Assessing Interactions Among Omnichannel Attributes, Customer Perceptions, Customer Experience, Channel Selection

Abstract

Purpose: This study aims at understanding the impacts of three omnichannel attributes (channel transparency, channel uniformity, channel convenience) and four customer perceptions (perceived innovativeness, perceived personalization, perceived risk, perceived credibility) on customer experience and channel selection decision. Research design and methodology: A quantitative online survey with 356 shoppers was executed. The partial least squares linear structural model (PLS-SEM) and Smart PLS were adopted to analyze the collected data and test the proposed hypotheses. Results: The research findings indicate four dominant results: (i) The customers' channel selection is directly determined by customer experience; perceived innovativeness; perceived personalization; perceived risk; and perceived credibility; and (ii) among these, the perceived risk shows negative impact on the customer's experience and customers' channel selection whereas others reveal the positive status; (iii) The customer experience represents the most decisive impact on the channel selection, then perceived personalization, perceived credibility, perceived innovativeness, and perceived risk. (iv) Three proposed channel attributes (transparency, uniformity, convenience) significantly influence the overall customer experience. Conclusions: This research adds to the body of knowledge in omnichannel retailing, customer experience, and customer channel selection. Furthermore, this research provides omnichannel retailers with practical implications for improving customer channel selection.

1.Introduction

Bezes (2019) stated that retailing sector had undergone significant changes since the beginning of the 21st century in which traditional retailers shifted from single-channel, multi-channel towards the omnichannel model. Initially, multi-channel retailers manage channels separately (Nеslin & Shankar, 2009), resulting in severe data discrepancy and, especially, the customer experience throughout the shopping journey. Therefore, omnichannel is a contemporary method to enhance customer’s shopping experience and overcome all shortcomings of the multi-channel approach – organization touchpoints (Li, Liu, Lim, Goh, Yang, & Lee, 2018) and promote consistency in product/service offerings (Shen, Li, Sun, & Wang, 2018). Omnichannel integrates customer experience and focuses all purchaser interactions with the enterprises via the shoppers’ perspective (Yrjölä, Spence, & Saarijärvi, 2018).

According to Verhoef, Kannan, and Inman (2015), omnichannel management is the synergetic management of the multiple available channels and customer touchpoints to maximize the customer experience across channels and channel performance. The successful integration and coordination between sales channels shall better meet customers’ needs. From there, businesses can improve their financial performance (Hübnеr, Wollenburg, & Holzapfel, 2016). In response to retailing evolution, retailers must adapt to the omnichannel strategy. Specifically, the goal of this strategy is to maximize the overall retail experience across all channels and total sales through the integration of all traditional and online channels (Verhoef et al., 2015). In addition, retailers must be flexible and agile in their ability to change the way orders are fulfilled to ensure cost effectiveness (Ishfaq, Gibson, & Defee, 2016). The complex interactions in the retail supply chain present challenges to ensuring customer familiarity and comfort with the entire shopping process (Verhoef et al., 2015). Now, customers will needto decide which products and retailers to choose in a traditional shopping environment and which channel to choose in an omnichannel shopping environment. However, very few studies examine customer perceptions of the omnichannel method and how this affects customer’s channel selection (Bilgicer, Jedidi, Lehmann, & Neslin, 2015; Vеrhоеf et al., 2015; Yе, Lau, & Teo, 2018).

Given the above gaps in the literature, this study aims to devote to the literature of customer’s channel selection and the customer experience in omnichannel retailing context by exploring the impact of channel attributes and customer perceptions on the customer experience and, subsequently, on the customer’s channel selection. There have been threeomnichannel attributes, including Transparency, Convenience, Uniformity, and four customer perceptions: Perceived Innovativeness, Perceived Personalization, Perceived Credibility, and Perceived Risk.

2. Literature Review

According to Holbrook and Hirschman (1982), experienceis defined as the totality of events that a person experiences, usually affecting emotions and feelings when there is interaction through the stimulation of goods and services consumed. From a marketing point of view, the customer experience was proposed as the interaction between an organization and a unique customer (different individuals will not have the same experience), which can be remembered as a memorable event and enduring over time. Customer experience with a retailer was interpreted as an intrinsic and subjective response to interacting directly or indirectly with a company. More recently, Lemon and Verhoef (2016) describe customer experienceas a multidimensional construct that focuses on customers’cognitive, behavioral, emotional, and socialresponses to a company’s products throughout the entire customer journey. Although evaluated from different perspectives, customer experience is an overall concept, and there are certain commonalities between different definitions. The common point of the concepts is that customer experience is often theorized as a psychological construct and considered a subjective variable arising from feelings and comparing what customers receive with what they expect to receive (Suchánek & Králová, 2018).

Recent research on omnichannel retailing has focused on the dynamics that influence consumers’ channel choices. Keen, Wetzels, Ruyter, andFeinberg (2004) analyzed the consumer decision-making process to know how product price, retail form, and degree of control affect channel selection behavior among several channels (e.g., online, brick-and-mortar). Ansari, Mela, and Neslin (2008) suggested that sociodemographic characteristics and consumer experience (previous channel experience, number of previous purchases, interval between last two last purchases) can influence a consumer’s channel choice. In addition, Xu and Jackson (2019) investigated customers’ channel selection intentions in the omnichannel retail environment by analyzing the impact of channel attributes (Transparency, Convenience, Uniformity) on customers’ perception.

Channel transparency is expressed through various forms in the channel environment, including product information and order tracking capabilities (Xu &Jackson, 2019) and service availability information (Lee, Chan, Chong, & Thadani, 2019). Once a retailer cannot transparently disclose information and services on its sales channels, customers will face many difficulties in the purchasing process, negatively affecting their experience with the retailer (Bitner, Ostrom, & Meuter, 2002). Besides, channel consistency among stakeholders is reflected in product and service information across the entire channel (Lee et al., 2019), responsibility, and ability to communicate with the seller during the sales process (Xu &Jackson, 2019). Seck and Philippe (2013) suggest that channel consistency positively affects customer satisfaction and experience. In addition, the consistency of the channel also improves perceived service quality and minimizes customers’ perceived risk. Finally, Aagja, Mammen, and Saraswat (2011) found that the higher the channel convenience, the greater the influence on the customer experience. In addition, convenience positively impacts customer satisfaction and repeat purchase behavior (Seiders, Voss, Godfrey, & Grewal, 2007). Thus, a retailer’s channel possessing the three above attributes will limit the disclosure of shopper information to third parties, help buyers save monetary and non-monetary costs (time, effort), and improve customer comfort when using the channel. Thereby promoting a positive purchasing experience of consumers on the channel. Therefore, this study proposes hypotheses as follows:

H1:Channel transparency positively affects Omnichannel customer experience

H2: Channel uniformity positively affects Omnichannel customer experience

H3:Channel convenience positively affects Omnichannel customer experience

Alpert (2015) investigated the impact of consumers’ perceived innovativenesson various consumer goods. He examined perceived innovativeness’impact on satisfaction levels, and their results showed that greater awareness of technology novelty increases customers’ satisfaction experience. For perceived personalization, personalization helps retailers better meetthe increasingly diverse needs of customers, which has a positive impact on their experience (Lemke, Clark, & Wilson, 2011). McLean, Al-Nabhani and Wilson (2018) proposed the customization element that directly affects the customer experience in retail mobile applications. Also, on the integrated channel retail environment, Tyrväinen, Karjaluoto, andSaarijärvi (2020) confirmed the positive influence of personalization on the components of the buyer experience, including emotional experience experiential perception. In addition, personalization increases the customer's sense of control and makes them part of the experience creation (Chang, Yuan, & Hsu, 2010). Besides, perceived credibilityplays a vital role in shaping customer experience towards a retailer, driving repurchase intention from that retailer. Higher perceived credibility has a more substantial impact on customer experience and purchase channel selection intention through the perception of high quality, low risk, and information cost savings (Baek &King, 2011).

Contrary to the above factors, perceived risk is an essential factor in hindering the formation of a positive customer experience and negatively influences consumer purchasing behavior in the retail sector. Chang, Chih, Liou, andYang (2016) concluded that online shoppers’ perceived risk significantly negatively influences their experience and purchase decision. The above conclusion is also accurate in e-commerce when Kim, Ferrin, andRao (2008) confirmed that perceived risk harms users’ purchase intention. Similarly, Nok, Suntikul, Agyeiwaah, andTolkach (2017) show a negative relationship between perceived risk and purchase intention. Therefore, this study proposes hypotheses as follows:

H4:Perceived innovativeness positively affects Omnichannel customer experience

H5:Perceived personalization positively affects Omnichannel customer experience

H6:Perceived credibility positively affects Omnichannel customer experience

H7:Perceived risk negatively affects Omnichannel customer experience

Like the customer experience, customer channel selection is also influenced by four customer perception factors. According to Erdem andSwait (2004), perceived credibility is significantly related to emotion and reason in the customer decision-making process, therefore, has a positive influence on consumers’ future channel choices and considerations. The channel with a high degree of credibility will ensure a long-term plan to provide products and services to consumers while developing customer satisfaction, loyalty, and retailers' commitments. This continues to deliver positive word-of-mouth results (Ghorban & Tahernejad, 2012), helping to improve retailers' profits and competitiveness (Sallam, 2015; Al-Baz et al., 2018). For perceived innovativeness, Slade, Dwivedi, Piercy, and Williams (2015) developed a research model to determine the relationship between perceived innovativeness and consumer's intention in the context of online mobile payments. They argued that innovativeness positively influences intention to use remote mobile payments services. Besides, Bilgihan et al. (2016) concluded that recommendation systems with personalized features could attract customers to channels. A personalized purchase funnel reduces product searches and product review costs, thereby increasingthe chances of a buyer staying on the channel.

Moreover, it helps to minimize the customer's shopping time and effort (Kim &Baek, 2018). Perceived risk is still the only factor that negatively impacts channel selection decisions. Chang et al. (2016) concluded that perceived risk significantly negatively influences satisfaction and purchase decision. The above conclusion is also correct in e commerce when Kim et al. (2008) confirmed that perceived risk harms consumers' purchase intention. When buyers perceive the risks on the purchasing channel, they may not be satisfied with their experience and hesitate to choose this channel for future transactions. Consequently, this study proposes the following hypotheses:

H8:Perceived innovativeness positively affects customer’s channel selection

H9:Perceived personalization positively affects customer’s channel selection

H10: Perceived credibility positively affects customer’s channel selection

H11:Perceived risk negatively affects customer’s channel selection

Figure 1: Proposed research model

Gounaris, Dimitriadis, and Stathakopoulos (2010) noted that customers would consider their previous retail purchase experience to decide whether to return or repurchase to the brick-and-mortar store. In the online retail environment, McLean et al. (2018) also assert that a complete customer experience will significantly increase the frequency of retailers using mobile applications. In the context of omnichannel, Shen, Li, Sun, and Wang (2018) suggest that customers’ previous experiences with specific shopping channels should be considered when assessing their omnichannel shopping intentions. Accordingly, we propose the following hypotheses:

H12:Omnichannel customer experience positively affects customer’s channel selection

3. Research Methods and Materials

The author surveyed 400 customers who already have shopping experience with omnichannel retailers to accomplish the research objectives. Moreover, to make the survey more accurate, we only focused on customers who bought electronic devices. In the actual market, electronic device retailers highly adopted the omnichannel distribution model.

An online survey with a structured questionnaire was conducted from October 2021 to December 2021. The entire questionnaire uses a 5-point Likert scale in ascending order of the respondent's level of agreement. First, the research team conducted a pilot-test interview with a small sample of respondents to check for the quality and validity of the questionnaire. Through receiving feedback and discussion, the research team made appropriate adjustments to develop the official scale and questionnaire, which have the most suitable level with Vietnam's actual business and culture context. Second, the questionnaire was delivered to 400 respondents by email social media accounts. At the end of the investigation, the questionnaires with errors such as missing value, repetitions, conflicts were deleted to ensure the research results were accurate. Finally, the study obtained 356 complete questionnaires (equivalent to the rate of 89.0%) to include in the subsequent analysis steps.

Collected data were cleaned and analyzed by Microsoft Excel for descriptive statistics purposes, Partial least squares path modeling (PLS-SEM), and Smart PLS 3.2.2 software is used to evaluate the scale and determine the importance of the factors test the hypotheses posed. PLS-SEM was adopted to support prediction models from empirical data when different measurement scales and small sample sizes are used in the research model (Birkinshaw & Morrison, 1995). The PLS-SEM analysis was executed to assess the measurement and structural models. The measurement model was appraised by examining the values of Cronbach Alpha, Internal composite reliability, convergent validity, and discriminant validity (Henseler, Ringle, & Sarstedt, 2009). The structural model was scrutinized both direct and indirect effects to test the proposed hypotheses through the values of path coefficients, R2, f2, Q2, and p-values.

All measured items followed preceding studies with several adjustments to suit the research context. The current study consists of nine multi-dimensional constructs. Three constructs belong to the channel attributions: channel transparency, channel uniformity, and convenience. The transparency and uniformity of channel measures were adopted from Lee, Chan, Chong, and Thadani (2019); and Xu and Jackson (2019). The items of channel convenience were inherited from Xu and Jackson (2019) and Yan, Chen, Zhou, and Fang (2020). Four customer perception attributes, including perceived innovativeness, perceived risk, perceived personalization, perceived credibility was, advanced from the scale of Lin (2016); Xu andJackson (2019); Chetioui, Benlafqih, andLebdaoui (2020); Hickman, Kharouf, andSekhon (2020); and Yan et al. (2020). The omnichannel customer experience was evolved from Le and Nguyen-Le (2020) and Nguyen (2021). The customer channel selection scale was adopted from Xu and Jackson (2019) andTruong (2020).

4. Results

When performing the descriptive statistical analysis procedure with the selected sample, the study obtained the results of sample structure distribution as follows:

Table 1: Sample demographic characteristics

4.1. Assessment of measurement models

To assess the measurement model, we first estimated the convergent validity by examining the outerloadings of each item and the Cronbach Alpha (CA), the composite reliability (CR), average variance extracted (AVE) of each construct. According to Ford andLarcker (1981), the AVE coefficient must be greater than .50 to confirm the convergence value. Theouter loadings of each item should exceed .70, and the CA of each scale is above 0.70 to achieve the significance level (Hair, Sarstedt, Matthews, & Ringle, 2016).

The channel attributes constructs would be more reliable after removing five items, including CT3, CU3, CU9, CC₆, and CC₇. The customer perception constructs would be more reliable when eliminating PI4, PI7, PR2, PR3, and PP3. Ten disqualified items possessed the outer loading values below the approved value of .70. All the remaining 51 items satisfied the levels of reliability. The nine constructs’ CA and CR values are more significant than .70, and AVE values greater than .50 indicate the consistency reliability and convergent validity.

Table 2:Consistency reliability & convergent validity

Fornell andLarcker (1981) suggested that the square root of the AVE of each variable should be greater than the correlation coefficients between the latent variables to achieve discriminant performance. As the results are shown in Table 3, the square root AVE of each variable (at the beginning of each column) is larger than the correlations between the latent variables (correlation coefficient is below the initial value in the column). Thus, we may conclude that the measurement model showed adequate discriminant validity. (Appendix 2)

4.2. Assessment of Structural models

Multicollinearity is a phenomenon where the independent variables are strongly correlated with each other. The model that occurs with multicollinearity will cause many indexes to be skewed, leading to the results of quantitative analysis no longer giving much meaning. Sarstedt, Hair, Cheah, Becker andRingle (2019) proposed that VIF indexes of 5 or more show a very high degree of multicollinearity, and below 3, there is no multicollinearity. The analysis indicated that the lowest VIF value is 1.100 and the highest is 2.137, all lower than 3. Thus, there is no crucial multicollinearity concern in the structural model. To assess the quality of the structural model, we used the Standardized Root Mean Square Residual (SRMR) value. Hu and Bentler (1999) consider a .08 or lower acceptable value. The analysis result of the model fit summary (Table 3) demonstrates the SRMR value of .056, indicating the model’s good fit for theory.

R2 is the primary way to measure the model’s predictive accuracy and represent the percentage of variance in the dependent variables as explained by the independent variables in the model. Three dimensions of channel attributes (transparency, uniformity, and convenience) and three dimensions of customer perception (perceived personalization, perceived risk, and perceived credibility) can be explained 60.7% of the variance of the customer experience. Four observed dimensions of customer perception and experience explained 61.4% of the customer’s channel selection variance. Q2 describes the model’s ability to predict the observed variables of a latent variable (Reinartz, Haenlein, & Henseler, 2009). If the value of Q2 obtained is more significant than .00, the model can predict a particular dependent variable (Hair, Sarstedt, Hopkins, & Kuppelwieser, 2014).

Hair et al. (2019) proposed thresholds to assess the predictability according to the Q2 index below 0.25 as low predictability, from .25 to less than .5 as medium predictability, and .5 or more as high predictability. Q2 values of CX is .390, and of CS is .408, indicating that the customer experience and channel choice demonstrate satisfactory predictive relevance.

Table 3: R2, Q2, SMRM

The consecutive criterion to assess the structural model is path coefficients (b values). The path coefficients (Gronemus, Hair, Crawford, Nyalwidhe, Cunnion, & Krishna, 2010) express the degree of shift in the dependent variable for each independent variable. The path coefficient value is suggested to be above .100. However, the relationship between Perceived innovation and customer experience represented the bof less than .100 (.094) so the study rejected this hypothesis. Table 4shows that the path coefficients for all relationships were statistically significant due to all p values < .05. Therefore, eleven over twelve proposed hypotheses were supported.

Table 4: Hypotheses testing

The results of table 4 indicate that there have been nine positive relationships and two adverse ones among twelve proposed interactions. The perceived risk shows both negative impacts on customer experience and channel selection. In addition, Cohen (1988) proposed the f2 index level to assess the importance of independent variables according to the following levels .02, .15, and .35 indicating small, medium, and high effects. The results of Table 4 illustrated the two highest impacts of perceived credibility and perceived personalization on customer experience as f2 =.289 and f2=.282. In comparison, perceived risk moderately affects customer experience as f2=.054. Three channel attributes have medium influences on customer experience as f2 values range from .027 to .072. Customer experience shows the highest degree of influence for customer's channel selection as f2=.145 whereas four customer perceptions showed slight to medium effects due to the f2 values ranging from 0.024 to 0.170.

Figure 2: PLS results

5. Discussions and Implications

Channel attributes with three dimensions and customer perception with three dimensions were scrutinized as the propulsive factors of customer experience. Three sub dimensions of channel attributes, comprising transparency, uniformity, and convenience, were approved to affect omnichannel customer experience positively. While perceived personalization and perceived credibility of customer perception have an intense and positive impact on customer experience, perceived risk harms this dependent variable. These findings are supported by the studies of Sunikka and Bragge (2012); Seck and Philippe (2013); Bilgihan, Kandampully, andZhang (2016); Chang, Chih, Liou, andYang (2016); Choi, Kwon, and Shin (2017); Oppong (2020). Among the six examined components, perceived personalization has the most significant impact on customer experience, followed by perceived credibility.Two findings suggest that to advance positive customer experiences, retailers need to increase trustworthiness and develop personalization features across their sales channels.

Moreover, the customer perception, which includes perceived innovativeness, perceived personalization, perceived risk, and perceived credibility, together with customer experience, positively affect the customer's channel selection. These findings support the studies of Murali, Pugazhendhi, and Muralidharan (2016); and Mahmoud, Hinson, andAdika (2018). Likewise, perceived personalization and credibility are two factors that have the most profound effect on customer channel selection. These findings suggest that consumers are interested in personalization capabilities and high-demand credibility from the retailer's channel. A reliable and personalized channel help customer reduce search costs and possible risks and increase customer belief in the quality of products/ services and retailers. (Erdem & Swait, 1998).

Theoretically, this study supports the two aspects of omnichannel retailing, customer experience, and especially customer’s channel choice. First, although there have been several studies on customer experience, and customer’s channel selection in omnichannel retailing, these two variables are only individuals affected by either the channel's attributes or the customer’s perception. However, this study combined the two above groups of factors above and examined their sub – dimensions’ influences on customer experience to understand better the role of each factor group in enhancing overall customer experience and customer’s channel selection. Second, this study examines the relationship that previous models have overlooked –the relationship between the customer experience and channel selection decisions.

Practically, we consider that this study has several implications for management. First of all, omnichannel retailers should focus on enhancing perceived personalization on both online and offline channels to deliver a seamless shopping experience to their customers. Retailers can apply information technology to collect customer behavior across channels and make appropriate recommendations to buyers. Specifically, brick-and-mortar store salespeople could view recent customers' purchases and behavior by a tablet and their loyalty card or email address. As a result, salespeople will make tailored recommendations to in-store shoppers based on data across channels. Second, posting high-quality and consistent content will increase brandawareness and strengthen the retailer's business image with customers. Besides, content and action must be persistent across channels, improving perceived credibility and customer experience. Finally, omnichannel retailers should also pay attention to decreasing customers' perceived risk of omnichannel shopping. Guarantee policy and customers' personal information are two critical aspects retailers need be considered. To limit the risk of customers' personal information, retailers need to issue regulations to classify groups of information decentralize the use of information groups to ensure information security.

6. Conclusions and future research

The fact that more and more businesses are applying the omnichannel retail model has been attracting the research focus of scholars and professionals. Channel selection will help businesses improve sales strategies, increase customer experience and retain loyal customers with companies. This study advances the literature on omnichannel retailing in theoretical and practical implications. However, this study stillhas several limitations, which suggest directions for further research. First, this study mainly depended on the quantitative survey method with self-reported data from the methodology perspective.

Future studies are suggested using other methods such as field experiments, data mining, or qualitative interviews to improve the validity of the proposed research model. Secondly, implemented in Vietnam and mainly focused on the consumer electronics retailing sector, the research result maybe not be generalized to other contexts. Hence, future studies could assess this theoretical framework concerning other industries. Third, this research was taken in Vietnam and did not mention significant cultural or demographic differences to anticipate customer behaviors. Future studies, thus, are suggested to assess the effects of cultural or demographic factors in its research to enhance the generalization of research results.