Brands extension to autonomous driving:
Identifying brand related success drivers
Liu Yang
Brands extension to autonomous driving:
Identifying brand related success drivers
Master Thesis
University of Groningen Faculty of Economics and Business
Marketing Intelligence
16th January 2017
First supervisor: dr. F. (Felix) Eggers
Second supervisor: prof. dr. T.H.A. (Tammo) Bijmolt
Liu Yang · Cederlaan 188 · 5616SC · Eindhoven · Netherlands
0
Abstract
It is estimated that there will be a significant growth for autonomous vehicle industry in the next two decades, and currently, consumer demand for a self-driving car is quite high among different countries. Consumers' adoption of an innovative product which can reflect their behavior and consumption decisions toward certain brands. From a theoretical perspective, this research is identifying what are the possible brand related factors will drive autonomous car successes. In a managerial view, this study provides insight for companies what brand factors will influence consumers' decision in evaluating a self-driving car, and which brand factors have the relatively strong effects compare with other brands drivers.
The literature reviews are given based on a summary of most frequently discussed successful brand extension factors from previous articles most from technology companies. These factors are brand familiarity, brand quality, extension attitude, image fit and category fit. Lastly, I also include consumer characteristic as moderators in consumers’ preference choice model.
The model is estimated using MNL models on an aggregate level with a maximum likelihood procedure. In total three models are performed; the basic model contains car attributes, the extending model include the brand related factors, and the moderating model contains the moderators of consumer characteristic. As the models are performed better from model 1 to model 2, it can be concluded that with adding covariates of brand factors, the model can better predict respondents' choices. However, the moderating model is not significantly improving compared with model 2. Therefore the moderating effect is not noticeable in this research.
0
Preface
1
Table of content
Table of content ... 1
Figures and Tables ... 2
Chapter 1: Introduction ... 3
Chapter 2: Literature Reviews ... 7
2.1 Brand familiarity ... 7
2.2 Quality of parent brand ... 8
2.3 Attitude toward the extension ... 9
2.4 Image fit and category fit ... 9
2.5 Moderating Effect of Consumer innovativeness ... 10
2.6 Moderating Effect of Perceived risk ... 11
Chapter 3: Conceptual model ... 13
Chapter 4: Methodology ... 14
4.1 Attributes and levels ... 14
4.2 Conjoint analysis ... 15 4.3 Experiment design ... 16 4.3 Choice Design ... 17 4.4 Moderating effect ... 17 Chapter 5: Results ... 18 5.1 Sample... 18 5.2 Conjoint analysis ... 19 5.3 Goodness of Fit ... 20 5.4 Models interpretation ... 21
5.3 Total brands effects Interpretation ... 22
5.3 Moderating effects ... 25
Chapter 6: Discussion ... 28
6.1 Conclusion ... 28
6.2 Hypotheses discussion ... 29
2
6.3 Managerial implication ... 31
Chapter 7: Limitations and further research ... 33
7.1 Limitations ... 33
7.2 Direction for further research ... 33
Reference list ... 35 Appendix A ... 38 Appendix B ... 40 Appendix C ... 42 Appendix D ... 43 Appendix E ... 45 Appendix F... 46
Figures and Tables
Figure 1: Conceptual model ... 13Figure 2 Vividness ... 19
Figure 3 Total Brand Effects ... 23
Table 1 Overview of literature review ... 12
Table 2 attribute level ... 14
Table 3 Brand related factors overview ... 19
Table 4 Models Overview ... 20
Table 5 Model Specification ... 21
Table 6 Model Importance ... 22
Table 7 Specific Brand Effects ... 23
Table 8 Partial Brand Effects ... 24
Table 9 Partial Brand Importance ... 25
Table 10 Model 3 Specification ... 25
3
Chapter 1: Introduction
An autonomous vehicle, also names as the driverless car, robotic car and self-driving vehicle (SDV). It is “A vehicle enabled with technology that has the capability of operating or driving
the vehicle without the active control or monitoring of a natural person" (Rish et al. 2008). It can
be classified to fully autonomous vehicles and partially vehicles. Fully autonomous vehicles can self-driving in virtually without driver intervention in all situations, whereas the partially autonomous vehicles can self-driving in virtually without a driver in selected situations, such as on highways, precise routes and auto free mouse valet parking (Rish et al. 2008). By applying autonomous vehicles, from the society perspective, it can significantly reduce road accidents and fatalities. Moreover, it can reduce the fuel consumption (Olivier et al. 2013) which contribute to our environment. From the driver perspective, autonomous cars can make commuting easier, and the driver will have more time doing other productive tasks and leisure in the car. (Rish et al. 2008)
As the tremendous opportunities could create in this innovation markets, it would be interesting to see what the critical drivers for customers adopted a driverless car are. It can be a quite general topic of which factors that influences customers to make the decision of selecting the autonomy cars finally. Such as consumer preference, technology maturity, regulation and mitigation of serious risks all play significant roles in adoption process (BCG, 2016). This thesis focuses on marketing related field; therefore, the researcher will only research on consumers’ preference that influences the autonomous vehicle adoption process. Without considering the technology, regulation, and mitigation factors, the main challenge here is to understand if consumers have the preference on certain brands in adopting an innovation product.
4
and safe vehicles. And there are also lots of them believe the technology company can provide them sophisticated software in SDV algorithm which is the critical determinant for safety issues. The rest consumers expect the new startup's company to provide the SDV as these new startups own an innovation business model which let consumers do not only benefit from the SDV product, but also receive a better service. Therefore, this research tries to reveal what are the significant determinants influence consumers to make the decision related to the brand preference.
5
categories. Therefore, in this paper, we will use preference analysis as a research methodology. I also investigate if consumers’ individual characters moderate the choice to different attribute’s in the conjoint analysis.
From the managerial view, the consumer demand for a self-driving car is quite high, approximately 55% of US drivers are interested in purchasing a partially autonomous car, and around 44% of them would like to have a fully autonomous car (Fortune, 2016). And the primary buying motivations from the customer are lower costs of fuel and insurance, also with the improved safety. According to BCG (2016) research, the partially autonomous vehicles will frequently appear on the road by 2017, there will be a significant growth for autonomous vehicle industry in the next two decades. Therefore, it will be beneficial for companies to see how branding will influence consumer decision toward a SDV or other similar innovation products. Whether the self-driving industry is following a traditional loop or it is belonging to a new brand extension model? Therefore, this paper aims to extend the current research by addressing brand related factors in the SDV introduction with respect successful brand extension. Particular, the researcher will focus on if brand preference will influence consumers' adoption to the autonomous vehicles. Moreover, under which types of companies' brand will have more advantage to launch a self-driving car successfully.
The objective of this research is to find out whether brand factors will influence consumers' perception to adopt a breakthrough new product. More specifically, what could be the strong drivers for brand extension in an innovation industry? So the research questions are:
How important is the brand name when evaluating autonomous vehicles?
Which attributes (brand extension factors, autonomous vehicle type and price) are critical in this adoption process also which levels of these attributes consumer prefer?
6
7
Chapter 2: Literature Reviews
From the problem statement above, the aim of this research is to find if brand factors will influence consumers’ adoption on an autonomous vehicle. So this literature review is to give an overview of potential drivers for successful brand extension among different industries. As there are few studies that specific in analyzing brand extension on innovation product. Based on theory, people should not view innovation product as an isolation sector, and innovation can be a bundle of interrelated technology (Roger, 2003). Therefore, the adoption of one innovation products may depend on the adoption of other related technology products. Accordingly, I summarize articles which are most from technology companies and published in recent ten years. And I selected the most popular drivers. Moreover, I select some factors are not frequently mentioned by these research, but may still relevant for autonomous industry. Therefore, I combined both drivers with high frequency appear in recent research but also some drivers that may have influence in a particular autonomous industry in my further analysis.
In Table 1, I list several important factors that illustrate if these factors are significant or not and explain these articles’ data collection details, and under which industries. Firstly, these factors are chosen because they are widely accepted by most authors. Secondly, in specific brand extension cases on innovation industry, some factors may have an effect in determining a successful brand extension are including in this table as well. In Appendix A I list some factors also studied by theses research but not include in my research, due to these factors are not frequently analyzed by recent research and irrelevant for an innovation industry.
Lastly, I also contain the moderate effect regarding consumer characteristic: innovativeness and perceived risk. Due to in an innovation cases, consumer characteristics have the high possibility in moderating consumers' preference in the choice set of brand related factors, SDV type and price.
2.1 Brand familiarity
8
extension than less familiar names (Martines & Pina, 2010; Milberg et al., 2010). This is because on the one hand, more familiar with the brand will stimulate consumers' ability to identify the relationships between the original brand and new products (Wanke, Bless, & Schwarz, 1998). Another explanation is because of the mere exposure effects, more familiar toward a brand, the brand name is more accessible to consumers, as consumers may use the parent brand familiarity as a heuristic to guide their evaluations (Hoyer & Brown, 1990). For autonomous car industry, consumers probably already develop preferences in certain brands as they have the prior using experiences with these brands so that they will generate more favorable attitudes toward a specific brand extension. For example, a consumer will generate a preferable attitude towards a familiar brand's autonomous car compare with the new, unfamiliar brand. This is because if this autonomous car launches by the company that the consumers are currently using their products or have previous using experiences, they may develop more knowledge in this brand and then have the ability to recognize the relationship between the brands. On the other hand, in an innovation industry, consumers may perceive high uncertainty toward a breakthrough technology car. Therefore, brand familiarity can greatly decrease perceive of risk hence to be a crucial determinant when evaluating a self-driving car.
Therefore, the first hypothesis is
H1A: Brand familiarity has a positive effect on the evaluation towards the extension.
2.2 Quality of parent brand
9
extension then benefits with this relationship. On the other hand, if this self-driving car is associated with some brands which consumers do not perceive with a good quality, and then the extension is less adopted by consumers.
Therefore, the hypothesis is formulated:
H1B: If perceived the parent brand as a good quality brand is associated with higher preference
of consumers’ adopting toward the self-driving car.
2.3 Attitude toward the extension
The more satisfy with the extension, the more satisfy customers feel about the brand (Martines & Pina, 2010; Albrecht et al., 2013). Therefore, by strengthening the attitude to the extension, consumers will feel less risky and develop more favorable evaluation to the extension (Alexander & Colagate, 2005). Currently, only two articles find this driver significant, however, according to the BCG (2015) report, there is a large part of customers show negative attitude to use an autonomous car under some brands due to quality problem. Therefore, the extension may not be successful if they have a negative attitude in certain brands that is launching a SDV. Therefore, the hypothesis is made:
H1C: The better attitude to the brand extension, the high evaluation to adopt the self-driving car.
2.4 Image fit and category fit
10
features or attributes similarity (Broniarczyk & Alba, 1994). For example, Ford and Mercedes-Benz belong to traditional car manufacturer; the consumer will feel fitter if they perceive autonomous cars is a classic car. The majority articles support that the fit between the parent brand and extend products can be the critical driver for a successful brand extension. (Table 1) In general, this article firstly investigates if product fit with the brand is still an important driver for autonomous car brand extension success. Specifically, I am interested in whether different fit will drive different effects in SDV brand extension. Therefore, based on image fit, if Apple or Google brand a driverless car, then consumer perceive both the product and parent brand is innovativeness. Therefore, consumers are more readily accept the extension. On the other hand, based on the category fit, if Ford or Mercedes-Benz name a driverless car, consumers then perceive both products and parent brand is category fit. Therefore, consumers are more readily accept the extension.
H1D: Image fit will positively influence the consumers’ choice to adopt the SDV, particular, if
consumers perceive the image fit, then Technology Company has higher evaluation of adoption H1E: Category fit will positively influence consumers’ choice to adopt SDV, particularly, if
consumers perceive the category fit, then automaker companies have a higher evaluation of adoption.
2.5 Moderating Effect of Consumer innovativeness
11
Therefore, it is interesting to see if the autonomous car is preferable by innovativeness consumer when branding by some ambitious start-ups, like an unknown brand name (Robocab).
The Hypothesis is made:
H1F: The consumers’ innovativeness moderates the SDV type, price and brand related factors
toward the brand extension preference.
2.6 Moderating Effect of Perceived risk
Perceived risk is categorized into two factors, uncertainty about the consequence of making a mistake and uncertainty about the outcome (Gronhaug & Stone, 1995; Mitchell, 1999). Brands extend to a new category and also increase consumers’ perception of risks. According to a case study of Public Perceptions of Self-Driving Cars (2014) the result finds that safety is the most attractive elements in evaluating a self-driving car, which means consumers, actually perceive high uncertainty toward this innovation car, and this uncertainty will matter their life. Similarities, due to consumers recognize the risk may vary per individual. For example, if consumers do not perceive to choose a SDV as a big decision to make, they do not have a strong preference in which brand in providing the SDV. Their final decision may not base on the brand related factors.
Therefore,
H1G: Consumer perceived risk moderates the SDV type, price and brand related factors toward
12 Table 1 Overview of literature review
Quality of the parent brand Brand familiarity Attitude to extension Image fit Category fit Perceive d risk Consumer innovativene ss Data Industry
Albrecht et al.(2013) √ √ √ √ √ 492 respondents Non/Luxury brand He &Li (2010) √ 200 participates in China High-Tech brand Hem, Chernatony
&Iversen(2003)
√ √ √ for service industry
701 respondents in Norway FMCG, Durables, Services Jin & Zou(2013) × × 320 participates In Shanghai Web-brands Martinez & Pina(2010) √ √ √ √ 699 respondents in Spain FMCG, Service
Milberg et al.(2010) √ √ √ 278 participates of business students
High involvement products Shen, Bei & Chu(2011) √ √ 160 respondents of University
students
13
Chapter 3: Conceptual model
14
Chapter 4: Methodology
The methodology based on the conceptual model will separate into two parts; the first part is using the conjoint analysis to prove if brand related factors will influence consumers’ adoption to SDV when corresponding with other attributes, and how important the brand influences the SDV adoption process. The second part is using a multiple-item measure to define which drivers play a role when evaluating a brand extension to an autonomous car.
4.1 Attributes and levels
Before start conjoint analysis, the attributes and levels should be defined. Due to the respondents should choose their preferred options regarding adopt a SDV in an assumption scenario. This scenario is to imagine in a certain situation; you need to choose an SDV which is provided by different brands. These brand names are differentiated with Apple & Google represent Technology Company, Ford & Mercedes-Benz represent traditional automakers, and Tesla is a brand between a technology company and mainstream automakers. Robocab stands for a new startups company which is unfamiliar for all respondents. The rest brands are chosen based on the well-established brands; this is because the brand extension is more attractive in a well-known brand instead of an unknown brand. The price range is given as a reference for the measure if brand switching will influence the price sensitive in Table 2. The SDV type is based on the recent BCG (2015), and the types can be classified into two levels. Therefore, the SDV adoption attributes are SDV type (half automated and fully automated), price per km (€ 10, € 12, € 14,€16,€18,and €20), brand (Apple, Google, Mercedes Benz, Ford, Robocab)
Table 2 attribute level
Attribute Attribute level
Brand Tesla Apple Google Ford & Mercedes-Benz Robocab Price Range €10 €12 €14 €16 €18 €20
SDV type Half automated
15 4.2 Conjoint analysis
Conjoint analysis is an approach to measure and understands how individual respondent to a product or services. It is based on the principle that consumers evaluate the value of a product or services by combining the different attributes' sum value toward this product or services. The utility is the unique preference among different individual and serves as a fundamental concept for measuring value in the conjoint analysis (Hair et al., 2006). In conjoint analysis, the importance of each attribute can be identified, as well as the utility of specific levels within the attributes can be researched. Therefore, for an innovation product, even if it is not on the market yet, we can still use conjoint analysis to measure consumers' preference correspond to consumers' wishes using different utilities per level per attribute. Furthermore, when asking the consumer if they would adopt an SDV, they will develop tons of factors that are important for their final decision. The scope of this research is only invested if brand factors will influence the SDV adoption. Therefore, they conjoint analysis can limit the consumers' response to certain factors, accordingly, to check what factors that the consumers find are important. According to the recent research (BCG, 2015), there are two important determinants in consumer preference concept that consumers concern for accept an SDV. By adding the different brand name as a new attribute, I assume that consumers will choose an option that maximizes their utility.
The mathematic foundation for this study is based on the random utility theory (RUT). RUT indicates that a participant utility U for individual product or product attributes are unobservable latent constructs, consisting of a systematic component V and a random component ε that captures non-systematic effects that are not accounted for (Eggers & Eggers, 2010)
U = V + ε
The formula above explains that the total utility for adopting a SDV is the sum of all individual choice attributes.
V links SDV attributes (X) to preference estimates (β), where ε is assumed to be individual and distributed as extreme value (Eggers & Eggers). Therefore, three utility models are implemented in the conjoint analysis. The benchmark model also named the traditional model includes the autonomy attributes, and the model can be found in Equation 1. To extend the transitional model with the brand extension success drivers (Equation 2), I also add the brand effects as residuals (γ) in the traditional formula (Eggers, Eggers, & Kraus 2014). Next, I include the interaction effect (ղ) between consumers’ characteristics with brand specific covariates, price and autonomous types (Equation 3).
16
Equation 1: Utility benchmark = β0+ β1 SDV type * type + β2 price *price + β3 brand *brand
Equation 2: Utility brand residual = β0 + β1 SDV type * type + β2 price *price + β3 brand *brand +γFam Famibrnad
+ γQuaQuaibrnad + γAttAttibrand + γImaImaibrnad +γCatCatibrand
Equation 3: Utility add moderating effects = β0 + β1 SDV type * type + β2 price *price + β3 brand *brand +γFam
Famibrnad + γQuaQuaibrnad + γAttAttibrand + γImaImaibrnad +γCatCatibrand+ղccha, famccha*Famibrand+ղccha, Quaccha*Quaibrand+ղccha,Attccha*Attibrnad+ղccha, Imaccha*Imaibrand+ղccha, Catccha*Catibrand
+ղccha,priceccha*priceiprice+ղccha, typeccha*typeitype
Next, the preferences are estimated based on multinomial logit model, the probability p will be chosen can calculated.
P(𝑦 = 𝑗|𝑥𝑖, 𝛽𝑖) =
𝑒𝑥𝑖𝑗𝛽𝑖𝑗
∑𝐽 𝑒𝑥𝑖ℎ𝛽𝑖
ℎ=1 4.3 Experiment design
For the first part of the survey, a pick any procedure is implemented to measure evaluation the brand. Therefore the results of respondents are coded as binary variables. Customers will choose one of the familiar brand, and a good quality brand among all provided brands. This is due to a 0/1 coded responses works well in a choice model (Holbrook & Moore, 1984). The advantage of this method is that by using pick-any procedure-based on the brand can better lend itself to the information on both brands and consumers in the preference than multidimensionally scaled (Holbrook & Moore, 1984).
Secondly, to ensure every respondent can vividness put themselves in making a decision to choose a self-driving car situation. The following text is added to the choice sets: “Picture yourself in the year 2025. Imagine that you are going on holiday to a destination that you have never visited before. Instead of renting a manually-driven rental car you are also considering the newly available self-driving cars. These autonomous, self-self-driving cars do not have to be rented for the entire time of your holiday but can be requested via an app whenever needed. You do not have to pick up the car because the car can drive to you and can park itself when you do not need it anymore. You only need to pay for every hour that you use the car. Fuel costs, car parking tickets, insurance and taxes are all included in the hourly rate.”
17
Lastly, the pick any procedure is used again to measure respondents' perception of autonomous brands with different brand residuals. The items are choosing based on a multiple-item measure which can be developed in a questionnaire using variables, and they have been summarized in the previous literature review that can best indicate respondents' state. Normally, most authors define one success factor based on previous studies (Aaker & Keller, 1990; Bottomley & Doyle 1996; Klink & Smith, 2001). Due to the length of the survey, in appendix A, I summarized previous authors’ items in one table and only pick one factor in survey questions.
4.3 Choice Design
The options for respondents are based on choices sets; therefore, a choice-based conjoint analysis technique is applied. Unlike the traditional conjoint analysis, respondents need to value one design per question (Hair et al., 2009)
The choice sets consist of three attributes with five levels, six levels, and two levels. In total, the choice design consists of 11 choice sets. Due to dual –response design, there are 22 choice sets in the data. This is because there are two choices per set. For the no-choice option, the dual-response design is implemented. The advantage of this design is that consumers have to choose one the alternatives and then indicate whether there are other more preferred options, so the test results are more realistic and robust.
For the choice analysis, the data will be merged into one file which contains both the choice data and survey items. Next, the brand-specific covariates will be created. It can be accomplished when using IF-Statement. Therefore, the brand quality, brand familiarity, brand extension preference, image fit and category fit covariates are created as binary data.
4.4 Moderating effect
Furthermore, the conjoint analysis cannot directly measure the moderate effect regarding consumer innovativeness and perceived risk, as these two factors remain constant when consumers are making the decision. Therefore, the last part of survey collects the data that address the moderate effect in the conceptual model. Here I use the rating scales to measure consumers’ individual characteristics by using the multiple items measurements (In Appendix B). This rating scale can best capture the multifaceted constructs (Eggers, Eggers, & Kraus 2014). For example, to measure if consumers are innovative, I ask respondents, to what extent you agreed/disagreed with the following statements (-3 = "strongly disagree," and 3 = "Strongly agree")
18
Chapter 5: Results
In this chapter, the results will be discussed. The first part of this section is the sample overview. Next, three models are estimated together with the part of the goodness of fit. I calculate adjust R2, Chi-square and hit rate to determine the model fit, and I conclude if adding attributes are significant improving the model. The following section is the conjoint analysis, which includes interpretation of the basic model estimation and extends model estimation. The third part of the results is the interpretation of moderating effect.
The data is collected online. Participates fill in the questionnaire through an online survey. In total, there are three students collect the data within the same survey. 339 respondents complete this survey (N=339), the sample size is enough for further analysis.
5.1 Sample
There are 51.9% female respondents and 48.1% male respondents who are almost equally divided. The average respondents are 26 years old with a standard deviation of 7 years old. The oldest respondent is 73 years old, and youngest respondents are 18 years old. The average time for respondents completing the survey is 9.4 minutes with a standard deviation of 5 minutes. Approximately 49% respondents are from Netherlands, and 8% respondents are from Colombia, and 7% respondents are from China. The rest respondents are from other countries of the world. In Table 3, it shows that 84.7% respondents are familiar with Tesla Motors, and 79.4% respondents think Tesla Motors offers a high-quality product. 90% respondents are familiar with Apple, and 87% of them find Apple has a good quality. 95% respondents know Google and 75.8% believe Google offers good quality. 94% respondents know Mercedes-Benz and 92% of them find it has a good quality. 91.7% respondents familiar with Ford, but 43% respondents consider Ford has a good quality. Only 1.8% participated know Robocab, and 1.5% find it has a good quality. And only one respondent's don't know any of these brands nor think these brands can offer a good quality product. According to Table 3, most respondents show a positive attitude to Tesla Motors which account for 86.4%, and next comes to the Mercedes-Benz which is 79.6%. Google ranks the third place with the proportion of 66.4%. Finally, Apple, Ford, and Rocab which are relatively less preferred by respondents if they want to extend to a SDV car. Less than 1% people indicate that they don't prefer all brands for SDV extension.
19
lower category fit. From this summary, I cannot conclude that the technology companies like Apple & Google received higher image fit than mainstream automakers like Mercedes-Benz &Ford. On the contrary, neither can prove that the mainstream automakers receive higher category fit than technology companies.
Table 3 Brand related factors overview
Tesla Motors Apple Google Mercedes-Benz Ford Robocab None Familiarity 84.70% 90.00% 95.00% 94.10% 91.70% 1.80% 0.03%
Quality 79.40% 87.00% 75.80% 92.90% 43.10% 1.50% 0.03%
Extension Attitude 86.40% 51.90% 66.40% 79.60% 38.90% 13.90% 0.09%
Image fit 86.10% 62.80% 76.70% 74.30% 44.80% 14.70% 5.00%
Category fit 80.20% 28.00% 61.70% 41.60% 23.90% 13.60% 5.00%
Next, the respondents are asked how vivid they can image the situation when using a SDV, and the survey is designed to a scale ranging from 1 to 6. According to Figure 2, most of the respondents can well imagine this situation as the most rating are 4 to 6. Therefore, the results will not be in influenced by the imagine scenarios which may cause respondents not providing a real preference as they are uncertain about the environment when making the decision.
Figure 2 Vividness
5.2 Conjoint analysis
The models are estimated using the MNL models on an aggregate level with a maximum likelihood procedure. In total I investigate three models (Table 4), Model1 is the basic model only include the car attributes. Model 2 I include brand related factors as an extending model to model 1. In model 3 I add in moderating effect in model 2.
Vividness
20 Table 4 Models Overview
LL BIC(LL) Chi-square p Npar df p-value R²adj Hit Rate
Model1 -5529,08 11128,08 5328.7318 <0.00 12 327 5,6e-2084 32.37% 64.77% Model2 -5257,30 10613,65 5332.292 <0.00 17 322 2,8e-1973 35.63% 67.24% Model3 -5204,75 10648,36 5977.4086 <0.00 41 298 1,7e-1969 35.98% 67.91%
5.3 Goodness of Fit
To measure if these three models fit, I calculate the Pseudo-adjust R2, Chi-square and the Hit Rate for these models. As Latent Gold uses another method to calculate the adjust R2. Therefore, the following formula is used:
𝑅2adj = 1 −LL(β∗) − npar LL(0)
With the n= number of respondents, c=choice sets and m=number of alternatives. According to Table 3, the three log-likelihood of the aggregate models are LL (1) 5,529,0845, LL (2) =-5,527.3044 and LL (3) =-5,204.7461. Therefore, we get the adjust R21=32.37%, R22=35.63%, and R23=35.98%. The adjust R2 is getting better from model 1 to model 3. Due to there may be other factors that influence the choice model, the value around 0.2 to 0.4 are considered acceptable. As all parameters in these three models are significantly different from null model, it can be concluded the aggregate model fit well with the observed choices. And also can prove that with the new factors add in, the model can better predicate the choices. But the difference between Model 2 and Model 3 is not significant, which also indicate that the moderation effect is not noticeable.
LL(0) = n × c × ln (1
𝑚) = 339 × 22 × ln ( 1
3) = −8,193.4504 Next, the Chi-square is calculated as follow:
Chi-square = -2(LL (0) – LL (β)), where Model1 =5328.7318, Model 2=5332.292 and Model 3=5977.4086. As all numbers are larger than the critical value, the p-value is <0.001. Therefore, all models I invest are better than the null model. And with more attributes add in, the prediction of choice is increasing.
21 5.4 Models interpretation
The first estimating model (Model 1) is the basic model; the attributes only include the car brand, level of autonomy, price per hour and non-option. The results indicate that all attributes are significant. Due to the coefficients are hard to interpret in MNL model. Therefore I obtain the marginal effects for predicting the real probabilities. According to the Table 5, the partially automated car shows a negative preference compare with the full level of the automatic car. This means respondents prefer the full level of autonomy. Furthermore, with the increase of price level the utilities become relatively less preferred. In general, Tesla Motor is the most preferred brand and followed by Mercedes Benz, Google, Apple, Ford, and Robocab. The non-option choice indicates more respondents choosing “Yes” option than “no” option. In the basic model, the brand of car has the highest impact on those interviewed' choice (37.53%), next come the Non-option (29.51%), price per hour (28.92%) and level of autonomy (4.04%). In Table 6, there is an overview of model importance comparing model 1 and model 2.
Next, I extend Model1 to Model 2 by including brand covariates inside, which are brand quality, brand familiarity, extension preference, image fit and category fit. Besides the non-option (p=o, o8), all attributes are significant in Model 2 (p<0.05). The result indicates that the brand effects are mostly influenced by the brand quality, brand familiarity, extension preference, image fit and category fit. Specifically, brand familiarity adds 0.07 to the whole brand effect; brand quality adds 0.10 to the total effect, extension preference adds 0.13. Image fit and category fit add 0.09 and 0.03 respectively. Therefore, they support the hypotheses H1A, H1B, H1C H1D and H1E. In model 2, Mercedes-Benz is the most preferred brand and followed by Tesla Motor, Robocab, Ford, Google and Apple. Due to Latent Gold provide a comparison of their effect magnitude, according to Table 6, the results illustrate that price per hour has the highest impact on choice (31.53%). Following by extension preference (16.27%), brand quality (11.68%), image fit (10.55%), car brand (10.48%), brand familiarity (8.50), the level of autonomy (3.97%), non-option (3.84%) and category fit (3.18%). Therefore, it can be concluded that respondents' preference to a SDV extension is the most important, and then brand quality serves as the second important brand related factors in choosing a SDV, however, the category fit is less relevant for respondents when evaluating a self-driving car.
Table 5 Model Specification
Model for choices 1 Model for choices 2 Attributes Class1 p-value Class1 p-value Car brand
Tesla 0.1185 5.5E-121 0.0278 2.20E-13
Apple -0.0216 -0.0504
Google 0.0175 -0.0414
M Benz 0.1124 0.0363
Ford -0.0351 0.0095
22
Level of autonomy
1 -0.0167 7.60E-06 -0.0164 2.00E-05
2 0.0167 0.0164
Price per hour
€10 0.107 3.80E-78 0.1214 4.00E-88 €12 0.0639 0.0705 €14 0.0244 0.023 €16 -0.0016 -0.0029 €18 -0.0615 -0.0722 €20 -0.1322 -0.1397 None option 0 0.122 7.40E-157 -0.0159 0.084 1 -0.122 0.0159
Brand related factors Brand familiarity 0.0704 0.00029
Brand quality 0.0967 5.00E-12
Brand extension preference
0.1348 1.10E-30
Brand image fit 0.0873 1.20E-12
Brand category fit 0.0264 0.017
Table 6 Model Importance
Model 1 Importance Model 2 Importance
Level of autonomy 4.04% Level of autonomy 3.97% Price per hour 28.92% Price per hour 31.53%
Car brand 37.53% Car brand 10.48%
None-option 29.51% None option 3.84% Brand familiarity 8.50% Brand quality 11.68% Extension preference 16.27% Brand image fit 10.55% Brand category fit 3.18%
5.3 Total brands effects Interpretation
After confirm all brand factors have a significant effect in customers' choice sets. Next I examine the total brand's effect in the extended model, for example, the total effect for Tesla is the parameter for brand quality times the percentage of consumers who associate Tesla with a high-quality brand. Therefore, the total effect of Tesla is
0.1252+0.07*84.70%+0.10*79.40% +0.13*86.40% +0.09*86.10% +0.03*80.20%=0.473
23
Therefore, the total brand effects are obtained in Figure 3. Mercedez Benz receives the most total brands effect consumers, following by Tesla, Ford, Google and Robocab. However, Apple receives the lowest total brand effects.
Figure 3 Total Brand Effects
Furthermore, the specific effect for every brand are calculated in Table 7 below, as all parameters are statistic significant, the γ estimates are significantly supported Eq.2
Next, I compare the importance of brand in model 1 and the importance of brand in model 2
The difference between 37.53%-10.48%=27.5%, therefore this percentage can be explained by the total brand effects for every brand.
Table 7 Specific Brand Effects
Tesla Motors Apple Google Mercedes-Benz Ford Robocab Familiarity 0.05963 0.063 0.067 0.066 0.065 0.001 Quality 0.077 0.084 0.073 0.090 0.042 0.002 Extension Attitude 0.1165 0.070 0.090 0.107 0.052 0.019 Image fit 0.075 0.055 0.067 0.065 0.039 0.013 Category fit 0.021 0.007 0.016 0.011 0.006 0.004
To see partial effects for every drive, I add brand familiarity, brand quality, extension preference, image fit and category fit separately in the basic model. The results indicate all marginal effects are significant. And then compare with these six models in Table 8. It displays the marginal change after adding every brand related factors separately. Specifically, with the covariates of brand quality, the price is more important than benchmark model if only consider the brand quality factors.
0.473 0.052 0.127 0.503 0.247 0.121
Tesla Apple Google M Benz Ford Robocab
24
And also the brand quality factors decrease the importance of non-option. Regarding brand familiarity, the preference ranking of the brand is consistent with the basic model, and similar to brand quality, the effect of familiarity of brand increase the importance of price and decrease the importance of non-option. Following with the extension preference, Benz is a preferable brand compare with other brands to launch a SDV, after that are Tesla, Google, Ford, Apple and Robocab. The importance of price is increase and non-option is decrease. Next, M Benz receives most image fit and following by Tesla, Google, Ford, Apple and Robocab. Lastly, for category fit, Benz, Tesla, and Google still rank before Apple, Ford and Robocab. And the importance change is not significant in category fit.
Table 8 Partial Brand Effects
Basic Model Brand Quality Brand Familiarity
25 Table 9 Partial Brand Importance
Basic Model Brand Quality Brand Familiarit y Exten.Prefe r Image Fit Category Fit Level of Autonomy 4.04% 4.59% 4.55% 3.68% 4.53% 3.98%
Price per Hour 28.92% 32.76% 33.23% 32.16% 30.99% 28.25% None Option 29.51% 16.90% 15.69% 15.12% 16.95% 22.97% Car Brand 37.53% 23.39% 23.90% 23.76% 26.84% 32.53% Brand Quality 22.35% Brand Familiarity 22.62% Extension Prefer. 25.27% Image Fit 20.69% Category Fit 12.27% 5.3 Moderating effects
To test the moderating individual consumer characteristic effects, the interaction effects of consumers' innovativeness and perceived risk are interacting with brand related factors, the level of autonomy and price per hour. I add variables in the choice sets that multiply consumers' innovativeness and perceived risk with each of the effects coded attribute level. Here I only include the significant moderating effect in the table. All insignificant interpretation can be found in Appendix D. According to Table 10, besides the non-option option, the results show all the main effects in model 2 are significant. Due to the moderating effect, the selection of car brand is changing. Unlike Model 2 the most preferred car is Tesla Motors, and next are Mercedes Benz, Robocab, Ford, Google and Apple.
Table 10 Model 3 Specification
Model for choices 3
Attributes Class1 p-value
Car brand Tesla 0.0445 5.70E-06 Apple -0.044 Google -0.0351 M Benz 0.026 Ford 0.0005 Robocab 0.0082 Level of autonomy 1 -0.0377 3.00E-12 2 0.0377
Price per hour
26
0 -0.0001 0.99
1 0.0001
Brand familiarity 0.0725 0.00022
Brand quality 0.0916 9.00E-11
Brand extension preference 0.1376 1.80E-31 Brand image fit 0.0864 3.00E-12 Brand category fit 0.0275 0.014 Innovation*Tesla Motor 0.0063 0.0096 Innovation*Level of auto -0.0025 0.023 Innovation*Non-option -0.0119 6.60E-06 Perceived Risk*€10 -0.0133 0.00046 Perceived Risk* Level of auto 0.0101 1.10E-08 Perceived Risk* Non-option 0.0129 0.0029
Table 11 Model 3 Importance
Model 3 Importance
Car brand 3.99% Perceived price*€10 7.21%
None option 0.01% Perceived price* Level of auto 5.44% Brand familiarity 3.27% Perceived price* Non-option 6.37%
Brand quality 4.13%
Extension preference 6.20% Brand image fit 3.89% Brand category fit 1.24% Innovation* Tesla Motor 3.38% Innovation*Level of auto 1.34% Innovation*Non-option 6.41%
Specifically, there is a significant moderate effect (p=0.02) between consumers’ innovativeness with the level of autonomy in respondents’ choice. If consumers are more innovative, they are less likely to choose the partially automated car than to choose the fully automated car. The utility of choosing the partial type is decreased by -0.0377-0.0025=--0.0402. Moreover, there is a significant interaction between innovations with Tesla Motor. The utility of choosing the Tesla Motor is increased by 0.0445+0.0063=0.0508, which means the most innovative of the respondents, the more likely to choose the Tesla Motor.
27
self-driving car, the less likely to choose the cheapest option. And this makes the price in total less important than the model that is not taking account with moderating effect, as the range gets narrower. There is a significant moderate effect between perceiving risk and level of autonomy. If respondents feel more risk in choosing an autonomous car, the more likely to choose a partial level selfdriving car. Specifically, the utility of choosing the partial level autonomy is increased by -0.0377+0.0101=--0.0276
28
Chapter 6: Discussion
6.1 Conclusion
The goal of this research is to examine if brand related factors will influence consumers' perception to adopt an autonomous vehicle. More specifically, what are the strong brands related drivers could influence the consumers' preference utilities. To address my research questions:
How important is the brand name when evaluating autonomous vehicles and which levels of these attributes consumer prefer?
According to the CBC study in this research, the results from the basic model indicate that brand name serves as the most important factors in evaluating a SDV compared with the rest attributes, such as the Non-option, the price, and the level of autonomy. Therefore, when evaluating a self-driving car, the brand name still plays a significant role, next important factor is the price, and the level of autonomy only has a little effect.
Which attributes are critical in this adoption process?
29 6.2 Hypotheses discussion
Brand familiarity
Consistent with the research hypothesis, more familiar with brands consumers will generate more positive attitudes to an extension than unfamiliar brands. Therefore, the brand effects can be explained by the familiarity with the brand name. To be more specific, if consumers perceive the brand as a familiar brand, the utility of choosing this brand will increase by 0.07 compare with a non-familiar brand.
However, the interesting finding of this driver is brand familiarity is not as important as extension preference, quality of the brand, and image fit. This is because consumers may be familiar with some brands but do not satisfied with these products. At certain points, being familiar with some brands can reduce consumers’ uncertainty to choose the SDV, but according to this SDV adoption research, other brand related drivers have more importance effects in overall brand effect.
Quality of parent brand
The hypothesis for the quality of the brand is supported by the research finding. If consumers perceive the certain brand as a high-quality brand, they will be more likely to choose this brand. Previous research mostly proves this driver in FMCG industry (Bottomley & Holden, 2001; Volckner & Scattler, 2006; Albrecht et al., 2013), and this research shows this driver is also significant influence consumers’ decision making in a technology innovation industry. Due to most consumers concern their safety of using the robotic, therefore a high-quality brand can signal a warranty of safety (P. Song et al.,2010). Therefore, when facing an incremental innovation product, consumers evaluate the brand utilities based on quality can reduce the cost and risks. Furthermore, according to the CBC model, I calculate that quality driver is the second highest total effect on brand related factors. If consumers associated the brand as a good quality, then it will increase about 10% possibility to choose the certain brand compare with they do not perceive this brand as a good quality brand.
Attitude toward the extension
Attitude towards the extension has the strongest total effect on brand based on this research’s finding. The hypothesis illustrates that the better attitude to the brand extension, the high evaluation to adopt the self-driving car is significant. According to the choice model, people associate with a preferable brand to an autonomous car will increase the utilities to choose the brand by 0.13 compare with people who do not associate with a preferable brand. Although, based on a large number of brand success extension articles, only two research have shown this driver is significant
30
the attitude toward a brand extension is a crucial role, which proves the current research (Science Daily, 2016) says that the psychological factors are the key factor that influences the SDV adoption. However, attitude toward an SDV extension brand may contain many factors, and may at the meantime interact with other brand extension drivers as well. Therefore, more specific details will be discussed in the further research section.
Image fit and category fit
Both image fit and category fit are found significant influencing the choice sets, which are if consumers perceive the SDV is image fit or category fit with the brand, the utilities choosing certain brand will increase. Image fit has the third highest effect on brand factor, where people associate with image fit toward a brand has higher utilities (0.08) to adopt the SDV than people who do not perceive this brand as image fit. However, category fit has the least impact on choice sets, the utilities of people who perceive the certain brand as category fit is 0.02 higher than people who to do not associate with the certain brand as category fit. The definition of category fit toward a self-driving car is the brand shares the concrete function features or attributes similarity (Broniarczyk & Alba, 1994). But the image fit is more relates to the brand name (Boush et al., 1987). These results show that when consumers are evaluating a SDV, they will concern more about the brand name instead of considering if this brand offers similar attributes.
In the literature review, I assume that Google & Apple can represent a high technology industry, hence can receive great image fit as autonomous car belongs to an incremental innovation technology. However, pick any procedure shows that only 69.75% of people perceive these companies perceive image fit and 60% people think the traditional automakers (Mercedez Benz & Ford) also perceive the image fit. Therefore, I cannot conclude that if consumers perceive the image fit to the technology company has a higher evaluation of adoption. Moreover, neither can I conclude that if consumers perceive the category fit, then mainstream automaker companies have a higher evaluation of adoption. This is based on the pick any procedure results, 44.8% people perceive technology companies as category fit, and 32.8% people perceive the automaker companies fit.
Moderating effect of consumer innovativeness and perceived risk
31
innovation car brand and more likely to choose Tesla. This brand is consistent to my assumption; as more innovative people will seek for the least image fit nor category fit (Xie, 2008). Based my hypothesis Robocab as an unknown brand should have significant moderating effect with consumer innovativeness. Therefore, for a totally brand new brand to entering the SDV industry is quite crucial. Another interesting finding is that if people perceive risk in choosing a SDV, the importance of the price is decreasing. In another word, if the consumer perceives a higher risk, they will pay less attention to the price. This is due to a higher price may signal a good quality of a brand. To reduce the perceive risk, respondents may spend relatively higher price in choosing a SDV. 6.2 Theoretic implication
This research provides detail information for brand extension to an innovation product, the successful brand drivers in FMCG industries are still found significant in an innovation product industry. Furthermore, this research extends the current research by adding brand related factors in the SDV introduction with respect successful brand extension. These specific drivers are brand familiarity, quality of parent brand, image fit and category fit. According to our basic model, the total effects for brands attributes have the largest effect on consumers' adoption. After including the brand related factors drivers in the model, it shows the band attributes are widely explained by these drivers (Indirect effects).
Furthermore, this research uses choice-based conjoint analysis to measure respondents’ preference instead of ranking or rating. Particularly, for an innovation product, by combining different attributes and levels can make an approximate prediction about the adoption rate of new products (Eggers, Eggers, & Kraus). This research combines most possible determinants for a SDV from the latest market research (BCG, 2016) and analysis how respondents value each component and come up a quantifiable result. Currently, only few research use CBC method to evaluate the successful brand extension drivers nor use CBC for adoption modeling. Therefore, these results indicate that how respondents' decision changing that involves a trade-off, and it also shows that what are the most crucial factors that are influencing the decision changing.
6.3 Managerial implication
32
Firstly, SDV brand extension success focuses mainly on consumers' preference for the brand extension brand. Therefore, compare with other drivers like image fit or brand familiarity, if consumers do not prefer this brand launching SDV, the brand extension may still not successes. Secondly, from pick any method, it is hard to category respondents in the different group based on how they perceive a SDV. Whether they see SDV is a technology product, or a mainstream car? Hence, it is difficult to conclude under which type kind of industries will successfully be launching the SDV. But it can conclude that for a start-up, the situation to enter the SDV market is difficult, as our models show a low utility for the SDV with an unknown name. But it is advisable for these new startups to firstly build their brand by increasing consumers’ familiarity, providing good quality products, improve the SDV extension preference, etc. And for an established company, it receives decent brand familiarity and brand quality. They should focus on finding out what drives consumers' preference to this extension. Moreover, this company should clearly be positioning in certain markets that can clearly inform consumers. Hence customers can develop clear image fit or category fit toward this extension product.
33
Chapter 7: Limitations and further research
7.1 Limitations
In order to shorten the survey, I only choose one item to measure a brand related factor. Therefore the single item may not fully explain different facets of every factor. However, according to literature review I summary items based on a multiple-item measure which can indicate the choosing item still represents the similar respondents' state. Therefore, a single item is still valid measuring respondents' perception of brand residuals.
According to the models, there are still unexplained brand residuals influencing the brand choice utilities as the brand residuals are not approaching “0”. For example, Apple receives the least total brand effects; this could be some other brand factors may not include in my model. Such as brand loyalty. As the goal of this research is trying to identify the key brand drivers for adopting an autonomous car, I choose these key drivers based on most frequently discussed drivers from previous studies. There may be some factors that will influence the brand extension to SDV but in very specific points. Therefore, these drivers may not be that significant important in a brand related factors in general.
Moreover, the driver of attitude to an extension is found to be the most important driver, but this factor may interact with other brand related factors. For example, an individual who perceives Mercedes-Benz perceive a good quality car manufacture, so they more prefer Mercedes-Benz offer him an SDV instead of other brands. As discussed before, the main goal of this research is firstly identifies what factors will influence the respondents' choice to an autonomous. So for the further analysis can generate a more insight to attitude to an extension or find out if this factor interacts with other factors.
Lastly, in designing the attributes, I try to generate Google and Apple as a technology company, Mercedes-Benz and Ford as mainstream automakers. However, due to Mercedes-Benz belong to a luxurious brand when respondents compare it with Ford, Ford will receive less preference. When directly compare these brands as technology companies with mainstream automakers will not valid, as consumers can hardly perceive the similarity between Mercedes-Benz and Ford.
7.2 Direction for further research
34
Moreover, it is also interesting to see to what extend this research can generate to other innovative industries, such as VR industry. Although this industry also can represent an incremental innovation industry, due to it is an entertainment product, consumers' perceiving risk may entirely different from consumers' perceiving risk to a SDV car.
Lastly, the samples are collecting based on the Netherlands, and most respondents are Dutch. According to BCG report (2016), different countries show significant different attitudes to SDV. Hence, different cultures may result in different respondents’ choice. Therefore, for latter research it can include the cultural heterogeneity as a factor, which allows compassion between different cultures.
35
Reference list
Aaker, D.A, Keller, K.L. (1990). Consumer evaluations of brand extensions, Journal of Marketing, 54, 27–41
Adell. E. (2014). Driver acceptance of car navigation systems: integration of locational accuracy, processing speed, and service and display quality with technology acceptance model, Personal and
Ubiquitous Computing, 18 (3), 503-513
Ajzen, I. (1991). The theory of planned behavior, Organ. Behav. Hum. Decis. Process. 50(2), 179– 211
Alba, J.W, Hutchinson, J.W. (1987). Dimensions of consumer expertise, Journal of Consumer
Research, 13, 55–59
Albrecht, C.M., Backhaus, C., Gurzki, H. and Woisetschläger, D.M. (2013) ‘Drivers of brand extension success: what really matters for luxury brands’, Psychology and Marketing, 30 (8), 647– 659.
BCG (2016). Autonomous vehicle adoption study. (Accessed on Sep.13, 2016)
http://www.bcg.com/expertise/industries/automotive/autonomous-vehicle-adoption-study.aspx
Bottomley, P.A.; Doyle, J.R. (1996) The Formation of Attitudes Towards Brand Extensions: Testing and Generalising Aaker and Keller’s Model, International Journal of Research in
Marketing, 13(4), 365-377.
Boush, David M., Shipp, S., Loken, B (1991), A Process-Tracing Study of Brand Extension,
Journal of Marketing Research 28, 16-28
Boush, David M., Shipp, S., Loken, B., Gencturk, E., Crockett, S., Kennedy, E., Minshall, B., Misurell, D., Rochford, L.& Strobel, J. (1987), II Affect Generalization to Similar and Dissimilar Brand Extensions", Psychology & Marketing, 4 (3), 225-237
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology MIS Quarterly, 13, 319–340
Eggers, F. Eggers, F & Kraus, S., (2014). Entrepreneurial branding: measuring consumer preferences through choice-based conjoint analysis. Int Entrep Manag J, 12: 427-444 Griffith (2016). Who will build the next great car company? (accessed Sep.25,2016)
http://fortune.com/self-driving-cars-silicon-valley-detroit/
Hair, Jr., J.F., Black, W.C., Babin, B.J., Anderson, R.E., Tatham, R.L., 2006. Multivariate data analysis (6th Ed.), Pearson-Prentice Hall, Upper Saddle River, NJ.
He, H, & Li, Y, 2011, 'Key service drivers for high-tech service brand equity: The mediating role of overall service quality and perceived value', Journal of Marketing Management, 27, 77-99
Hem, L.E., de Chernatony, L., & Iversen, N.M. (2003). Factors influencing successful brand extensions. Journal of Marketing Management, 19, 781-806
36
Jin, Zou. (2013). Extend to online or offline? The effect of web-brand extension mode, similarity and brand concept on consumer evaluation. Journal of Marketing Mangement.29(7-8), 755-771 Keller, K.L. (2003). Strategic Brand Management: Building, measuring, and managing brand equity, 2nd ed., Prentice-Hall, New York, NY.
Kent, Robert J. and Chris T. Allen (1994), "Competitive Interference Effects in Consumer Memory for Advertising: The Role of Brand Familiarity," Journal of Marketing, 58 (July), 97-105
Klink, R. R., Smith, D. C., 2001, Threats to the external validity of brand extension research.
Journal of Marketing Research 38, 326-335
Louho, R., Kallioja, M. & Oittinen, P. (2006). Factors affecting the use of Hybrid media applications. 35 (3) 11-21.
Mao, H., Krishnan, H.S. (2006). Effects of prototype and exemplar fit on brand extension evaluations: A two-process contingency model. Journal of Consumer Research, 33(1), 41–49
Matines, M. Pina, (2010). Consumer response to brand extension: a comprehensive model.
European Journal of Marketing, 44 (7-8), 1182-1205
Midgley, D. F. and Dowling, G. R. (1993), "A Longitudinal Study of Product Form Innovation: The Interaction between Predispositions and Social Messages," Journal of Consumer Research, 19, 611-625.
Milberg, Sinn, Goodstein. C. (2010). Consumer Reactions to Brand Extensions in a Competitive Context: Does Fit Still Matter? Journal of Consumer Research. 37(3), 543-553
Holbrook, B. William L. (1984), The Pick-Any Procedure Versus Multidimensionally-Scaled Correlations: An Empirical Comparison of Two Techniques for Forming Preference Spaces, in Association for Consumer Research 11, 56-62,
Park, Kim. (2014). Driver acceptance of car navigation systems: integration of locational accuracy, processing speed, and service and display quality with technology acceptance model. Pers Ubiquit
Comput, 18, 503-513.
Queensland University of Technology (2016). People not technology will drive success of autonomous vehicles. (Accessed Oct.5, 2016).
https://www.sciencedaily.com/releases/2016/08/160802103706.htm
Rangaswamy, A, Raymond B and Terence (1993), “Brand Equity and the Extendibility of Brand Names,” International Journal of Research in Marketing, 10, 61-75.
Rishi, Sanjay, Stanley & Gyimesi. (2008) “Automotive 2020: Clarity beyond the chaos.” IBM Institute for Business Value. Accessed on Oct.3, 2016) http://www-304.ibm.com/
easyaccess/fileserve?contentid=164523
37
Samaradiwakara; C, G, Gunawardena. (2014). Comparison of existing technology acceptance theories and models to suggest a well improved theory/model. International Technical Sciences
Journal, 1 (1), 21-36
Sheinin, D.A. and Schmit, B.H. (1994), "Extending brands with new product concepts: the role of category attribute congruity, brand affect, and brand breadth", Journal of Business Research, Vol. 31(1), 1-10
Yung-Cheng Shen, Lien-ti Bei and Chia-Hsien Chu. (2011) Consumer evaluations of brand extension: The roles of case-based reminding on brand-to-brand similarity. Psychology and
Marketing 28(1), 91-113
Steenkamp, Jan-Benedict E. M., and Katrijn Gielens (2003). “Consumer and Market Drivers of the Trial probability of New Consumer Packaged Goods,” Journal of Consumer Research, 30(3), 368-384
Volckner F, Sattler H (2006). Drivers of brand extension success. Journal of Consumer Research, 70(2): 18-34
Wanke, M., Bless, H., & Schwarz, N. (1998). Context effects in product line extensions: Context is not destiny. Journal of Consumer Psychology, 7(4), 299–322
Xie, H (2008). Consumer innovativeness and acceptance of brand extensions, Journal of Product &
38
Appendix A
Factors not list in previous table Conclusion Volckner &
Sattler(2006)
Marketing support** Retailer acceptance** History of previous extensions**
Parent-brand conviction** Parent –brand experiences
Fit, marketing support, retailer acceptance, parent-brand conviction and parent-brand experiences are critical drivers
for brand extension. Martinez &
Pina(2010)
N.A The most important aspects driving brand extension is brand image. Brand familiarity does not directly influence
extension attitude. Albrecht et
all(2013)
Hedonic value(Luxury)** Category involvement** Prestige ,affiliation, and uniqueness
not significant
The influence of functional value is more important for non-luxury brands, the hedonic factor is only drive non-luxury
brands’ extension, Milberg et
al(2010)
N.A Without consider the fit, extension performs better paired with relatively unfamiliar versus familiar competitors. Shen, Bei &
Chu(2011)
Brand-to brand similarity An existing brand with similar brand concepts in the new product category can improve consumers’ acceptance of
brand extension. Jin & Zou(2013) Purchase intension
Evaluation of the extension
39
He &Li (2010) Brand loyalty ** Fit moderates brand loyalty on brand extension, and moderating fit is further moderated by technologic direction Hem,
Chernatony &Iversen(2003)
Brand reputation** Similarity, reputation, perceived risk and innovativeness are crucial factors that influencing the brand extension
40
Appendix B
Scale/Item Source
Brand familiarity
Extend to which partisans agreed/disagreed with the following statements (-3 = “strongly disagree”, and 3 = “Strongly agree”)
Regarding <Product brand>, I am familiar with this brand. I am knowledgeable with <Product brand>
Kent & Allen, 1994
Quality of current brand
Extend to which partisans agreed/disagreed with the following statements (-3 = “strongly disagree”, and 3 = “Strongly agree”)
the [ product name] offers high quality products
the quality of [ product name] product is far above average.
Aaker & Keller, 1990; Sheinin & Schmitt,1994
Attitude toward extension
Extend to which partisans agreed/disagreed with the following statements (-3 = “strongly disagree”, and 3 = “Strongly agree”)
I like [ product name] <extension> [ product name] <extension> is attractive
Volckner & Sattler, 2006)
Image fit
Global similarity between the parent brand and the extension product (-3 = “not very similar”, and 3 = “very similar”)
Would the people, facilities and skills used in making the original product be helpful if the manufacturer were to make the extension (-3 = “not very helpful”, and 3 = very helpful”)
Aaker & Keller, 1990; Volckner & Sattler, 2006
Category fit
Extend to which partisans agreed/disagreed with the following statements (-3 = “strongly disagree”, and 3 = “Strongly agree”)
[ product name] is closely tied to the attributes of the original product category
My association with [ product name] are closely tied to the attributes of the original product category
Ranggaswamy, Burke & Oliva, 1993
Innovativeness
Extend to which partisans agreed/disagreed with the following statements (-3 = “strongly disagree”, and 3 = “Strongly agree”)
Klink & Smith, 2001
41
Overall I enjoy buying the latest products I like to buy new products before others
Perceived risk
If I buy <extension>, I would feel very uncertain of the level of quality that I am getting
I prefer buying a well-known [product name] <extension>, because I need the reassurance of an established brand offering this product.
43
