 4 inch 16 mp China 2 Huawei 4 inch 8 mp U.S 3 Apple 5 inch 16 mp China 4 Apple 5 inch 16 mp Korea 5 Samsung 5 inch 16 mp U.S 6 Huawei 5 inch 16 mp Korea 7 Samsung 4 inch 8 mp Korea 8 Samsung 4 inch 16 mp Korea 9 Apple 4 inch 8 mp U.S 10 Huawei 4 inch 16 mp Korea 11 Apple 5 inch 8 mp China 12 Samsung 4 inch 8 mp China 13 Apple 5 inch 8 mp Korea 14 Samsung 4 inch 16 mp China 15 Apple 5 inch 8 mp U.S 16 Huawei 4 inch 16 mp U.S 17 Huawei 5 inch 16 mp China 18 Apple 4 inch 16 mp Korea 19 Apple 4 inch 8 mp Korea 20 Huawei 5 inch 8 mp China 21 Samsung 4 inch 16 mp U.S 22 Huawei 5 inch 8 mp U.S 23 Samsung 5 inch 16 mp Korea 24 Apple 4 inch 8 mp China 25 Huawei 5 inch 8 mp Korea 26 Huawei 5 inch 16 mp U.S 27 Samsung 5 inch 8 mp U.S 28 Huawei 4 inch 8 mp China 29 Apple 4 inch 16 mp U.S 30 Samsung 5 inch 8 mp China 31 Apple 5 inch 16 mp U.S 32 Huawei 4 inch 8 mp Korea 33 Samsung 5 inch 8 mp Korea 34 Samsung 4 inch 8 mp U.S 35 Huawei 4 inch 16 mp China 36 Samsung 5 inch 16 mp China

(20)

16

As Apple and Samsung are from U.S and Korea respectively, it is possible that some of their smart phones are made in their home countries, although, most of the ones are currently manufactured/assembled in China. As for Huawei, which is a Chinese brand, it makes less sense to label the phone as Made in U.S or Korea. So, the combinations were processed with the most reasonable COO information to all the brands, with Apple linked to both U.S and China; Samsung linked to both Korea and China and Huawei linked to China only. Another consideration is the length of the questionnaire. To avoid the questionnaire being too long and meaningless, options that are not representing real market offers are excluded. All the options presented should be perceived realistic and similar to the real market offers, so the choices participants make can represent real consumer choices. A Huawei phone made in Korea or the US, or a Samsung smart phone made in the US, or an iPhone made in Korea, is not likely to be offered in the current market, and are thus excluded. As a result, 20 combinations remain. However, this modification reduces the randomness of the orthogonal design, and creates correlation between CoM and Brand, which will be analyzed in later session.

Table 4 Orthogonal Design excluding unrealistic CoM and Brands matches

Card ID Brand Screen size

Camera resolution

Country of Manufacture 1 Apple 4 inch 16 mp China 2 Apple 5 inch 16 mp China 3 Samsung 4 inch 8 mp Korea 4 Samsung 4 inch 16 mp Korea 5 Apple 4 inch 8 mp U.S 6 Apple 5 inch 8 mp China 7 Samsung 4 inch 8 mp China 8 Samsung 4 inch 16 mp China 9 Apple 5 inch 8 mp U.S 10 Huawei 5 inch 16 mp China 11 Huawei 5 inch 8 mp China 12 Samsung 5 inch 16 mp Korea 13 Apple 4 inch 8 mp China 14 Huawei 4 inch 8 mp China 15 Apple 4 inch 16 mp U.S 16 Samsung 5 inch 8 mp China 17 Apple 5 inch 16 mp U.S 18 Samsung 5 inch 8 mp Korea 19 Huawei 4 inch 16 mp China 20 Samsung 5 inch 16 mp China

(21)

17

To make the options more “real”, prices are added. The next step is to set the most reasonable price for these combinations. As Price is impacted by the attribute levels, below logics are used:

1. Smartphone with both bigger screen and higher camera resolution should be matched to highest price level;

2. Smartphone with both smaller screen and lower camera resolution should be matched to lowest price level;

3. Smartphone with mixed quality, eg. bigger screen but lower camera resolution or smaller screen but higher camera resolution should be matched to the price in the middle.

iPhone 5c is taken as a reference for price. It features 4-inch screen and 8 megapixel camera, Mediamarket offers it at 325 euro. This is the starting point. For the research design, the big screen ones with high camera resolution are priced at 600 euro. Smart phones with mixed quality are priced in between as 450 euro. Furthermore, as the price is set based on iPhone 5c, based on market reality, iPhone is more expensive than Samsung, and Samsung is more expensive than Huawei, so the price level is adjusted: when it is Samsung, the price is reduced by 50 euro, when it is Huawei, the price is reduced by 100 euro. In this way, Price is not independent, but rather dependent on smart phone’s screen size, camera quality and brand. Table 5 Processed Orthogonal Design with Prices added

Card ID Brand Screen size

Camera resolution

Country of

Manufacture Price 1 Apple 4 inch 16 mp China 450 2 Apple 5 inch 16 mp China 600 3 Samsung 4 inch 8 mp Korea 275 4 Samsung 4 inch 16 mp Korea 400

5 Apple 4 inch 8 mp U.S 325

6 Apple 5 inch 8 mp China 450 7 Samsung 4 inch 8 mp China 275 8 Samsung 4 inch 16 mp China 400

9 Apple 5 inch 8 mp U.S 450

10 Huawei 5 inch 16 mp China 500 11 Huawei 5 inch 8 mp China 350 12 Samsung 5 inch 16 mp Korea 550 13 Apple 4 inch 8 mp China 325 14 Huawei 4 inch 8 mp China 225 15 Apple 4 inch 16 mp U.S 450 16 Samsung 5 inch 8 mp China 400 17 Apple 5 inch 16 mp U.S 600 18 Samsung 5 inch 8 mp Korea 400 19 Huawei 4 inch 16 mp China 350

(22)

18

20 Samsung 5 inch 16 mp China 550

As can be seen, the Prices attached to the options are not randomly assigned, rather, they are specifically calculated for each option to make them look realistic and close to the real offers in the market.

These 20 options are presented to participants in the questionnaire. Respondents are positioned in a situation that they are searching for a smart phone to buy and these 20 combinations are the options provided in the market and they are asked to provide a rating on how likely they are going to purchase each option.

5.3 Measures

In the conjoint analysis, the independent variables of Brand, Screen Size, Camera Quality and Country of Manufacture are measured jointly in the same model using multiple regression analysis. Price, which is dependent on the four independent variables, is analysed separately using regression analysis. As the orthogonal design outcome has been modified and the randomness of the design has been impacted, Pearson correlation analysis is conducted to test the correlation between all the variables. The dependent variable, purchase intention is a common effectiveness measure and often used to anticipate response behaviour. The method of asking participants to evaluate a product and then indicate an intention is prevalent throughout the literature (Li et al., 2002). Thus, an established seven-point scale was used to measure the likelihood that participants would purchase the evaluated product. The question “How likely are you going to purchase this smart phone?” was asked, with scale 1 being very unlikely and 7 being very likely. To measure how consumers view a smart phone being hedonic or functional, participants were asked to characterise smart phone as primarily a functional product or primarily an entertainment/enjoyable product, along a scale from 1 (primarily for functional use) to 7 (primarily for entertainment/enjoyable use) (Palazon and Delgado-Ballester, 2013).

(23)

19

6. Survey outcomes and analysis

6.1 Response analysis

156 people participated in the study and 144 of them filled in the questionnaire completely. Out of the total population, 66 were female (46%) and 78 were male (54%). Majority of the participants are between 26 and 35 years old, counting for 65% of the total number. This is followed by the age group 36-45, which is 15% of the population. Furthermore, there are 8% respondents younger than 26 and the rest are older than 45 years old. The distribution of age group can be found in Table 6.

Table 6 Respondents’ Age Distribution

Age Frequency Distribution

18-25 11 8% 26-35 94 65% 36-45 22 15% 46-55 10 7% 56- 7 5% Total 144 100%

Among all the respondents, 141 are already smart phone users, which make up 98% of the population. Same as the real market situation, most of the current smart phone users use phones from the brands Apple or Samsung, while the portion is different: with 54% of them being iPhone users and 21% Samsung users. Other popular brands such as HTC, Sony, LG and Huawei are also presented. The distribution can be found in Table 7:

Table 7 Respondents’ Smart Phone Brands Distribution

Brand Frequency Distribution

Apple 78 54% Samsung 30 21% HTC 10 7% Sony 7 5% LG 6 4% Huawei 3 2% Nokia 1 1% Other 9 6% Total 144 100%

(24)

20

Among the 144 respondents, 67 or 47% indicate that a smart phone is more a functional product to them by providing a score between 1-3; 42 or 29% of the participants prefer to categorize a smart phone as more a hedonic (enjoyable) product (score 5-7); 24% (35) choose a neutral score of 4. So it is fair to say that most consumers do not view smart phones as hedonic products (76%).

6.2 Multiple regression analysis in SPSS and R

As each participants provided purchase intention score (preference score) for all the 20 profiles, the data is reworked in such a way that all the attributes levels and all the preference scores are shown, as a result, 2880 rows of data were created (144*20=2880), instead of the original 144 rows.

Price is not included in the analysis together with the four independent variables, because it is not an independent attribute in the study. As mentioned, it is designed to be completely dependent on the level of the four attributes. So price levels are expected to be full reflections of the attribute levels, as a result being redundant information. It is presented to participants for the sake of completeness of an offer. The impact of price is calculated separately in a later section.

As there are several attributes with different levels, a multiple regression analysis is conducted using statistical software SPSS. A check of frequencies is done to ensure there is no error in the data. Attributes Brand and Country of Manufacture are first dummy coded, then the dependent variable Purchase Intention is added, after which the independent variables are entered (Screen Size, Camera resolution, Apple, Samsung, Huawei, China, Korea and U.S.A). “Excluded cases listwise” is selected so only cases with no missing data in any variable will be calculated.

To test the full model, the default method is selected. As the ANOVA tests suggests (Table 9), the model is significant with F(6,2879)=22.648. However, less than 5% of the variance can be explained by the model (Table 8).

Table 8 SPSS Full Model Summary

Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,213a ,045 ,043 1,774

(25)

21

a. Predictors: (Constant), USA, Camera, Screen, Korea, Huawei, Samsung

Table 9 SPSS Full Model ANOVA

ANOVAa

Model Sum of Squares df Mean Square F Sig. 1 Regression 427,891 6 71,315 22,648 ,000b

Residual 9046,553 2873 3,149 Total 9474,444 2879

a. Dependent Variable: PurInt

b. Predictors: (Constant), USA, Camera, Screen, Korea, Huawei, Samsung

At the same time, high correlations are found between Countries of manufacture and Brands, which is expected. As per design, USA and China are the manufacturing countries for Apple, Korea and China are for Samsung, and China is for Huawei. So correlation between USA and Apple is 0.612, between Korea and Samsung is 0.612 and between China and Huawei is 0.408, all being statistically significant. This could be part of the reasons why R2 is rather small. The result of Pearson correlation analysis can be found in Table 10.

Table 10 SPSS Full Model Correlations

Correlations

PurInt Apple Samsung Huawei China Korea USA Screen Camera Pearson Correlation PurInt 1,000 ,178 -,038 -,171 -,096 -,001 ,119 -,022 ,021 Apple ,178 1,000 -,667 -,408 -,167 -,408 ,612 ,000 ,000 Samsung -,038 -,667 1,000 -,408 -,167 ,612 -,408 ,000 ,000 Huawei -,171 -,408 -,408 1,000 ,408 -,250 -,250 ,000 ,000 China -,096 -,167 -,167 ,408 1,000 -,612 -,612 ,000 ,000 Korea -,001 -,408 ,612 -,250 -,612 1,000 -,250 ,000 ,000 USA ,119 ,612 -,408 -,250 -,612 -,250 1,000 ,000 ,000 Screen -,022 ,000 ,000 ,000 ,000 ,000 ,000 1,000 ,000 Camera ,021 ,000 ,000 ,000 ,000 ,000 ,000 ,000 1,000 Sig. (1-tailed) PurInt . ,000 ,021 ,000 ,000 ,469 ,000 ,121 ,129 Apple ,000 . ,000 ,000 ,000 ,000 ,000 ,500 ,500 Samsung ,021 ,000 . ,000 ,000 ,000 ,000 ,500 ,500 Huawei ,000 ,000 ,000 . ,000 ,000 ,000 ,500 ,500 China ,000 ,000 ,000 ,000 . ,000 ,000 ,500 ,500 Korea ,469 ,000 ,000 ,000 ,000 . ,000 ,500 ,500 USA ,000 ,000 ,000 ,000 ,000 ,000 . ,500 ,500

(26)

22

Screen ,121 ,500 ,500 ,500 ,500 ,500 ,500 . ,500 Camera ,129 ,500 ,500 ,500 ,500 ,500 ,500 ,500 .

A further step is taken where “Stepwise method” is selected, while everything else remains the same to run the model, so only the statistically significant variables are shown in the Model Summary, which shows that Brand Apple and Brand Huawei are the only statistically significant variables in this model (Table 11&12). Also, consistent with the previous finding, less than 5% of the variables can be explained by the model.

Table 11 SPSS Stepwise Model Summary

Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 ,178a ,032 ,031 1,785 2 ,208b ,043 ,043 1,775

a. Predictors: (Constant), Apple

b. Predictors: (Constant), Apple, Huawei

Table 12 SPSS Stepwise ANOVA

ANOVAa

Model Sum of Squares df Mean Square F Sig. 1 Regression 299,515 1 299,515 93,952 ,000b Residual 9174,929 2878 3,188 Total 9474,444 2879 2 Regression 410,384 2 205,192 65,129 ,000c Residual 9064,061 2877 3,151 Total 9474,444 2879 a. Dependent Variable: PurInt

b. Predictors: (Constant), Apple

c. Predictors: (Constant), Apple, Huawei

These results imply that Screen size, Camera quality and Country of Manufacture variables do not have significant impact on consumers’ smartphone purchase intention in the model. Brands, however, do significantly influence the purchase intention. However, as according to the hypothesis, CoM can either be significant or insignificant depends on the consumer’s view of smart phone being utilitarian or hedonic, this variable will be further examined. Table 13 is a summary of the testing results of hypothesis 1-3.

(27)

23 Table 13 Hypotheses Testing Results H1-H3

Hypothesis Prediction Result

H1 Screen size of the smart phone impacts consumers’ purchase

decision.

Rejected.

H2 Camera quality of the smart phone impacts consumers’

purchase decision.

Rejected.

H3 Brand of the smart phone impacts consumers’ purchase

decision.

Confirmed.

A further step is taken to calculate the importance of each attribute. Another multiple regression analysis is conducted. The preference score (Y) as the dependent variable and the ten attribute levels are independent variables (X). b0 is the intercept, and b1,…b10 are parameters of regression model (part-worth utilities of attribute levels). “s” is the number of respondent (s=1,2,…,144) The mathematical expression of the model is as follows:

Y s= b0s + b1s(Brand Apple) + b2s(Brand Samsung) + b3s(Brand Huawei) + b4s(Screen 4 inch) + b5s(Screen 5 inch) + b6s(Camera 8mp) + b7s(Camera 16mp) + b8s(China) + b9s (Korea) + b10s (USA)

By using this formula, it is assumed that all the independent variables are also independent from each other. However, as the orthogonal design was reworked to reduce unrealistic combinations, Brand and CoM are highly correlated with each other. Due to the limitation of statistic software available, this is not taken into account in this model. This is also one of the limitations of this research.

Calculation is done using the Conjoint Package in Statistic software R, in order to make this possible, data needs to be translated into formats that R can work with. Four separate text files are created, including “profile”, “attribute levels”, “preference score matrix” and last but not least, “preference score in one column”.

The first calculation uses the function caPartUtilities(). It calculates and returns matrix of individual part-worth utilities (parameters of regression) for all artificial variables (with intercept on first place) for every respondent. Taking the first respondent as an example (s=1),

Y1 = 3.167 + 0.417 (Apple) -0.583 (Samsung) +0.167 (Huawei) -1.7 (4 inch) +1.7 (5 inch) + 0.5 (8 mp) -0.5 (16 mp) + 0.417 (China) + 0.417 (Korea) – 0.833 (USA)

(28)

24

Table 14 R Multiple Regression Results for Respondent 1&2

Respondent intercept Apple Samsung Huawei 4 inch 5 inch 8 mp 16

mp China Korea USA

1 3.167 0.417 -0.583 0.167 -1.7 1.7 0.5 -0.5 0.417 0.417 -0.833

2 4.333 0.333 -0.417 0.083 -0.25 0.25 -0.15 0.15 0.083 -0.167 0.083

The result shows that for this respondent, the most attractive smart phone option is a phone from Apple with screen size 5 inch and an 8 megapixel camera. It does not really matter for this respondent whether a phone is made in China or Korea, but rather not made in the USA. The second respondent shows some different preference: he/she prefers an Apple iPhone with 5 inch screen and a 16 megapixel camera, and prefers a phone to be made in China or the USA.

The complete list of individual part-worth utilities for all variables for every respondent can be found in Appendix B.

Next, the parameters for all 144 respondents are calculated using function Conjoint (). And the result is the following:

Residuals:

Min 1Q Median 3Q Max -2,6490 -1,4979 -0,4934 1,5049 4,5604 Coefficients:

Estimate Std. Error t value Pr(>|t|) (Intercept) 3,07407 0,04269 72,013 <2e-16 *** factor(x$Brand)1 0,50000 0,06037 8,282 <2e-16 *** factor(x$Brand)2 -0,02083 0,06037 -0,345 0,730 factor(x$Screen)1 0,03958 0,03307 1,197 0,231 factor(x$Camera)1 -0,03819 0,03307 -1,155 0,248 factor(x$CoM)1 -0,07755 0,04929 -1,573 0,116 factor(x$CoM)2 0,08044 0,07794 1,032 0,302 --- Signif. codes: 0 ‘***’ 0,001 ‘**’ 0,01 ‘*’ 0,05 ‘.’ 0,1 ‘ ’ 1 Residual standard error: 1,774 on 2873 degrees of freedom Multiple R-squared: 0,04516, Adjusted R-squared: 0,04317 F-statistic: 22,65 on 6 and 2873 DF, p-value: < 2,2e-16

[1] "Part worths (utilities) of levels (model parameters for whole sample) :" levnms utls 1 intercept 3,0741 2 Apple 0,5 3 Samsung -0,0208 4 Huawei -0,4792 5 4inch 0,0396 6 5inch -0,0396 7 8mp -0,0382 8 16mp 0,0382

(29)

25

9 China -0,0775 10 Korea 0,0804 11 USA -0,0029

[1] "Average importance of factors (attributes):" [1] 42,40 22,99 16,88 17,72

[1] Sum of average importance: 99,99

The model is statistically significant, F(6,2873)=22.65, p < 0.001, R2=0.04516. So

although the model is significant, less than 5% of the variance is explained by the model, which is consistent with the finding using SPSS.

These results also show that on a total level, the most attractive options are Brand Apple, Screen Size 4 inch, Camera resolution 16 megapixels and Country of Manufacture Korea. The respective importance of attribute is also calculated: with Brand being the most important (42.4%), followed by Screen Size (22.99%), Country of Manufacture (17.72%) and Camera Resolution (16.88%). The relative importance of attributes is presented in Figure 2.

(30)

26

6.3 Impact of Price

Although not included in the initial analysis of the study, it is still interesting to see if and how much Price alone has influenced the purchase intention. A regression analysis is done with SPSS and the result shows that Price does have a statistically significant impact on purchase intention, with F(1,2879)=15.188, R2=0.005. This result (Table 15 & Table 16) suggests that although being significant, Price can only explain 0.5% of the purchase decision. One reason of the low R2 can be that Price itself is not independent in this study. It is designed to be a reflection of the combination of other attributes.

Table 15 SPSS Impact of Price Model Summary

a. Predictors: (Constant), Price

Table 16 SPSS Impact of Price ANOVA

ANOVAa

Residual 9424,707 2878 3,275 Total 9474,444 2879

a. Dependent Variable: PurInt b. Predictors: (Constant), Price

6.4 Testing the moderating effect

A moderator changes the relationship between the independent variable and the dependent variable, so if the relationship changes because of the introduction of consumer’s view of smart phone being utilitarian or hedonic, then the moderating effect can be proven, otherwise not. Figure 3 shows the hypothesis of the variables.

(31)

27 Figure 3 Proposed Moderating Effects

Firstly, the relationship between CoM and Purchase intention was tested using regression analysis in SPSS. Model was run using default method. The result shows that the model is

significant, F(2,2877)= 21.99, R2= 0.015. So although the model is significant, only 1.5% of

the variance can be explained (Table17 & Table 18) Table 17 Model Summary CoM and Purchase Intention

a. Predictors: (Constant), USA, Korea

Table 18 CoM and Purchase Intention ANOVA

ANOVAa

Residual 9331,767 2877 3,244 Total 9474,444 2879

a. Dependent Variable: PurInt b. Predictors: (Constant), USA, Korea

In the following step, the model was run again with the “stepwise” method selected to show the significant variables. The result shows that USA is statistically significant, with F(1,2878)=41.46, R2=0.014 (Table 17, Table 19 & Table 20).

Table 19 CoM Stepwise Model Summary

Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate Country of Manufacture Smart phone purchase decision Consumer’s view of smart phone being utilitarian or hedonic

(32)

28

1 ,119a ,014 ,014 1,801

a. Predictors: (Constant), USA

Table 20 CoM Stepwise ANOVA

ANOVAa

Residual 9339,894 2878 3,245 Total 9474,444 2879

a. Dependent Variable: PurInt b. Predictors: (Constant), USA

Table 21 CoM Stepwise Model Coefficients

Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. 95,0% Confidence

Interval for B Correlations

Collinearity Statistics B Std. Error Beta Lower Bound Upper Bound

Zero-order Partial Part Tolerance VIF 1 (Constant) 3,031 ,038 80,756 ,000 2,957 3,104

USA ,540 ,084 ,119 6,439 ,000 ,376 ,705 ,119 ,119 ,119 1,000 1,000 a. Dependent Variable: PurInt

It is obvious that the result is different from the previous finding, when being measured together with the other attributes, County of Manufacture was found statistically not significant. While being measured individually, CoM is statistically significant. Again this could be caused by the high correlation between CoM and Brand. In another word, CoM is dependent on Brand, when being measured together with Brand, mainly Brand is showing the impact and being statistical significance.

Last but not least, the moderating effect of Consumer’s view of smart phone being utilitarian or hedonic was tested. In order to perform the task, the script “Process” was installed in SPSS. As this is a simple moderation model, model number 1 was selected. Purchase intention is the dependent variable (Y), CoM is the independent variable (x) and Consumer’s view of smart phone being utilitarian or hedonic is the moderator (M) being tested. The full result can be seen in Appendix D.

(33)

29

While the whole model is significant, with F(3,2876)=14.6265, p<0.001, R2=0.02, the

moderation effect (interaction between Attitude& CoM) is not, with p=0.28. Table 22 Moderator Testing Result

Outcome: PurInt Model Summary

R R-sq MSE F df1 df2 p ,1278 ,0163 3,2405 14,6265 3,0000 2876,0000 ,0000 Model

coeff se t p LLCI ULCI constant 3,1389 ,0336 93,5015 ,0000 3,0731 3,2047 Attitude ,0489 ,0220 2,2214 ,0264 ,0057 ,0920 CoMNO ,2686 ,0433 6,1954 ,0000 ,1836 ,3536 int_1 ,0304 ,0281 1,0832 ,2788 -,0247 ,0856

So the moderating effect of consumer’s view of smart phone being utilitarian or hedonic is not proven. In table 23, a summary of the testing results of hypotheses 4 and 5 are provided.

Table 23 Hypotheses testing result H4-H5

Hypothesis Prediction Result

H4a Country of origin of the smart phone impact consumers’

purchase decision when consumers consider a smart phone as a hedonic product.

Rejected.

H4b Country of origin of the smart phone does not impact

consumers’ purchase decision when consumers consider a smart phone as a utilitarian product.

Rejected.

H5 Consumer’s view of a smart phone being utilitarian or

hedonic moderates the relationship of COO and consumers smart phone purchase decision.

(34)

30

7. Discussions, limitations, and future research

7.1 Conclusions and discussions

In this research the different attributes of smart phones and their impact on consumers’ purchase intention is studied. Both qualitative and quantitative methods are used to conduct the study. Firstly a qualitative face to face interview is conducted to gather consumers’ opinions of attributes that are important for them while considering purchasing a smart phone. From the interview, four attributes are identified: screen size, camera quality, brand and price. While the first three attributes are independent, price is dependent on the others. Country of manufacture is also of interest due to a statement from literature reviews. As per design, CoM is highly dependent on the brand of the smart phone. Based on the input from phase one study, a conceptual model is developed and the hypotheses are formed. Then a quantitative questionnaire is conducted and sent out to collect consumers’ preference scores for certain smart phone offers. The technique of conjoint analysis is used to conduct the questionnaire and to collect the responses.

In the questionnaire, after some basic demographic questions such as gender, age group, participants are asked to indicate on a scale whether a smart phone is more hedonic or more utilitarian for them, and then are asked to provide a score of how likely they are going to purchase a certain smart phone given the 20 options provided. The hypotheses are tested based on the results from the questionnaires.

SPSS is the main tool used for hypotheses testing. When the model is tested as a whole, excluding the proposed moderator, the model is significant as a whole with attribute Brand being statistically significant and others not. When CoM is tested alone as an independent variable together with dependent variable Purchase intention, the impact is significant, although with low percentage of variance explained. “USA” as a country of manufacture is statistically significant with 1.5% variances explained. The hypothesis of “Consumers view a smart phone being hedonic or utilitarian is a moderator to CoM and purchase intention” is rejected. The impact of Price is also analyzed separately from the other attributes. And the result is statistically significant with 0.5% of the variables explained. One of the explanations

(35)

31

can be that Price is calculated based on the other independent variables and the main impact is reflected by the independent variables.

Statistics program R is also used. Using the conjoint package of R, the part-worth utility scores are presented for each respondent. The total importance of each attribute is also calculated with Brand being the most important (42.4%), followed by Screen Size (22.99%), Country of Manufacture (17.72%) and Camera Resolution (16.88%). It is worth mentioning that less than 5% of the variances can be explained by this model. This can partially be explained by the fact that CoM is dependent on Brand.

These results raise some questions:

1. Although being statistically significant, the model consisted of Screen Size, Camera, Brand and CoM together can only explain less than 5% of consumers’ smart phone purchase decision, apart from the high correlation between CoM and Brand, what could be the reasons for such low level of variables being explained? Is it the model, is it the statistics methods used, or are there more factors to be considered? Will the inclusion of other attributes mentioned by participants in the interview such as Memory Size, Operating Speed, Ease of use, Size of the phone, Operating system, Battery life and even Color of the phone lead to a better model to predict consumers’ smart phone purchase decision? If so, how and to what degree will they impact consumers’ smart phone purchase?

2. When being tested separately, CoM shows a statistically significant impact on consumers’ smart phone purchase intention. 1.5% of the variances can be explained by this attribute. While the percentage is not high, it still indicates that CoM information influences some consumers’ purchase decision. Given the still increasing market size of smart phones, this may make a difference of millions of euros. So it is a question for the management, should smart phone companies consider and provide some alternatives in regard of production locations?

3. While in the interviews most consumers claim the importance of Screen Size and Camera resolution, these two factors are found not statistically significant in the quantitative research. Why is this the case? One explanation could be that even the

(36)

32

“worse” options provided in the questionnaire are already good enough for most consumers. For example, a phone with a camera of 8 megapixels resolution is considered very good already, and consumers do not really need a 16 megapixels camera on their phone. So under this circumstance, they put the focus on other attributes. This might be useful information for smart phone companies, not every attribute needs to be extremely outstanding for the phone to sell.

Another possibility is that many consumers do not have much idea what exactly the number means. They do not really know what the difference is between an 8 megapixels or 16 megapixels camera, or how big is a 4 inch screen or 5 inch screen. It may be better to show them several real models instead of some combinations on paper; or at least provide better explanations of what exactly the difference attribute levels mean.

7.2 Limitation and further research

There are several limitations and possible improvement of this study. First of all, a fairly small sample size (9) is taken for the interviews and only 4 product attributes are selected as the attributes for this research. Further research could extend the study by including the impact of more attributes or factors that impact consumers’ smart phone purchase behavior, such as the ones being mentioned by consumers in the interviews: Memory Size, Operating Speed, Ease of use, Size of the phone, Operating system, Battery life and Color of the phone. The presentation of the options can also be improved by including visual images of the different smart phones and better illustration of the meaning of all the different attribute levels should also be provided.

Secondly, due to the design of the study, price is not being part of the analysis together with the independent variables. Although the impact of Price is analyzed separately, it is less a reflection of how it really works. Price certainly plays a big role in consumers’ purchase decision. The impact of price is worth looked into more in depth in future researches.

Thirdly, even though some of the models that are tested are significant, the variances being explained is fairly small in the research (all below 5%), while this can be partially explained by the high correlation between some variables (CoM and Brand, Price and other attributes),

(37)

33

there are other factors that should be investigated, maybe from other angles other than product attributes, such as product life cycle, consumer profile etc.

Although the technique of conjoint analysis is used to conduct part of the study, the analyses are done mainly using multiple regression and simple regression analysis in SPSS and R. This is due to the lack of access to real conjoint analysis statistics program. By using professional conjoint analysis analyzing tools, more accurate and insightful results might be concluded.

Last but not least, some attribute levels are more preferred than the others, it does not mean the better the attribute level, the more preferred it is. In this study, it seems sometimes a less advanced level is more favored. It is interesting for further researches to find out what attributes fall into this category and to what degree are the levels optimal.

(38)

34

References

Ajzen, Icek and Martin Fishbein (1980), Understanding Attitude and Predicting Social Behavior. Englewood Cliffs, NJ: Prentice Hall.

Bilkey, W.J. and Nes, E. (1982), “Country-of-origin effects on product evaluations”, Journal of International Business Studies, Vol, 13 No.1, pp.41-58.

Bodog, S. and Gyula, F. (2012), “ Conjoint Analysis in Marketing Research”, Journal of Electrical and Electronics Engineering, Vol 5, No.1.

Chandon, Pierre, Vicki G. Morwitz, and Werner J. Reinartz (2005), “Do Intentions Really Predict Behavior? Self-Generated Validity Effects in Survey Research,” Journal of Marketing, 69 (April) 1-14.

Chitturi, Ravindra, Rajagopal Raghunathan, and Vijay Mahajan (2007), “Form Versus Function: How the Intensities of Specific Emotions Evoked in Functional Versus Hedonic Trade-Offs Medate Product Preferences,” Journal of Marketing Research, 44 (November), 702-714.

Chu Po-Young, Chang Chia-Chi, Chen Chia-Yi and Wang Tzu-Yun, (2010), “Countering negative country-of-origin effects”, European Journal of Marketing, Vol. 44 Iss7/8pp.1055-1076.

Cordell, V.V. (1992), “Effects of consumer preferences for foreign-sourced products”, Journal of International Business Studies, Vol.23 No.2, pp.251-69.

Gill, Tripat (2008), “Convergent Products: What Functionalities Add More Value to the Base?” Journal of Marketing, 72 (March), 46-62.

Godey, B. Pederzoli, D. Aiello, G. Donvito, R. Chan, P. Oh, H. Singh, R. Skorobogatykh, I. Tsuchiya, J. and Weitz, B. (2012), “Brand and country-of-origin effect on consumers’ decision to purchase luxury products”, Journal of Business Research, 65 (2012) 1461-1470. Green, Paul E. and Srinivasan, V. (1978), “Conjoint Analysis in Consumer Research: Issue and Outlook”, Journal of Consumer Research, Vol. 5, No.2 , pp.103-123.

Green, Paul E. and Srinivasan, V. (1990), “Conjoint Analysis in Marketing: New Developments with Implications for Research and Practice”, Journal of Marketing, Vol.54, No.4.

Kotler, P. Gertner, D. (2002) “Country as brand, product and, beyond: a place marketing and brand management perspective.” Journal of Brand Management, 9 (4): 249 -61.

(39)

35

Leclerc, France, Bernd H. Schmitt, and Laurette Dube (1994), “Foreign Branding and Its Effects on Product Perceptions and Attitudes,” Journal of Marketing Research, 31 (May), 263-70.

Li, H. Daugherty, T. and Biocca, F. (2002), “Impact of 3-D Advertising on Product Knowledge, Brand Attitude, and Purchase Intention: The Mediating Role of Presence”, Journal of Advertising, Vol. 31, No. 3, Advertising and the new media, pp 43-57.

Lim, L. K. Han, K.S. and Chan Y.F.B (2013), “Factors Affecting Smart phone purchase decision among Malaysian Generation Y”, International Journal of Asian Social Science, 2013, 3 (12): 2426-2440.

Miller, John W. (2011), “Country Labeling Sets Off EU Debate,” The Wall Street Journal, (June 6).

Nagashima, A. (1970), “A comparison of Japanese and US attitudes toward foreign products:, Journal of Marketing, Vol.34 (January), pp. 68-74.

Okechuku, C. (1994), “The importance of product country of origin: a conjoint analysis of the United States, Canada, Germany, and The Netherlands”, Journal of Marketing, Vol.28 No.4, pp.5-19.

Palazon, M. and Delgado-Ballester, E. (2013), “Hedonic or utilitarian premiums: does it matter?”, European Journal of Marketing, Vol. 47 No. 8, 2013 pp. 1256-1275.

Petruzzellis, L. 2010. “Mobile phone choice: technology versus marketing, the brand effect in the Italian market.” European Journal of Marketing, Vol.44, No 5, 2010

Piron, F. (2000), “Consumers’ perceptions of the country –of-origin effect on purchasing intention of inconspicuous products”, Journal of Consumer Marketing, 17(4):308-21. Quester, P. and Smart, J. (1998), “The influence of consumption situation and product involvement over consumers’ use of product attribute”, Journal of Consumer Marketing, 15 (3): 220-38.

Roth, MS. Romeo, GB. (1992) “Matching product category and country image perceptions: a framework for managing country of origin effects.” Journal of International Business Studies, 23 (3): 477-97.

Samiee, S. (1994), “Consumer evaluation of products in a global market”, Journal of International Business Studies, Vol.25 No.3, pp.579-604

The impact of COO and other attributes on consumers’ smart phone purchase