Brand Experience Clustering and Depiction of Brand Satisfaction, Brand Loyalty and Emotional Confidence

Widespread brand choice and consumers empowered by digital gadgets have posed a considerable challenge for marketing practitioners to sustain the long-term brand-consumer relationship (Das, et al., 2018). To understand the long-term brand-consumer relationship, different evaluative (i.e., brand attitude) and motivational (i.e., brand attachment) marketing constructs have been tested (Ajzen, 2005; Park, et al., 2010). These constructs have certainly enriched our understanding of the relationship. However, emotional confidence, despite its recognition and importance, has received little or no attention in the academic literature (Kidwell, et al., 2008; Rizvi and Oney, 2018). The emotional confidence (EC) is defined as affective certainty based strong emotional response, which is essential for a long-term brand-consumer relationship (Rizvi and Oney, 2018).

The EC as a construct has just surfaced in the literature and its antecedents and consequents are not well documented. Therefore, this study explores its relationship within the existing marketing frameworks. Rizvi and Oney (2018) argue that emotional confidence is a consequence of overall brand experience; presumably, stronger the brand experience, stronger is the affective certainty towards a brand.

Brand experience is considered a core interaction that may lead to various consequences such as brand satisfaction and brand loyalty (Brakus, et al., 2009). Brand experience as a construct is well documented in the academic literature, and it is further explored for possible consumer profiling, namely there are four dimensions of brand experience: sensory, affective, intellectual and behavioral (Zarantonello and Schmitt, 2010) and how it influences other constructs like brand commitment (Das, et al., 2018).

This study investigates the influence of brand experience on emotional confidence and whether or not emotional confidence influences brand satisfaction and brand loyalty. Further, clusters of brand experience dimensions were elicited to analyze the relevance of the elicited clusters concerning emotional confidence, brand satisfaction and brand loyalty.

To the best of our knowledge, no study has investigated the influence of brand experience on emotional confidence. It is important to note that consumer profiling based on the brand experience dimensions exists and its relative impact on purchase intentions is affirmed. However, such consumer profiles are not analyzed regarding emotional confidence, brand satisfaction, and brand loyalty. This paper mainly deals with two major questions: 1)

Does brand experience influence emotional confidence, and its relationship with brand satisfaction and brand loyalty is investigated?

1.1

Background

Brand experience has captured the attention of both academics and practitioners. Brand experience is a multi-dimensional construct consisting of sensory, affective, intellectual and behavioral dimensions (Brakus, et al., 2009). Although brand experience can be positive or negative, its measure implies positive experience towards a brand. It is neither an evaluative construct like brand attitude, nor a motivational construct like brand attachment. The dimensions of the brand experience are also used to profile consumer and to investigate its influence on purchase intentions (Zarantonello and Schmitt, 2010). The brand experience is used to assess its impact on brand satisfaction, brand loyalty and brand commitment (Das, et al., 2018; Brakus, et al., 2009).

Confidence as a construct is widely used in behavioral science. It is considered a key and one of the most critical indicators of economic health. Despite its widespread use, the conceptual domain of confidence remains uncharted. The conceptual ambiguity poses a challenge regarding how to measure it, mostly confidence as a construct is conceptualized as a cognitive construct; despite highlighting urge and importance of emotional confidence (Kidwell, et al., 2008), very little progress has been made (Rizvi and Oney, 2018). Emotional confidence is conceptualized as an affective certainty based on strong emotions. However, emotional confidence is neither an evaluative construct like brand attitude, nor a motivational construct like brand attachment; conceptually, it is similar to brand experience. However, emotional confidence is a consequence of brand experience. Positive or negative brand experience may result in affective certainty (i.e., emotional confidence) to act. Both strong positive and negative brand experience can trigger affective certainty to act or not to act. But it is important to note in this study: emotional confidence is measured based on “strong positive emotions”. From a measurement perspective, brand experience implies positive experience, even though brand experience can be positive or negative.

Stemming from recent development, Rizvi and Oney (2018) argue that emotional confidence can be a consequence of overall brand experience. Since antecedents of emotional confidence are not yet specified in the literature, it will be interesting to see whether or not brand experience influence emotional confidence. Moreover, Brakus, et al. (2009) have shown that brand experience could affect brand satisfaction and brand loyalty. In this study, we employ the same framework only by adding emotional confidence as a consequence of brand experience and as an antecedent of brand satisfaction and brand loyalty. Further, consumers are clustered based on brand experience dimensions to depict the relationship among emotional confidence, brand satisfaction, and brand loyalty. Based on the theoretical model as well as the theoretical background presented above, the following hypothesis has been developed for the research study (the conceptual framework can be seen in Fig. 1):

H₁: Brand experience has a direct and a positive influence on emotional confidence.

H₂: Emotional confidence has a direct and positive influence on consumer brand satisfaction.

H₃: Consumer brand satisfaction as a direct and positive influence on consumer brand loyalty.

H_4: Emotional confidence has a direct positive influence on brand loyalty.

Figure 1

The focus of this study was two-pronged, first to assess the influence of brand experience on emotional confidence, second clustering of brand experience dimensions to depict the relationship between emotional confidence, brand satisfaction and brand loyalty. A self-administered questionnaire was used to collect data (n = 240) in the context of car buying. Judgment sampling was used to obtain the data as prior experience (users of the brand) was warranted; so data was collected from various car service centers and the participation in the study was voluntary.

For this, validated scales of all constructs were used and the measures included brand experience, brand satisfaction, emotional confidence and brand loyalty (Brakus, et al., 2009; Oliver, 1999; Rizvi and Oney, 2018).

Structural equation modeling (AMOS) was used to analyze the data for the first part of the paper. For the second part of the paper, the four dimensions of the brand experience: sensory, affective, behavioral and intellectual were clustered to further assess the relationship among emotional confidence, brand satisfaction and brand loyalty within those clusters. Hierarchical cluster analysis was used (SPSS), as there were no prior assumptions regarding the number of clusters; the clusters were extracted by employing Ward's method to get a sizable cluster for further analysis. To validate the solution, both sub-sample and discriminant analysis were performed.

Table 1 contains a descriptive analysis of demographic profile obtained from 251 respondents based on four demographic aspects, namely age, marital status, family income and driving Toyota experience.

Accordingly, the prime number of respondents, that is, 85 (34%) respondents from the age group 36–50 years, while 66 (26.4%) respondents from age group 17–25 years. Likewise, 54 (21.6%) respondents from age group 26–35 years. Also, the results indicate that 156 (62.4%) respondents were married and 94 (37.6%) respondents were single. Regarding family income, 95 (38%) respondents were from the family income group 75,001–125,000 PKR. Also, 78 (31.2%) respondents were from the family income group 125,001–200,000. About driving Toyota experience, the demographic profile results indicate that 120 (48%) respondents were 1–3 years’ experience of driving Toyota as compared to 74 (29.6%) respondents with 4–5 years’ experience.

3.3

Confirmatory Factor Analysis

For factor extraction, a principal component analysis method was used to establish an independent linear combination of observed variables. Extraction values provided in Table 2 show the amount of variance that is explained by each factor. The variables with a high amount of variance indicate the strength of variable regarding their highest representation, whereas, variables with the lowest amount of variance indicates the low representation in the factor space. For example, BEX4 = 0.896 and BEX3 = 0.859 indicates the high proportion of variance. In contrast, BL4 = 0.630 and BEC7 = 0.649 indicates the low proportion of variance in the factor space. Scree plot also indicates the percentage of “variance” in a factor structure (space). A scree plot exhibited in Fig. 2 shows the maximum variation in 6 factors as the eigenvalues on the y-axis flatten at 6^th factor and remaining factors explain a minimal variation.

Figure 2

3.4

Confirmatory Factor Analysis

KMO value greater than 0.7 is an indication for “confirmatory factor analysis” (Hair, et al., 2010). Therefore, IBM-AMOS v19 was used to test the “confirmatory factor analysis” (see Fig. 3) independently before the “structural equation model” (Anderson and Gerbing, 1988). In order to use the “maximum likelihood” estimation method, data normality was confirmed as the values of Kurtosis (−0.548 & 0.825) and Skewness (−1.485 & −0.838) shown to be normally distributed and consistent with the suggested limits (Kline, 2015). Refer to Table 4, all factor loadings (λ values) of 04 variables and 22 items are measured above-suggested limit 0.50 statistically significant at p < 0.000.

Figure 3

Table 4

CFA based Reliability & Validity (Source: Own research)

Variables	Factor Loading	t-value	R2	α (alpha)	CR	AVE
Emotional Confidence
BEC1	0.855	Fixedᵒ	0.731	0.920	0.944	0.875
BEC2	0.906	19.856	0.820
BEC3	0.898	19.536	0.806
BEC4	0.888	19.099	0.788
BEC5	0.827	16.788	0.684
BEC6	0.813	16.319	0.662
BEC7	0.718	13.400	0.515
Brand Satisfaction
BPQ1	0.772	Fixedᵒ	0.596	0.925	0.963	0.682
BPQ2	0.761	11.818	0.580
BPQ3	0.844	12.933	0.713
BPQ4	0.742	11.504	0.551
Brand Experience
BEX1	0.775	Fixedᵒ	0.600	0.882	0.815	0.572
BEX2	0.755	6.821	0.570
BEX3	0.985	4.999	0.970
BEX4	0.794	Fixedᵒ	0.631
BEX5	0.645	5.806	0.471
BEX6	0.964	Fixedᵒ	0.928
Brand Loyalty
BL1	0.805	10.754	0.649	0.838	0.802	0.851
BL2	0.837	11.072	0.701
BL3	0.873	11.395	0.763
BL4	0.652	Fixedᵒ	0.425
BL5	0.766	10.337	0.586

ᵒ

Based on ML estimation fixed parameter @ 1.0

α – value: internal consistency; CR – composite reliability; AVE – average variance extract

The composite reliability (CR) values (0.886 and 0.960) and average variance extraction (AVE) values (0.568 and 0.82) also measured between the suggested limits thus satisfied the “structural reliability” and “convergent validity” of the construct(s) used in this study (Bagozzi and Yi, 2012; Fornell and Larcker, 1981).

Refer to the results summarized in Table 5, the χ² value attained significance at p < 0.000. In case of all four variables (i.e., emotional confidence, brand experience, brand satisfaction and brand loyalty), model fit indicators (e.g., AGFI, RMR, CFI, TLI, NFI and RMSEA) are methodically compatible with the sample data.

Table 5

Confirmatory Factor Analysis (Source: Own research)

Fit Indices
	χ2	df	χ2/df	AGFI	RMR	CFI	TLI	NFI	RMSEA
Emotional Confidence	87.324	34	2.56	0.798	0.075	0.954	0.931	0.946	0.145
Brand Experience	16.034	06	2.67	0.925	0.080	0.981	0.952	0.970	0.082
Brand Satisfaction	2.568	02	1.28	0.975	0.020	0.999	0.996	0.994	0.034
Brand Loyalty	19.016	05	3.80	0.915	0.049	0.980	0.960	0.973	0.106

Probability level = 0.000

χ² - chi-square; AGFI - adjusted goodness-of-fit; RMR - root-mean-square; CFI - comparative fit index; TLI - Trucker-Lewis Coefficient; NFI - normed fit index; and RMSEA - root-mean-square error of approximation

3.5

Structural Equation Model

Fig. 4 shows the path diagram of a Structural Model 1. The model fit results appear in Table 6. The χ² achieved significance at p < 0.000. Likewise, the level of discrepancy was measured with χ² / df. The measured value (i.e., 600.738 / 203 = 2.95 between 2 or 3) also shows the model fit. Besides, other 06 model fit indices such as, AGFI = 0.780; RMR = 0.250; CFI = 0.897; TLI = 0.883; NFI = 0.853 and RMSEA = 0.089 indicates an “adequate model fit” as per the suggested brackets.

Figure 4

Table 6

Fit Indices based on Model 1 (Source: Own research)

χ2	df	χ2/df	AGFI	RMR	CFI	TLI	NFI	RMSEA
600.738	203	2.95	0.780	0.250	0.897	0.883	0.853	0.089

Probability level = 0.000

χ² - chi-square; AGFI - adjusted goodness-of-fit; RMR - root-mean-square; CFI - comparative fit index; TLI - Trucker-Lewis Coefficient; NFI - normed fit index; and RMSEA - root-mean-square error of approximation

Hypothesis testing results based on Model 1 are listed in Table 7.

Table 7

Hypothesis Testing Results based on Model 1 (Source: Own research)

R	Estimate	R2	S.E	ᵒt-value	ᵒᵒp-value
H1: Brand experience has a direct and a positive influence on emotional confidence
0.786	1.228	0.617	0.178	6.890	***
H2: Emotional confidence has a direct and positive influence on consumer brand satisfaction
0.689	0.465	0.474	0.047	9.947	***
H3: Consumer brand satisfaction as a direct and positive influence on consumer brand loyalty
0.630	0.783	0.396	0.103	7.572	***

ᵒ

t value = 1.228/.178 = 6.890

ᵒᵒ

The probability of attaining a t-value (critical ratio) as high as 6.89 in absolute value is ≤ 0.001. More specifically, the regression weight for Brand Experience in the prediction of Emotional Confidence is significantly different from zero at the 0.001 level (two-tailed).

Fig. 5 shows the path diagram of a structural Model 2. The model fit results appeared in Table 8. The χ² achieved significance at p < 0.000. Likewise, the level of discrepancy was measured with χ² / df. The measured value (i.e., 560.370/202 = 2.77 between 2 or 3) also shows the model fit. Besides other 06 model fit indices, such as AGFI = 0.786; RMR = 0.186; CFI = 0.907; TLI = 0.894; NFI = 0.863 and RMSEA = 0.084, indicate an “adequate model fit” as per the suggested brackets.

Figure 5

Table 8

Fit Indices based on Model 2 (Source: Own research)

χ2	df	χ2/df	AGFI	RMR	CFI	TLI	NFI	RMSEA
560.370	202	2.77	0.786	0.186	0.907	0.894	0.863	0.084

Probability level = 0.000

χ² - chi-square; AGFI - adjusted goodness-of-fit; RMR - root-mean-square; CFI - comparative fit index; TLI - Trucker-Lewis Coefficient; NFI - normed fit index; and RMSEA - root-mean-square error of approximation

Hypothesis testing results based on Model 2 are listed in Table 9.

Table 9

Hypothesis Testing Results based on Model 2 (Source: Own research)

R	Estimate	R2	S.E	ᵒt-value	ᵒᵒp-value
H1: Brand experience has a direct and a positive influence on emotional confidence
0.797	1.260	0.635	0.181	6.955	***
H2: Emotional confidence has a direct and positive influence on consumer brand satisfaction
0.647	0.439	0.418	0.047	9.340	***
H3: Emotional confidence has a direct and positive influence on consumer brand loyalty
0.496	0.412	0.246	0.070	5.857	***
H4: Consumer brand satisfaction as a direct and positive influence on consumer brand loyalty
0.258	0.315	0.066	0.098	3.216	0.001

ᵒ

t value = 0.439 /0.047 = 9.340

ᵒᵒ

The probability of attaining a t-value (critical ratio) as high as 9.359 in absolute value is ≤ 0.001. More specifically, the regression weight for Emotional Confidence in the prediction of Consumer Brand Satisfaction is significantly different from zero at the 0.001 level (two-tailed).

Fig. 6 shows the path estimates based on both Model 1 and Model 2.

Figure 6

The first research hypothesis assumes that “brand experience (BEX) has a direct and a positive influence on emotional confidence (BEC)”. In the case of both Model 1 and Model 2, the results exhibited that (BEX) had a significant impact on the (BEC). Statistically, the influence of BEX on the BEC was found to be significant at p ≤ 0.001 (M1: γ = 0.786, t = 6.890; M2: γ = 0.797, t = 6.955).

The second research hypothesis assumes that “emotional confidence (BEC) has a direct and positive influence on consumer brand satisfaction (BPQ)”. The results exhibited that BEC had a significant impact on the BPQ. Statistically, the influence of BEC on the BPQ was found to be significant p ≤ 0.001 (M1: γ = 0.689, t = 9.947; M2: γ = 0.647, t = 9.340).

The third research hypothesis assumes that “Consumer brand satisfaction (BPQ) has a direct and positive influence on consumer brand loyalty (BL)”. The results exhibited that BPQ had a significant impact on the BL. Statistically, the influence of BPQ on the BL was found to be significant at p ≤ 0.001 (M1: γ = 0.630, t = 7.572, M2: γ = 0.258, t = 3.216).

However, the fourth research hypothesis intentionally included in Model 2 assumes that “emotional confidence (BPQ) has a direct and positive influence on consumer brand loyalty (BL)”. The influence of EC on BL was found to be significant at p ≤ 0.001 (M2: γ = 0.496, t = 5.857).

Table 10 shows the elicited clusters. Four clusters were extracted based on brand experience dimensions using hierarchical cluster analysis. Fig. 7 shows a graphical representation of the brand experience dimensions along the elicited clusters. Cluster 1 shows above-average scores on all the dimensions, whereas Cluster 2 shows below average on all the dimensions. Cluster 3 indicates the highest intellectual aspect with relatively lower, but above average sensory experience; in this cluster behavior, the experience is the lowest with relatively higher but below average emotional experience. Cluster 4 shows that behavioral, sensory and emotional experience is above average with below average intellectual experience. The primary objective of these clustering was to observe how the clustering of brand experience dimensions depicts the relationship among brand loyalty, brand satisfaction and emotional confidence.

Figure 7

Fig. 8 shows scores of emotional confidence, brand loyalty and brand satisfaction. As expected, the above-average scores of all the brand experience dimensions depict higher brand loyalty, and emotional confidence and brand satisfaction are slightly below average. Similarly, that was the case in the second cluster blow average brand experience pulled the brand loyalty and emotional confidence and brand satisfaction downwards; however, correspondents of the brand satisfaction was less. Interestingly, in the third cluster, higher intellectual and sensory experience pulled the brand loyalty and emotional confidence upwards but that was not the case with brand satisfaction.

Figure 8

This study affirms that brand experience in the context of car buying influenced emotional confidence, and there was a positive and statistically significant influence of brand experience on emotional confidence. The results show that within the framework, emotional confidence plays an essential role as there is a positive and statistically significant influence on brand satisfaction. The inclusion of a direct path in the model increases the relative model fit (SEM) as the value of chi-square further reduced. The affirmation of the hypotheses paved the way for further analysis, the use of structural equation modeling just showed the holistic impact of brand experience on emotional confidence. The relatedness of the sub-dimensions (i.e., sensory, affective, intellectual and behavioral) was not fully explored. The relationship between emotional confidence, brand satisfaction, and brand loyalty was further analyzed through clustering of the brand experience sub-dimensions. Four clusters were elicited (see Fig. 7).

In Cluster 1, all BE dimensions were above average, whereas in Cluster 2 all BE dimensions were below average, both clusters were pretty distinct. However, Cluster 3 and Cluster 4 showed some exciting patterns and the main contrast was based on intellectual experience. In Cluster 3, intellectual experience was highest and sensory experience was above average, whereas behavioral experience was lowest and affective experience was below average. In Cluster 4, intellectual experience was most depressed and behavioral experience was highest, and both affective and sensory experiences were below average. The relationships between EC, BS and BL within the elicited clusters were observed (see Fig. 8).

The observation suggests that higher BE (Cluster 1) depicts higher EC, BS and BL. Similarly, lower BE (Cluster 2) suggests lower EC and BL; the BS satisfaction dipped down but not as much as the EC and BL. The observation of Cluster 3 was interesting as both EC and BL was the highest, and BS was below average. Contrary to the assumption that higher affective experience should translate into higher EC and BL. Similarly, in Cluster 4, when intellectual experience was below average, the EC and BL dipped; it seems like the intellectual experience drives the EC and BL.

The use of judgmental sampling reduces the scope and generalizability of the study. Sample size (n = 240) for performing cluster analysis is not sufficient as more clusters can be observed with higher sample size. However, this study shows some interesting pattern that can further either be replicated or refuted with a higher sample size.