Theoretical foundation

Logistics service quality

Logistics service quality refers to the extent to which logistics services meet or exceed the expectations of customers. Recently, scholars studied LSQ extensively. The term “logistics service quality” was first introduced and explored by Mentzer et al. [17]. They argued that customers’ perceptions of LSQ include not only physical distribution services but also other aspects that are of great importance to them. As many studies have verified, service performance is an important driver element in the logistics domain that creates value and helps obtain competitive advantages [8, 18, 19]. In most studies, LSQ is examined and estimated in various dimensions. The LSQ framework proposed by Mentzer et al. [8] includes nine service quality dimensions: PQ, IQ, ordering procedures, order quality, ordering release quantities, order discrepancy handling, order accuracy, timeliness, and order conditions. Furthermore, Mentzer et al. [19] explored logistics services in three dimensions: condition, timelessness, and availability. The viewpoint of Mentzer et al. [8] was also accepted and explored by Bienstock et al. [20] in their study on the SQ of physical distribution as a double-order construct consisting of timeliness, availability, and conditions. Previous findings also revealed that LSQ has a two-dimensional variable—operational and relational performance—in line with Stank et al. [18] and Daugherty et al. [21]. Rafele [22] proposed a framework for estimating logistics service performance, including logistics quality, in three dimensions: fulfillment methods, tangible components, and information actions. In line with Saura et al. [23], LSQ is mainly determined by empathy and reliability. Bouzaabia [24] studied 12 items of the LSQ from Mentzer et al. [19] and eliminated three items; the remaining nine items were structured by accounting for two dimensions: operational LSQ and relational LSQ. Gil-Saura et al. [25] proposed personal quality, information quality, order quality, punctuality, and order form [25]. In line with Gaudenzi [26], the key variables that characterize the wide LSQ concept have been discussed in seven dimensions: IQ, PQ, ordering procedures, order accuracy, order condition, timeliness, and order discrepancy handling. The perspectives of Gil-Saura et al. [25], Gaudenzi [26], and others were partially adopted in this study. Hafez et al. [27] explored six dimensions of LSQ such as information quality, product quality, product condition, delivery services, reverse logistics, customer services, while Revindran et al. [28] identified four dimension like product availability, Timeliness, product condition and reverse logistics. Choi et al., [29] proposed five dimensions like quality of delivery, quality of information, price of delivery quality of order, and customer service. Likewise, Rajendran et al. [30] suggested five dimensions such as Delivery services, reverse logistics, product quality, customer service and service recovery.

Owing to the rapid increase in B2C and other online commerce, a limitation due to insufficient LSQ is outstanding, whereby examining electronic LSQ is crucial in a B2C commerce environment [31]. Concerning electronic LSQ (e-LSQ), Rao et al. [32] introduced a measurement framework for e-LSQ in the B2C environment, which included order tracking, on-time delivery, item availability, and shipping options. However, in present research logistics service quality refers to the overall evaluation of the quality of logistics services provided by a company, which comprises five dimensions: personal contact quality, resource quality, operational quality, information quality, and customization quality (Fig 1). These dimensions collectively describe the attributes of logistics services that are crucial in meeting customers’ needs and expectations, such as the quality of communication, resources, operations, information, and the degree to which services are customized to meet individual customer requirements.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Conceptual framework of re-use intention.

https://doi.org/10.1371/journal.pone.0286382.g001

Customer satisfaction

CS is the emotional response of customers when they perceive a distinction between their past expectations and the real performance of a product or service that they have used [33]. It is considered an overall judgment of performance [34, 35]. CS in logistics industry context indicates that the customers’ delightfulness, well matching services with expectations, positive or better feelings towards LSP for choosing them. Many studies have shown that improvements in LSQ contribute to CS [21, 36], and ensuring CS with physical delivery services is a crucial objective for LSP. Extensive research has been conducted on CS in relation to logistics services [3, 19, 37]. Previous studies have considered CS to be a result of LSQ and investigated how LSQ affects CS [19, 38]. Stank et al. [18] found that relational performance has a considerable effect on CS, whereas operational performance has no such effect. Furthermore, optimizing LSQ can lead to improved CS [39]. As CS and SQ are interconnected concepts that are extensively studied in the LSQ literature, it is essential to identify which factors of LSQ have a significant impact on CS.

Numerous studies have confirmed the importance of LSQ for achieving CS. Rafiq and Jaafar [38] described the link between customer perception of LSQ and its impact on CS, considering functional measures, such as PQ, IQ, and ordering procedures. Most studies have verified the relationship between LSQ and CS in a B2B environment between industrial purchasers and vendors [17, 19, 40, 41]. Murfield et al. [31] studied the effect of LSQ on CS and customer loyalty in an omnichannel commerce environment and concluded that three components of LSQ—timeless, availability, and condition—have positive associations with the CS of logistics services. Studies in a B2C environment have confirmed a positive correlation between LSQ and CS in online retail [42]. Rao et al. [32] also found that item availability and online delivery are critical elements of electronic physical distribution service quality. However, most studies have focused on customer purchase satisfaction rather than CS with logistics services.

Re-use intention

The RI is an important method for measuring customer behavior in marketing studies. The business is committed to strengthening customer loyalty by increasing repurchases and re-use, preventing customer loss, and increasing the retention ratio by maintaining high customer loyalty. In the past few years, the intention of reusing has received a significant stream of attention from researchers. Furthermore, it has been extensively used in studies on customer loyalty. Different studies concerning different industries, such as food delivery [43], mobile bookstore applications [44], e-commerce [45], and smart tourism [46], have indicated that customer satisfaction affects re-use intention. Jen et al. [47] and de Oña et al. [48] specified that excellent CS is directly connected to customer RI. Furthermore, as Chou and Kim [49] and Jen et al. [47] concluded, customer satisfaction is an interlink between service quality and re-use intention.

For logistics service providers, the re-use intention of customers is very important from a long-standing perspective, which ensures that both receivers and providers of logistics services have a favorable relationship [50]. In line with the existing research, reserving existing customers in other domains is considerably more meaningful than achieving new customers. The intention to re-use a logistics service can be defined as the customer’s belief and willingness to make up a product and express it as a specific future behavior i.e., willing to use or recommend other about the logistics services or service providers in future when needed to use. To some extent, logistics service re-use intention is critical to the success of sellers, not only from the viewpoint of the sellers but also from that of the buyers. Through an online investigation of online sellers, Ngah et al. [45] concluded that satisfaction significantly relates to the re-use intention of certain third-party logistics services and that the satisfaction index was a mediator between reliability and RI. At present, there is limited literature on the logistics service re-use intentions from the perspective of online shopping customers.

Hypotheses development

Following Mentzer et al. [8], nine constructs have been proposed, including information quality and ordering procedures. CS and quality of service are related concepts that have been widely studied by various scholars. Mentzer et al. [19] insisted that CS was a consequence of the LSQ, time of delivery, and settlement of orders’ puzzles and problems, significantly reducing satisfaction. Studies have stated that operational and relational performances related to logistics services have significant and positive relationships with CS [21]. According to Ramanathan [51], the primary cause of customer dissatisfaction stems from either late or non-delivery of products, inaccurate orders, or damaged products. Yeo et al. [52] and Gupta et al. [11] showed a significant link between RQ and CS logistics. Based on this reasoning, the following hypotheses (H₁, H₂) are proposed:

1. H₁: OQ positively influence CS.
2. H₂: RQ positively influences CS.

There is insufficient personal contact between logistics service providers and buyers; therefore, information delivered by logistics service providers through the web or apps is especially important. When logistics service providers provide high-quality information to their customers, they enhance the customers’ overall experience, leading to greater CS. For example, when customers receive authentic and upgraded information about the status of their shipments, they feel more informed and in control of the delivery process [10, 53]. This can help reduce uncertainty and anxiety, which, in turn, leads to greater satisfaction with logistics services.

Similarly, when logistics service providers provide complete information about the range of services they offer, they help customers make informed decisions about which services to use [3]. This can lead to greater customer satisfaction as they can select the services that best meet their needs. Additionally, relevant information tailored to customer-specific requirements can lead to greater satisfaction [10]. For instance, providing customers with information about the expected delivery time, delivery cost, and options for rerouting or rescheduling their shipments can help them meet their individual needs and preferences, ultimately leading to a more satisfying customer experience. Thus, high-quality information provided by logistics service providers can lead to greater customer satisfaction by reducing uncertainty, helping customers make informed decisions, and meeting their individual needs and preferences. Thus, we propose the following hypothesis:

1. H₃: IQ positively influences the CS.

Bouzaabia [24] compared the perceptions of retail logistics service quality among Romanian and Tunisian customers and found that relational logistics service quality was the most vital interpreter of satisfaction. According to Uvet’s [9] findings on the correlation between consumers and employees in their awareness of SQ in logistics, it was determined that CS with logistics services is also influenced by the behavior and attentiveness of LSP staff. And they specifically found that PQ has a significant impact on CS. Similarly, Gupta et al. [11] revealed that personal contact quality affects CS positively in the case of logistics services. Based on these arguments, this study hypothesizes the following:

1. H₄: PQ positively influences CS.

Coelho and Henseler [54] declared that customization strengthens service quality, perceived customer trust, CS, and customer loyalty toward service providers. For logistics service providers, customer-specific or customized requirements are important for satisfaction and establishing a close relationship. In line with Hu [55], customized logistics services are a predictor of CS. Gupta et al. [11] discovered that the customization and innovation have a positive effect on CS for logistics services. Thus, the following hypothesis was derived:

1. H₅: CQ positively affects CS with logistics service.

In line with Oliver [33], CS impacts customer attitudes and leads to a significant positive effect on potential behaviors in the future, such as RI. The customer intention to re-purchase (re-use) products (services) covers specific and cautious attitudes because of experiences, particularly when customers concern a specific product or service [35]. Positive attitudes originate from satisfaction with products or services, and later impact intentions [56–58]. According to Wang et al. [59] and Alalwan [60], CS has a significant impact on RI in service industries, such as mobile food ordering online [60] and urban rail transit [59]. In the Malaysian context, Ngah et al. [45] found that satisfaction positively influences the intention to reuse third-party logistics services. Especially for online shopping linked to physical distribution, customers are concerned not only with online apps or websites but also with logistics services to complete purchases. Based on the above discussion, to test the effect of CS on the RI of logistics services, the following hypothesis was proposed:

1. H₆: CS positively affects logistics service RI.

According to the SOR model, the two dimensions of customer emotion and customer cognition are included in the organism variables, whereas customer satisfaction is included in customer emotion. As a response variable, research on customer behaviors in the theoretical area generally revolves around two paths: first, as per the Mehurabian-Russell (M-R) framework, stressing the intermediary role of emotion in environmental stimulus and customer response [61]; second, using perceived service quality as the intermediary variable, emphasizing that customers generate self-cognition through external stimulation, which then affects purchase intention [62]. In Zajonc’s [63] opinion, the two paths should be considered together; that is, both the cognitive and emotional responses should be studied.

Based on the M-R model, emotions play a mediating role in environmental stimuli and customer responses. This has been previously verified by several researchers. In line with Dabholkar et al. [64], elements relating to the quality of service are better designed as antecedents than their components, and CS is a significant mediator between SQ and behavioral intentions, such as the intention to re-use or re-purchase. In line with Yilmaz et al. [65] and Wang [59], positive CS created by the quality of service provided by the metro operator led to re-use intention. In line with Preacher et al. [66], a mediating effect occurs when a predicting variable indirectly affects a dependent variable through at least one intervening variable or mediator. Based on the previous literature reviews, this study concludes that customer satisfaction as a mediating variable significantly affects customers’ willingness to re-use logistics services. Thus, the following hypotheses were proposed:

1. H_M1: CS is a significant mediator within the OQ and logistics service RI relationship.
2. H_M2: CS is a significant mediator within RQ and logistics service RI relationship.
3. H_M3: CS is a significant mediator within IQ and logistics service RI relationship.
4. H_M4: CS is a significant mediator within PQ and logistics service RI relationship.
5. H_M5: CS is a significant mediator within CQ and logistics service RI relationship.

All association hypothesized above, presented in Fig 1 below.

Population and sample

The targeted respondents of this study were Chinese customers older than 18 years, representing approximately 65% of the Chinese population. The minimum sample size in this study was calculated using G*Power software. With an effect size (f²) of 0.15, a power of 0.95, and six exogenous variables, the minimum sample size required was 146 [67]. A sample size of 810, obtained from the distribution of 832 questionnaires, resulting in a response rate of 95.08 percent, was deemed sufficient for this study. Consequently, the subsequent investigation was considered valid. Our continuous effort to remind the participants to complete the questionnaire was the reason for the high response rates. The human research ethics committee of Nantong Institute of Technology approved this study (Reference number: BS-NIT-2023-0402). This study has been performed in accordance with the Declaration of Helsinki. This study collected written informed consent from all respondents. This study collected and reported no identifiable information. The information was gathered between November 15 and December 22, 2022.

Survey instrument

The questionnaire comprised two parts. The first was the demographic profile of the participants, such as educational level, age, average monthly income, gender, marital status, employment status, online shopping expense cost, and location. In addition, the interviewees were required to reply about the frequency with which they purchased products online and used logistics services each month. The second section focused on the exogenous and endogenous constructs of logistics services, including OQ, RQ, IQ, PQ, CQ, CS, and RI. All 36 measurement items of the above constructs were rearranged in accordance with previous studies. Five items of operational quality were adopted from Bouzaabia et al. [24]; Juga et al. [5]; and Gronroos [68]. Following the research of Gupta et al. [11]; Yeo et al. [52]; Thai [69], five items of resource quality were chosen. The five items of information quality were extracted from the study of Thai [70]; Huang et al. [71]; Saura et al. [23]; and Rafiq and Jaafar [38]. The study picked 5 items for personal contract quality from the studies of Bienstock et al. [20]; Saura et al. [23]; Rafiq and Jaafar [38]; and Mentzer et al. [19], while the five items of customization quality were adapted from the study of Vu et al. [4]; Hu et al. [55]; and Coelho and Henseler [54]. The study of Kwak et al. [72] was followed for the items (five) of reuse intention construct, whereas the study of Ali et al. [37]; Vu et al. [4]; Klaus and Maklan [73]; Dagger et al. [74]; Mentzer et al. [19] were consulted for the six items of customer satisfaction. The respondents were suggested at the beginning of the questionnaire to fill in the questionnaire according to their latest or most impressive service experience from logistics providers. To measure all indicators of the variables, we used a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). Full information on the questionnaire used to measure all variables is presented in Survey Questionnaire in S1 File.” The questionnaire was developed in English and later translated into Chinese to inquire about adult Chinese customers. To ensure that the responses were not distorted or misunderstood, professional translators were invited to assist in the translation work.

Data collection method

The online questionnaire was uploaded at the link “https://www.wjx.cn/vm/eBmy8hA.aspx#.”. A convenience sampling strategy was adopted to obtain adequate responses to the questionnaire via both WeChat and QQ software from November 12, 2022, to November 23, 2022, and a qualifying question was used to filter the potential respondents. Consequently, 832 responses were successfully obtained and 21 (2.52%) respondents chose “never use the service from the logistics service providers,” which was set aside. Consequently, 810 valid samples were retained for subsequent analysis. Informed consent was obtained from all participants using a questionnaire for ethical approval. To ensure the accuracy and efficiency of data collection during the survey, we conducted a pilot test. We closely monitored the distribution and completion of questionnaires to ensure their validity and organization. Our pre-test sample consisted of 35 participants, from whom we collected the completed questionnaires. The results from the pre-test confirmed the initial validity and reliability of the survey items. Complete data presented in Dataset in S2 File.

Common method bias

In this study, a self-administered questionnaire was adopted for data collection, and a structural model was used to assess the hypothesis. As a result, common method bias (CMB) could occur. Hence, both procedural and statistical remedies were used to avoid CMB [75]. In case of procedural measures, terminology definitions were clearly described and provided clear and concise instructions for respondents to follow when completing surveys in the questionnaire. This is helpful to reduce confusion and errors that can lead to bias in the data. Also, we told the people who answered them that there were no absolute right or wrong response and that their answers would be assessed anonymously. This can help to reduce the potential for social desirability bias and CMB that can arise when respondents feel pressured to provide certain answers. Besides, we randomized the order of questions to reduce the potential for order effects or other biases that can arise when questions are presented in a certain order. Regarding the statistical measures, first, the study conducted a Harman’s Single factor test and revealed that a single factor explained 39.44% of the overall variance, which is less than 50%, and CMB did not exist [76]. Second, the CMB was evaluated using Kock’s [77] full collinearity test. All the latent variables in this study were regressed on a commonly created variable. The results of the full collinearity test and variance-inflated factor (VIF) values for OQ (1.378), RQ (1.727), IQ (1.821), PQ (1.854), CQ (1.888), CS (2.065), and RI (2.634) were all below 3.3, which are shown in Table 1, suggesting that the model did not meet the problem of CMB [77].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Full collinearity test.

https://doi.org/10.1371/journal.pone.0286382.t001

Multivariate normality

Conducting proper data analysis by checking for multivariate normality is vital. In this study, multivariate normality was estimated using the Web Power online tool [78]. Results of the multivariate normality test showed that the p-values for Mardia’s multivariate skewness and kurtosis were below 0.05, which certified the issue of non-normality [79].

Data analysis method

The analysis was conducted in three stages. First, we quantified the validity and reliability of the measurement model. In the next phase, structural equation modeling (SEM) was used to elaborate the connection between the predictor and latent variables, including mediation and moderation effects. It is widely accepted that SEM provides better estimates than regression when executing mediation and moderation [80]. Therefore, this study applied SEM, specifically partial least squares-based SEM (PLS-SEM), using Smart-PLS 4.0. PLS-SEM was considered the best choice because of its effectiveness in evaluating complex frameworks that involve moderation effects [81]. Finally, to ensure the robustness of the results, a necessary condition analysis (NCA) using the bottleneck analysis approach was performed to achieve a comprehensive understanding of the relationships among the constructs. Necessity logic is critical for achieving a specific outcome, and its absence guarantees failure [82]. NCA is a statistical analysis approach that identifies single necessary causes or independent variables to achieve a target outcome [82]. Combining PLS-SEM and NCA has been suggested to further theorize and validate the theory (Richter et al., 2020). This technique is used as a complement to the existing conventional techniques i.e., here PLS-SEM identifies the factors that provide the best possible outcome while, NCA explore for necessity logic, and provides us necessary conditions to be meet for a particular outcome. This combined use leverage researchers to improve the accuracy and efficiency of the PLS analysis, as it reduces the number of variables that need to be included in the model.

Demographic characteristics of respondents

This study obtained valid results from 810 respondents. Table 2 presents the participants’ demographic profiles. Especially, the respondents were approximately equally split between males (49.8%) and females (50.2%). Besides that, 35.9% of participants were aged between 36 and 45 years, followed by participants aged 26–35 years (25.9%), 18–25 years (18.8%), 46–55 years (13.5%), and 56–65 years (4.7%). The remaining 1.2% of respondents were over 65 years old. Of the respondents, 33.5% were from East China, followed by North China (17.5%), Central China (15.7%), south China (11.9%), South China (11.9%), Northwest China (5.4%), South China (2.7%), and others (1.5%). In terms of monthly shopping frequency, 38.4% reported shopping 1–2 times per month, followed by 37.5% shopping 3–4 times per month, 13.5% shopping 6–10 times per month, and 9.3% shopping more than ten times per month, while only 1.4% reported rare shopping. The participants who used the logistics service 3–5 times were 42.7%, followed by 1–2 times (36.4%), and 6–10 times (12.1%). Of the respondents, 79.2% worked, and the remainder (20.9%) did not work. Of the respondents, 66.1% had a monthly income below RMB 4500, while the remaining 33.9% had an income above this threshold. Similarly, 85.8% of participants spent less than RMB 4500, while the remaining 14.2% spent more than that.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Demographic characteristics (N = 810).

https://doi.org/10.1371/journal.pone.0286382.t002

Reliability and validity

To confirm the construct reliability utilized in this study, internal consistency reliability was evaluated by computing both Cronbach’s alpha and composite reliability (rho_a and rho_c) (as presented in Table 3). According to Hair et al. [81], the composite reliability values (rho_c) for all constructs met the recommended range of 0.70–0.90. Additionally, the composite reliability values (rho_a) for all variables were within the lower and upper bounds of Cronbach’s alpha and composite reliability (rho_c), thus confirming a high level of internal consistency among the constructs. In addition, the average variance extracted (AVE) for all items for each variable was higher than 0.50, establishing adequate convergent validity to affirm the unidimensionality of each variable [83]. Therefore, we found sufficient reliability in the constructs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Reliability and validity.

https://doi.org/10.1371/journal.pone.0286382.t003

Once reliability was confirmed, the discriminant validity of the model was assessed by applying both Fornell and Lacker’s criteria and the heterotrait–monotrait ratio (see Dataset in S2 File. Loading, Cross-Loading, and the Fornell-Larcker criterion; and Fig 2). The discriminant validity of the framework was evaluated using Fornell and Larcker’s criterion, and the results satisfied this criterion, as presented in Table 4. According to the analysis outcome, the model displays discriminant validity, which is evident from the higher value of the square root of the AVE in the diagonal compared to the off-diagonal constructs [84]. This finding suggests that the construct measures in the model are more closely related to their respective variables than to other variables, indicating a good level of discriminant validity. According to Henseler et al. [85], an optimum heterotrait-monotrait ratio of correlations (HTMT) below 0.90 indicates high levels of discriminant validity (Table 4). Although the value below 0.85 is sufficient, we used higher threshold value so that it ensures that the constructs are sufficiently distinct and do not overlap in terms of the underlying constructs they are measuring. Also, by using a higher threshold, researchers can be more confident that their measures are truly distinct and can provide more accurate and reliable results. In this study, all values in the HTMT matrix were below this threshold, confirming a greater extent of discriminant validity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Measurement model.

https://doi.org/10.1371/journal.pone.0286382.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Heterotrait-monotrait ratio (HTMT) matrix.

https://doi.org/10.1371/journal.pone.0286382.t004

To determine multicollinearity, this study quantified the variance inflation factors (VIF) for all measurement items. All VIF scores were below the recommended cut-off value of 5.0, indicating the absence of multicollinearity. Therefore, the framework used in this study does not suffer from multicollinearity issues. Finally, the findings suggest that the model used in the study is reliable and valid and that the factors are independent of each other.

Hypothesis testing

The determination coefficient R² showed the explanatory power of the prediction model equations in explaining the endogenous constructs. According to Cohen’s [86] recommendations, an endogenous construct’s R² value ranging from 0.02 to 0.12 is generally deemed weak, while a value between 0.13 and 0.25 is regarded as medium, and a value of 0.26 or higher is considered substantial. As shown in Fig 2, the proposed model showed substantial explanatory power for CS (0.473) and logistics RI (0.361).

For the direct hypothesis (Fig 2 and Table 5), OQ (beta = 0.126, p < 0.01), RQ (beta = 0.220, p < 0.01), IQ (beta = 0.241, p-value < 0.01), PQ (beta = 0.222, p-value < 0.01), and CQ (beta = 0.119, p < 0.01) have significant relationships with CS for logistics services. CS (beta = 0.601, p<0.01) is significantly connected to logistics service RI. Notably, the results for H₁, H₂, H₃, H₄, and H₅ indicated that R² was 0.473 and OQ, RQ, IQ, PQ, and CQ explained 47.3% of the variance in the CS. Meanwhile, the results of H₆ also revealed that CS explained 36.1% of the variance in logistics service RI, with beta = 0.601, p < 0.01, and R² (0.361). Therefore, H₁, H₂, H₃, H₄, H₅, and H₆, which were proposed to test the direct effect, were all supported at the 1% significance level.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Hypothesis testing.

https://doi.org/10.1371/journal.pone.0286382.t005

The study also explored the mediation hypothesis by applying bootstrapping to the indirect influence, in line with Preacher et al. (2008). The result (Table 5) reveals that CS contributed a significant and positive mediating role for the connection among OQ → CS → RI (β-value = 0.076, t-value = 4.592, p-value<0.01, CI Min = 0.543, CI Max = 0.657), RQ → CS → RI (β-value = 0.133, t-value = 4.762, p-value<0.01, CI Min = 0.088, CI Max = 0.180), IQ → CS → RI (β-value = 0.145, t-value = 5.186, p-value<0.01, CI Min = 0.099, CI Max = 0.191), PQ → CS → RI (β-value = 0.133, t-value = 4.581, p-value<0.01, CI Min = 0.087, CI Max = 0.182), and CQ → CS → RI (β-value = 0.072, t-value = 2.790, p-value<0.01, CI Min = 0.031, CI Max = 0.115). As a result, at the 1% significance level, the hypotheses H_M1, H_M2, H_M3, H_M4, and H_M5 that CS mediates the relationships of OQ, RQ, IQ, PQ, and CQ with RI were supported.

Necessary condition analysis.

NCA can identify whether the existence of a specific factor is a necessary condition for the emergence of a certain result, and it can also analyze the bottleneck extent of the necessary condition [82]. Fig 3 reveals the scatter plots for all connections, and Table 6 reveals the NCA effect size d value, which is a “quantitative reflection of the magnitude of some phenomenon that is used to address a question of interest” [87]. For general studies, 0<d<0.1 is a “small effect,” 0.1<d<0.3 is a “medium effect,” 0.3<d<0.5 is a “large effect,” and d≥0.5 is a “very large effect.” [82]. In this study (Table 6), the effect sizes were between 0.1 and 0.3, except for information quality, which had the lowest effect size (0.045). The effect sizes were deemed to be both theoretically and practically meaningful, as concluded by Van et al. [88]. As directed by Kelley & Preacher [87], with the effect size of 0.1 as the standard, X “information quality” is not necessary to Y “customer satisfaction.” The NCA’s results manifest that OQ, RQ, PQ, and CQ are meaningful (d≥0.1) and significant (p<0.05) necessary conditions for CS; CS is logical and significant necessary conditions for logistics service RI.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. NCA charts.

https://doi.org/10.1371/journal.pone.0286382.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. NCA effect size.

https://doi.org/10.1371/journal.pone.0286382.t006

Table 7 reports the bottleneck-level analysis results of the NCA. As can be seen, to achieve 50% CS, a 0.494% of RQ, 0.37% of IQ, 1.111% of PQ, and 1.111% of CQ are required, while there is no bottleneck level in OQ; to achieve 90% CS, all 5 components of LSQ are bottlenecks for CS. The study also investigated whether customer satisfaction is a necessary condition for re-use intentions. Table 7 highlights this while considering the relationship between CS and RI, to achieve 30% re-use intention, a 1.111% CS is required; to achieve 40% or above levels of re-use intention, a 1.235% CS is required.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Bottleneck percentages.

https://doi.org/10.1371/journal.pone.0286382.t007

This study explored how OQ, RQ, IQ, OQ, and CQ, constituting LSQ components, influence CS, how CS influences logistics service RI, and whether the CS of LSQ mediates the connection between the elements of LSQ and logistics service RI. In this study, five components of the LSQ—OQ, RQ, IQ, PQ, and CQ—were used to determine the influence of the LSQ on customer satisfaction. Overall, 11 hypotheses were proposed and all relationships were found to be significant in the empirical investigation. The explanatory power of the exogenous constructs on the endogenous construct was also substantial, indicating that the model fit the investigation well. The following is a specific discussion of these relationships.

Hypothesis 1 posited that the operational quality of LSQ positively influences satisfaction. The results of this study support this hypothesis. This outcome is supported by the previous research conducted by Sorkun et al. [9]. They studied LSQ on omnichannel distribution for logistics services and identified that OQ influences the overall experience of the customer, from the order placement process to the delivery of goods. The faster and more accurately the logistics service provider can deliver goods, the greater the possibility of gaining CS via the services. Hence, LSP must emphasize OQ to enhance CS and achieve business success. Similar findings were revealed for Hypothesis 2, for which the results of this study claimed that customers are likely to be more satisfied with LSPs that have modern, well-maintained equipment. This includes everything from trucks and ships to warehouses and equipment handling. LSPs must have well-trained, experienced, and customer-focused staff that can promptly respond to and effectively address customers’ concerns and needs. This includes everyone from sales representatives and customer service personnel to drivers and warehouse workers. This result is consistent with early studies by Yeo et al. [52] and Gupta et al. [11]. This means that when logistics service providers invest in the latest technology, equipment, and employees, they can help ensure that goods are handled efficiently and effectively, thus reducing the likelihood of delays and damage during transportation. This, in turn, helps enhance CS by providing higher-quality services.

This study revealed that IQ offers a significant connection with CS for logistics services (H3). The scholars [10] supported this finding. However, the effect size of IQ on CS was relatively small compared to the other stimulus (LSQ) factors. This suggests that while IQ is a factor that contributes to CS, it falls short of being a necessary condition for the emergence of CS. IQ refers to the timeliness, accuracy, completeness, relevance, and clarity of information provided to customers. Logistics services rely heavily on the information exchange and communication between service providers and customers, making information quality a crucial factor in CS.

The outcome of this study suggests that PQ has a significant connection with CS in logistics services, supporting the hypothesis that PCQ directly affects CS (H4). As PQ increased, customers were more likely to be satisfied with their received logistics services, which is in accordance with previous studies [9, 10] that highlighted the importance of human factors in service delivery. Uvet [9] studied logistics service quality in the USA based on student views and found that PQ is a significant predictor of CS. These findings indicate that in addition to LSQ, the quality of personal interactions between customers and LSPs is a necessary factor in determining CS.

This study also found that CQ has a significant positive effect on CS for logistics services (H5). This outcome is in line with earlier research [10, 55] that showed that customization is a crucial factor in satisfying customer needs and preferences. Customization quality involves meeting customers’ specific requirements and tailoring services to their unique needs. When logistics service providers provide customized services to customers, they can increase the likelihood of CS, leading to loyalty and repeat business. The positive influence of CQ on CS suggests that LSPs should focus on offering customized solutions to meet customer needs. In doing so, they can differentiate themselves from their competitors and increase their value propositions to customers.

In addition, the findings for hypothesis (H6) illustrated that CS with logistics services is significantly linked with logistics service RI. This finding implies that if customers are satisfied with the logistics services they receive, they are more likely to reuse them in the future. This corroborates previous research [45] showing that CS is a significant predictor of customer RI. The findings suggest that satisfied customers are more likely to reuse logistics services, which can lead to increased customer loyalty and profitability for logistics service providers. One of the key takeaways from the study is the distinction between loyalty and reuse intention. While the two concepts are related, they are not interchangeable. Loyalty refers to the long-term commitment of a customer to a brand or company, which is often influenced by factors such as trust, emotional attachment, and perceived value. Conversely, reuse intention refers to a customer’s intention to use a specific service again in the future, which may or may not be influenced by loyalty.

By mediating analysis, the research hypotheses (H_M1, H_M2, H_M3, H_M4, and H_M5) for testing the customer satisfaction of logistics services mediate the impact between the five components of LSQ and the re-use intention of logistics services. This study indicates that CS plays a significant role in mediating the link between operational quality, responsiveness, information quality, customization quality, personal contact quality, and logistics service RI. However, all the mediating relationships are partially mediated as all the relationships got significant. The positive and significant indirect effects of each of these quality dimensions and RI through CS suggest that enhancing customer satisfaction can lead to increased repeat business.