Service journey quality: conceptualization, measurement and customer outcomes

3.1 Theory-based view of service journey quality

The first step of the conceptualization process focuses on forming a theory-based view of the SJQ concept's key content. As our literature review demonstrates, research on service process quality and customer journeys can together offer tentative insights into SJQ. First, many studies stress the importance of designing the journey as a whole and integrating touchpoints such that the journey runs smoothly (e.g. Lemon and Verhoef, 2016; Homburg et al., 2017; Sousa and Voss, 2006). However, due to the increasing functional and operational touchpoint interdependencies of the service processes, mere excellence in individual interactions may not be enough for successful service delivery if the service delivery touchpoints do not align and work in concert (see Rawson et al., 2013).

Second, many authors highlight that touchpoints and their cues should be thematically consistent and coherent along the journey (e.g. Homburg et al., 2017; Kuehnl et al., 2019). Customers experience and evaluate service cues across touchpoints, and with an increasing number of channels and partners involved in service delivery, they may feel lost if the service touchpoints are very dissimilar (cf. Berry et al., 2006).

Third, research emphasizes the individual and subjective nature of journeys, suggesting that they should be sensitive to customers' individual needs and channel preferences (e.g. Barwitz and Maas, 2018; Patricio et al., 2011). High-quality service delivery should therefore adapt to the needs of the individual customer throughout the whole service process (cf. Dhebar, 2013). These three themes – touchpoint integration, thematical coherence and sensitivity to individual customer needs – were adopted as the initial theory-based view for SJQ conceptualization, which was refined and complemented with empirical insights from service-intensive contexts.

3.2 A qualitative study for establishing a field-based view of service journey quality

Next, we conducted a qualitative study to generate an understanding of SJQ by drawing from human experiences (Gioia et al., 2013). Following Sharma and Conduit (2016), we took the Gioia approach to first conduct open coding to identify themes emerging from data, to capture any SJQ aspects that are relevant for consumers' lived service experiences. In accordance with the abductive approach, we next integrated the first-order codes into second-order, theory-centric themes iterating between the data and the tentative conceptualizations derived from previous research (Dubois and Gadde, 2002). The qualitative data thus served the purpose of substantiating and refining the theory-based view and formulating service process specific definitions for the tentative themes.

3.3 Conceptualization: synthetizing theory-based view and empirical findings

By synthesizing previous research with findings from qualitative study, we define SJQ as the degree to which customers perceive the combination of provider-owned service process touchpoints functioning as a (1) seamless, (2) coherent and (3) personalized whole. With provider-owned touchpoints, we refer to touchpoints controllable by the service providers, i.e. brand- and partner-owned touchpoints (Lemon and Verhoef, 2016). The rationale is that while modern multichannel service processes may involve a network of service providers and outsourcing partners (Tax et al., 2013), the customer nevertheless evaluates the process as a whole and often sees each touchpoint representing, and remaining the responsibility of the focal service provider, whether outsourced or not (Kranzbühler et al., 2019). Thus, the SJQ construct focuses on service delivery touchpoints that service providers are able to design and manage (see Becker and Jaakkola, 2020). We identify three key SJQ dimensions: seamlessness, personalization and coherence discussed in detail below.

The first identified dimension we label journey seamlessness: the focus group participants described positive service journeys as simple and hassle-free, wherein moving from one encounter to another is easy – even between different types of journey touchpoints. Participants discussed many instances of positive and negative service experiences emerging from the efficiency of integration and information flow across organizational silos or service delivery partners that created feelings of a smooth and continuous service, or lack of it:

I hit a car at the parking lot and called my insurance company. Their call center connected my call to the insurance payout services and also transferred my details since the person picking up already knew who I was and why I was calling! She told me what would happen next and helped me pick a repair shop. After the call, I got a text and an email with instructions to the repair shop, and the owner of the other car immediately got a call from my insurance company about his compensation, and I did not have to worry about it at all. (I2)

I had ordered from their online store but I could return items at the [physical] store…that was very convenient, not having to pack and send the clothes that did not fit. (R2)

I was trying to activate my new mobile phone subscription, but it was very difficult. I got a letter that instructed that I should register in one place, then confirm in another site, and then activate it in a third place! The salespersons gave totally different instructions. (T1)

These findings find support in existing research, pinpointing that journey touchpoints should be functionally integrated to enable a smooth end-to-end journey (Homburg et al., 2017). Edelman and Singer (2015) note that firms should streamline journeys so that customers are able to execute complex service processes quickly and easily; and Rawson et al. (2013) recommend shifting from siloed to cross-functional approaches in developing journeys. Combining the qualitative insights with extant research, we define journey seamlessness as the degree to which touchpoints are integrated allowing a customer's smooth transition between various service process touchpoints.

The second SJQ dimension identified is journey coherence. According to the focus group participants, good service journeys feel consistent both in terms of visual as well as service style elements of different touchpoints. The participants appreciated a journey during which everything from the service personnel's style and manner of delivering the service to the visual elements within the service setting are aligned and convey the same message. The below quotes illustrate coherence:

The [mobile phone] service provider sent me an ad to promote its special package for young adults; it looked trendy and playful. The firm was also present in our student event; they organized a funny game and gave away energy drinks…Even their sales reps were cool; they called themselves “social media ninjas”—so everything was about the trendy, bold, and youthful brand… (T1)

It is easy since every Ikea store is almost identical, and it feels the same since you can see Ikea colors all the time…their website looks like Ikea since there is blue and yellow. You also see yellow and blue signs in the stores, and you collect stuff in a yellow bag, and they pack them in a blue bag! (R1)

Different service employees had completely different styles of doing things...So their service principles seem to depend on who you happen to ask! (R3)

These findings resonate with the notion of brand cues' thematic cohesion and consistency across touchpoints (see e.g. Berry et al., 2006; Kuehnl et al., 2019). Previous research on customer experience management has indeed recommended that firms should systematically manage and orchestrate “experience clues” across various touchpoints (Berry et al., 2006; Zomerdijk and Voss, 2010). Synthesizing the qualitative insights with previous research, we define journey coherence as the degree to which service process touchpoints provide consistent experience cues.

The third SJQ dimension identified is journey personalization that builds upon the notion of the individual and subjective nature of service journeys (e.g. Dhebar, 2013; Barwitz and Maas, 2018). Many focus group participants described how excellent service journeys feel tailor-made and adapted to each customer's unique situation and needs. The participants appreciated the freedom to choose the course of their journeys – that is, the touchpoints through which they preferred the service interaction to be held. Following quotes illustrate journeys that felt personalized for the customer:

We went to a big concert and stayed in a hotel close by. There were many concert attendees staying there, and the hotel had decided to extend the breakfast time the next morning so that we could sleep in after the concert! That seemed like excellent service. (H2)

I needed to file an application for compensation for a broken sink in my bathroom. I could've done that online, by phone, or by visiting their office, but I chose the mobile application since I like to do everything with my phone. (I1)

We were two families with small kids traveling together [on a cruise ship] and had booked family cabins. It was really convenient that all the family cabins were located along the same corridor, and party people were at the other end of the ship. Passengers with prams had their own entrance to the ship, and there was a ship mascot greeting us and handing coupons for free ice cream at the play area, our kids loved it! (H1)

Also previous research has noted the importance of designing touchpoints that are sensitive to the customer's situational context (Homburg et al., 2017; Kuehnl et al., 2019), highlighting that customers should find the touchpoint architecture adaptive to their changing needs (Dhebar, 2013). Combining the empirical and research insights, we define journey personalization as the degree to which the combination of service process touchpoints is tailored to fit the customer's preferences and situational context.

4.1 Qualitative scale development procedure

The conceptualization of SJQ enabled us to proceed into building measures for the construct. The scale development follows the established procedures for building new scales (Churchill, 1979; MacKenzie et al., 2011). Specifically, on the basis of the construct definitions and insights from the qualitative study, we built an initial indicator pool for all three SJQ constructs, coming up with 23 total indicators that have been designed to evenly reflect the key domain of each SJQ dimension. Next, we assessed the indicator content validity through a qualitative item-sort task test suggested by Anderson and Gerbing (1991). Twelve senior scholars were asked to assign each scale item under one SJQ construct definitions or option “other” if it could not be accurately assigned to any one definition. All but one indicator received more than ten out of twelve correct responses, thereby passing the suggested threshold criterial of psa (>0.5) and csv (>0.7). When a reviewer felt unclear about or rated an item incorrectly, the indicator was carefully evaluated, and some wordings were consequently fine-tuned. Finally, we selected the six representative indicators from the item sort test for each dimension to sustain the final scale at a reasonable length – see Appendix for the final scale indicators.

4.2 Data collection for the scale validation

To validate the developed measure, we conducted a consumer survey concerning SJQ in financial services where the process of accessing and using the service typically involves a multitude of touchpoints in different channels. The data were collected in a centrally located downtown shopping center in a northern European town. We randomly approached adult customers and asked them to participate in an academic study that concerned their experiences with their primary financial service providers. We implemented a movie ticket draw as an incentive for consumers to participate and guaranteed full anonymity for each respondent and received N = 278 responses. The data were deemed adequate for the initial scale validation purpose of the first survey study.

4.3 Assessment of the SJQ scale validity and reliability

The SJQ scale validity was assessed based on confirmatory factor analysis using AMOS 24 software. The initial analysis of the proposed three-factor SJQ model with 18 indicators led to the elimination of two items from the seamlessness construct due to problems with discriminant validity. The elimination of problematic items is possible since the reflective indicators are interchangeable and because the construct is unchanged when an indicator is removed (Bollen and Lennox, 1991). A purified three-dimensional model with 16 items demonstrated a satisfactory fit. The chi-square statistic was significant (222.4; p = 0.00), but the critical ratio of the chi-square over degrees of freedom was close to 2 (χ²/df = 2.2), thus indicating a reasonable fit. The central fit indices provide support for the scale validity: goodness of fit index (GFI) = 0.91; comparative fit index (CFI) = 0.95; Tucker–Lewis index (TLI) = 0.94; standardized root mean residual (SRMR) = 0.042 and root mean square error of approximation (RMSEA) = 0.066 [0.054; 0.078] (c.f. Hu and Bentler, 1999). All indicator loadings were above 0.70 and significant at p < 0.01. Figure 1 summarizes the CFA results.

The Fornell and Larcker (1981) test further supports the discriminant validity because the average variance extracted (AVE) values for all constructs exceeded 0.50, and the squared AVE values of each construct exceeded correlations with other constructs. Construct reliabilities were also satisfactory, as all Cronbach's alpha and composite reliability values were higher than 0.70 (see Table 3 for scale details).

Finally, we compared the proposed three-dimensional model against alternative models, including a null model, a single-factor SJQ model and three two-dimensional SJQ models, to further assess the dimensionality of the construct. All alternative models had poor fit (see Table 4), thereby supporting the proposed three-dimensional conceptualization of SJQ. All in all, the initial scale validation stage provided support for the validity and reliability of the scales.

5.1 Research model and hypotheses

The study's final stage focused on testing the nomological and predictive validity of each SJQ construct. For this purpose, we identified the two seminal constructs of customer loyalty and service quality from earlier service research, which are conceptually related to SJQ. Figure 2 below presents this study's research model. Next, we discuss the research model and its hypotheses as well as justify the use of composite constructs to test the hypotheses.

First, extant research has widely emphasized the importance of well-functioning customer journeys in attaining a more favorable business performance in today's markets (Dhebar, 2013; Lemon and Verhoef, 2016; Halvorsrud et al., 2016). The customer journey's positive business impact is likely to occur through an improved customer performance, which culminates in customer loyalty (Zeithaml et al., 1996). This link can be theoretically justified on the basis of the model of behavioral outcomes of service quality as proposed by Zeithaml et al. (1996). The framework states that customers' positive service quality assessments strengthen their relationships with the company and manifest in behavioral changes that in turn drive the company's performance. This logic can be extended to SJQ because the offered SJQ conceptualization builds upon the idea that seamless, coherent and personalized journeys represent central elements for superior service delivery in contemporary business. Accordingly, customer SJQ should positively drive customers' intentions to exhibit behaviors that manifest loyalty, such as positive word of mouth, recommendations and the continued purchasing from a firm. Thus, we hypothesize that:

Second, SJQ should be a concept closely related to, although conceptually distinct from service quality. Both constructs measure customers' quality assessment of service delivery but with different foci; service quality focuses on a customer's expected level of service and perceptions of the service level received, which are typically evaluated with regard to one particular touchpoint or the service as a whole (Parasuraman et al., 1991). On the other hand, SJQ focuses on customers' quality assessments related to the various interdependencies between service process touchpoints (cf. Dhebar, 2013). Studies have indicated that, in the complex service delivery networks prevalent in today's markets, incongruities between touchpoints that emerge during the end-to-end journey with a firm can ruin the customer's “total experience” (Rawson et al., 2013; Banerjee, 2014). We therefore propose that customers' service quality perceptions should not be exclusively driven by their evaluations of individual service encounters but also by connections between various contacts through which the service occurs as each encounter builds customer expectations for the next one. Therefore, we hypothesize that seamless, coherent and personalized service journeys should improve customers' service quality perceptions:

Service research has shown that service quality is a fundamental driver of customer loyalty in service business (Parasuraman et al., 1991). We also hypothesized above that SJQ has a positive relationship with customer loyalty (H1) and service quality (H2). These notions enable us to theorize that the hypothesized positive connection between SJQ and loyalty takes place because SJQ influences service quality which in turn drives customer loyalty. In other words, service quality is a mediator that theoretically explains why SJQ has a positive relationship with customer loyalty. However, recent research suggests that the quality of journeys can also drive other loyalty-building customer outcomes such as customer experiences (see Rawson et al., 2013; Lemon and Verhoef, 2016; Kuehnl et al., 2019). Therefore, we expect that service quality is a significant but not the sole mediator of the relationship between SJQ and loyalty, implying partial mediation:

5.2 Use of composite constructs in the research model

Theoretical constructs are not per se multidimensional or unidimensional, but can usually be operationalized either way, thus representing various levels of theoretical abstraction (Law et al., 1998). SJQ and service quality represent two theoretically distinctive, although complex concepts in the research model, as both possess multiple unique dimensions. Against this background, both constructs can be meaningfully operationalized at a higher level of abstraction by means of second-order composite constructs (Hair et al., 2017). The use of hierarchical component models offers two important benefits for this study. First, they can notably reduce model complexity by decreasing the number of relationships in the full mediation model, thereby increasing parsimony (Hair et al., 2017, p. 281). Second, although all studied SJQ and service quality constructs are independent by nature, they are likely to correlate with one another due their conceptual closeness concerning quality in service business settings. Establishing a higher-order structure can reduce potential collinearity issues (Hair et al., 2017, p. 281).

Research should always carefully consider specification of the studied constructs because model misspecification can severely bias structural parameter estimates and lead to inappropriate conclusions about the hypothesized relationships between constructs (Jarvis et al., 2003). We argue that SJQ and service quality are best modeled as first-order reflective, second-order formative (Type II) composite constructs. We build this argument upon the following four criteria that define whether a construct is more efficiently measured by a reflective or formative perspective: (1) causality between the construct and its dimensions, (2) interchangeability of the dimensions, (3) covariation among the dimensions and (4) whether or not all dimensions have the same antecedents and consequences (cf. Jarvis et al., 2003, p. 203). The qualitative study suggests that the three SJQ dimensions form the overall level of the construct rather than a uniformly reflect the construct as dropping one dimension would alter the SJQ's conceptualization. Similarly, the dimensions can but are not required to correlate since a service firm might score high in seamlessness, but at the same time fail to provide personalized service encounters. Finally, one may logically expect that the various independent dimensions can have different drivers and outcomes.

5.3 Data collection

The data collection for testing the research model was conducted in collaboration with a Northern European bank that is classified as the second-largest bank in the country wherein the study was conducted. The bank provided us access to its consumers in one of its branches and drew a sample of 4,757 customers from its customer base. A customer experience-labeled survey was sent to the customers in the university's name as an academic survey. A movie ticket draw was used to incentivize the selected customers to participate in the study. All responses were highlighted to be fully confidential, and the respondents were guaranteed full anonymity for their responses. The data collection with two reminders led to N = 239 customer responses. The respondents' characteristics are summarized in Table 5.

5.4 Measures

We employed established scales for all the constructs in the research model. Specifically, we measured customer loyalty using the scale by Zeithaml et al. (1996), service quality with the measure developed by Cronin and Taylor (1992) and SJQ with the developed scale. In addition, since the study relied on a single-respondent design, we included a common method variance (CMV) marker variable to the questionnaire. For this purpose, we used an a-priori “Consumer Orientation Toward Sporting Events” scale developed by Pons et al. (2006) with four indicators and no nomological relationship with other study constructs, as recommended by the established guidelines (Lindell and Whitney, 2001; Chin et al., 2013).

5.5 Analytical procedures

Since the maximum-likelihood-based SEM is problematic for testing models with formative constructs, we analyzed the research model with PLS modeling using the SmartPLS3.0 software (Hair et al., 2012). This method is closely suited to studies that build upon formative constructs and complex models with higher-order constructs and mediation effects. We implemented the guidelines established by Hair et al. (2012) to estimate the research model. The modeling of second-order composite constructs follows the guidelines of Becker et al. (2012) as well as Cadogan and Lee (2013). Specifically, the simulation study conducted by Becker et al. (2012) demonstrates that reflective-formative, hierarchical type II constructs are most effectively modeled by the repeated indicator approach, path weighting scheme and mode B measurement. We further tested the SJQ construct's relationship with the second-order formative service quality construct through its lower-order dimensions (see Figure 2), which represents a conceptually superior way to estimate antecedent relations for formative constructs (Cadogan and Lee, 2013). To account for the total effects on service quality, the explained variance in each dimension was multiplied by its weight, and the individual contributions of each dimension were added together (Becker et al., 2012). The statistical significance of the PLS parameter estimates were tested with a bootstrapping procedure based on 5,000 subsamples. For clarity, we estimated the research model in two parts; we first estimated a simple baseline model with exclusively direct relationships between SJQ and customer loyalty and second, estimated a full model with the service quality construct as a mediator.

5.6 Common method variance

Because the study relies on a single-respondent design, common method variance (CMV) must be taken into account. We relied on both procedural and statistical approaches to assess and control CMV as recommended by Podsakoff et al. (2003). The procedural means include respondents' guaranteed anonymity, careful scale item development for added clarity and the use of different scale anchors for IV and DV. We also used several statistical means to assess CMV. First, Harman's single-factor test produced seven factors with eigenvalues greater in an un-rotated factor analysis and no single factor explained above 50% of the covariance. Second, we applied Lindell and Whitney's (2001) partial correlation technique to assess the magnitude of CMV effects in construct correlations. Importantly, the marker variable was found to have low (0.06–0.12) and insignificant relationships with other study constructs (see Table 6), while the specific data analysis procedures (see Lindell and Whitney, 2001) did not exhibit any substantial changes in their correlation coefficients or their significance when controlling for the CMV marker variable. Finally, we applied a measured latent marker variable approach when testing our research model, which has been proven to effectively detect and correct CMV in a PLS analysis (see Chin et al., 2013). Specifically, our full research model includes separate CMV marker variables for all constructs in the model (see Figure 2). All marker variable paths were again close to zero and insignificant (see Table 7 for full details). Importantly, the marker variable's inclusion in the model did not change path coefficients or significances; overall, the results indicate that a common method bias is not a major problem for this study.

5.7 Testing the research model and hypotheses

We began testing the research model by assessing the scale validities and reliabilities. The outer model results show that all standardized indicator loadings exceeded the recommended threshold of 0.70 and were all significant at the p < 0.01 level (see Appendix). The construct reliabilities were supported, as all Cronbach's alpha and composite reliability values were higher than 0.70. Convergent and discriminant validity were supported because all AVEs were higher than 0.50 and because the AVE's square root exceeded the construct correlations (see Table 6 for scale details). The scale correlations were high, although closely aligned with earlier service quality research findings that have found key service quality constructs and customer performance outcomes to be tightly interrelated (Brady and Cronin, 2001; Dagger and Sweeney, 2007). The correlations are also theoretically meaningful since all the studied constructs measured diverse aspects of service performance, including service journey quality, overall service quality and customer loyalty. Finally, a separate test of PLS cross-loadings confirmed that all indicators loaded highest on the construct that they were intended to measure (Hair et al., 2012).

The inner model evaluation also supported the quality of the model. We examined VIF values of the first-order constructs in relation to higher-order formative constructs and concluded that all values were below the suggested threshold of 5 (Hair et al., 2012).

The structural model results provide support for all three hypotheses. First, the direct effect model confirms that the developed SJQ measure has a strong, positive relationship with customer loyalty, thus supporting H1 (see Table 7). Specifically, the second-order SJQ construct had a path coefficient of 0.76** for loyalty in the direct effect model, thus explaining 60% of its variance. All first-order dimensions had positive and significant weights: personalization (0.40**), coherence (0.38**) and seamlessness (0.29*), thus supporting the predictive validity of all three dimensions for customer performance. The first-order path coefficients to the second-order constructs can be interpreted similarly to formative indicator weights – that is, by indicating the relative contribution of the lower-order construct to the higher-order construct while predicting outcomes in the studied nomological network (Becker et al., 2012).

Second, the full structural model results support the second hypothesis that SJQ is a central enabler of service quality in today's complex business environment. Interestingly, SJQ was positively related to all dimensions of service quality: reliability (0.85**), assurance (0.73**), tangibles (0.68**), empathy (0.65**) and responsiveness (0.57**). Yet, only two first-order dimensions of service quality contributed significantly to the second-order construct when predicting loyalty. By calculating the total indirect effects (Becker et al., 2012), SJQ explains 53% of the variance in service quality.

Third, the full mediation model results also support the partial mediation of H3 because the relationship between SJQ and customer loyalty became weaker, yet remained significant when controlling for the service quality mediation link. A separate, bootstrapping-based Preacher Hayes mediation analysis confirms the significance of the hypothesized partial mediation (see Table 8).

Specifically, when controlling for the service quality link, SJQ's direct effect on customer loyalty (0.26**) remained substantial and highly significant. SJQ also had a strong indirect effect (0.50**) on customer loyalty through the service quality link. Interestingly, SJQ's total effect on loyalty (0.76**) due to its direct and indirect effects was stronger than service quality's effect on loyalty (0.58**). Finally, a closer look at the dimension weights of the second-order constructs provides insight into the relative importance of the dimensions in the broader nomological network. When broadening the study focus to the full research model with both service quality and loyalty as dependent variables, the SJQ dimension weights slightly changed as the relative impact of personalization (0.56**) and seamlessness (0.28**) became more important compared to coherence (0.21*). In line with earlier findings in banking contexts (see Bloemer et al., 1998; Choudhury, 2013), reliability (0.70**) is the dominating dimension of service quality when predicting customer loyalty. Beyond reliability, the tangibles (0.18*) represent the only other significant 2nd order service quality dimension.

6.1 Theoretical implications

This research responds to calls for the development of a more thorough understanding of what makes service journeys excellent and supportive of superior customer outcomes (Ostrom et al., 2015), as well as to the managerial need for easily applicable tools for measuring performance in service delivery through complex journeys. The customer journey is considered a key concept for understanding the emergence of customer experiences (e.g. Lemon and Verhoef, 2016), but extant research has overlooked the development of this concept to fully capture the essence of service processes. This paper informs research and managerial practice by conceptualizing SJQ, developing measures for the construct and studying its relationship with service quality and customer loyalty. Extant journey research spans various literature fields (Table 1), thus rendering the findings relevant for many areas of service research. This study makes three main contributions, which are discussed in detail below.

First, the conceptualization and measure development for SJQ offers researchers constructs that capture the key aspects of functional customer journeys from the service process point of view. This contributes to the service management and experience research by moving the focus from individual encounters or overall service or brand evaluations (e.g. Parasuraman et al., 1991; Brakus et al., 2009; Lemke et al., 2011; Kumar et al., 2014) to the potentially multichannel and multiprovider service journeys as whole. As such, the SJQ construct provides a modern service quality tool that tackles service delivery through complex journeys, whose quality cannot be exclusively understood as an aggregate sum of individual service encounters (Rawson et al., 2013). For service design research, the SJQ construct offers a set of concrete goals for designing journeys to ensure that cues and touchpoints operate in concert and allow for personalized configurations, and it enables the measuring of to what degree these designed journey qualities are achieved. For customer experience research, SJQ offers a journey conceptualization that captures the key elements that affect CX formation in service-intensive contexts. This supports customer experience management (Homburg et al., 2017) and complements the constructs developed by Kuehnl et al. (2019) that link brand-focused journey qualities to brand attitudes, brand experiences and loyalty, but scarcely address the operational and functional touchpoint interdependencies relevant for service processes. For multichannel customer management research, the SJQ construct provides a tool for analyzing the performance of channel integration in service-intensive contexts, complementing the existing journey perspective that is purchase-process focused.

This study's second contribution is in confirming that SJQ is a critical driver of customer performance in contemporary service businesses. Specifically, we demonstrate that service journey seamlessness, personalization and coherence act as central drivers for both perceived service quality and customer loyalty intentions. The positive connection between the quality of journeys and service quality is intuitive, yet no earlier study has demonstrated this link. This study's findings indicate that customers' overall service quality assessments are affected by their perceptions of how successfully various service delivery touchpoints operate in concert. This link is evident even though many service quality dimensions focus on personal interactions at a single encounter. It is likely that encounters at earlier stages of the service journey build expectations that affect the customer's interpretation of later service encounters. In other words, high-quality journeys reduce negative surprises and enable the consistent meeting of customer expectations, which builds better service quality perceptions. Thus, this study demonstrates that SJQ represents an important service quality enabler and provides evidence for the assumption that, in today's multichannel, digitalized markets, service excellence warrants the perfection of service journeys (cf. Halvorsrud et al., 2016; Lemon and Verhoef, 2016). This finding is underscored by the empirical finding that SJQ has a stronger total effect on customer loyalty than on service quality.

Finally, this study contributes to the existing literature by clarifying the nomological network of service journeys and the theoretical mechanisms through which journey quality affects loyalty. Our findings reveal that service quality mediates the direct relationship between SJQ and customer loyalty. In other words, improved service quality represents a central mechanism that explains why high-quality journeys can help secure more loyal customers. This result complements earlier studies that have attributed the link between journey quality and loyalty primarily to improved brand attitudes (Kuehnl et al., 2019). However, this mediation is merely partial, thus indicating that SJQ contributes to customer loyalty above and beyond service quality when controlling for the mediation link. This finding is of great importance because it indicates that the relationship between SJQ and customer loyalty cannot be explained by improved service quality perceptions alone, but there may be other potential mediators as well, such as customer experience (see Lemon and Verhoef, 2016; Becker and Jaakkola, 2020). This proposition is discussed further in the implications for future research section.

In sum, this research provides new measures for service research, as well as for managers who wish to set goals, understand and develop their performance in service journeys. This study is among the first to develop the concept of service journeys to capture the nature of the process customers go through to access and use offerings in service-intensive contexts, in order to complement the customer journey concept that is predominantly anchored in the consumer decision-making process. In doing so, this study helps organize the pieces of the customer experience and journey puzzle by clarifying and nuancing the nomological network of some of these key concepts of contemporary service research.

6.2 Managerial implications

Contemporary service businesses rely increasingly on numerous interconnected online and personal service delivery channels, as well as diverse partners, in providing a service. This means that firms need to consider and manage various expectational, operational and functional interdependencies between various service delivery touchpoints (Dhebar, 2013). While service design tools, such as blueprinting, have been developed to help in designing functional service journeys, very few tools have been developed for assessing the quality of service journeys from the customer's viewpoint. This is problematic because measuring perceived service quality at individual touchpoints says little about the quality of the service process as a whole, and simple aggregate measures, such as the Net Promoter Score, do not reveal the cause of customer satisfaction or dissatisfaction. This study develops constructs to assess customer-perceived SJQ and provides empirical evidence for the relevance of high-quality service journeys, demonstrating that consumers' SJQ perceptions drive both service quality and customer loyalty. These findings indicate that service providers should invest in the development and monitoring of SJQ instead of just focusing on individual service encounters or overall satisfaction.

Specifically, our results highlight that service journey quality from the customer's perspective is determined by the three key dimensions of journey seamlessness, coherence and personalization. These SJQ dimensions provide managers concrete goals in designing their service journeys. First, journey seamlessness requires that the various provider and partner-controlled touchpoints along the service process are integrated and aligned so that the customer can smoothly transition between the touchpoints, whether outsourced or not. Second, journey coherence can be achieved by thematically integrating all touchpoints and the related “experience cues” to provide a consistent impression of the firm or the brand throughout all service encounters. Third, journey personalization stresses the tailoring of the combination of service process touchpoints to fit the customer's service delivery preferences and the customer's situation. Firms can use the SJQ measures developed in this study to analyze, monitor and develop the quality of their service journeys, as well as to set collaboration and quality goals for partners (see Appendix).

The three SJQ dimensions can also be used to analyze the functionality of customer relationship models for different customer segments that typically follow distinct journeys (e.g. when customer segments are served through different service channels). Measuring SJQ perceived by different customer segments helps the internal benchmarking of journeys, in order to reveal potential pain points in different journeys and improve performance, particularly for mass segments wherein the customers navigate their journeys independently. We also urge companies to assess whether or not SJQ might be a source of differentiation and thereby a competitive advantage in their industry and to develop their journeys accordingly.

6.3 Limitations and implications for future research

This study naturally has some limitations. Noteworthy is that the SJQ construct does not capture social and customer-owned touchpoints that are outside a firm's influence (Lemon and Verhoef, 2016; Becker et al., 2020), but instead focuses on provider- and partner-owned touchpoints, i.e. the parts of the journey that firms can seek to manage (Becker and Jaakkola, 2020). Furthermore, as the SJQ construct focuses on the quality of the service process, it does not cover what the customer receives as the outcome of the service. Therefore, when the aim is to study both technical and functional quality (Grönroos, 1984), the SJQ construct should be used together with complementary measures.

While the conceptualization builds upon a broad empirical basis that includes a qualitative study and two quantitative studies, further research related to the SJQ construct's nomological network is needed. This study has identified initial evidence that service journey quality predicts customer loyalty, but those findings were based on a single-respondent study design with subjective performance measures. Future research should confirm the findings using stronger research designs with objective performance measures and multiple-respondent study designs. Also, since the quantitative study findings are based on data exclusively from the financial sector, future research should investigate the relative importance of the construct dimensions and test the results' generalizability in other empirical contexts. We assume that the developed measure is applicable across service-intensive industries, but this assumption warrants further research. Researchers could study not only brand-focused contexts, such as tourism and hospitality but also contexts such as healthcare, public services, and B2B settings, wherein more functional interdependencies between touchpoints should be a key. Future studies could also examine the SJQ construct's application to experience-centric and hedonic versus mundane services, as well as to different types of journeys.

We call for a systematic research effort that focuses on SJQ and its antecedents, outcomes, mediators and moderators. Various areas of service research – including service management, service design and customer experience management – should more closely study which marketing activities drive seamless, coherent and personalized service journeys. On the outcome side, future research should study SJQ's effects on objective performance and examine whether or not improved customer experience also mediates the performance link. The role of experiential outcomes is emphasized by the study finding that the SJQ-loyalty link is not fully mediated by service quality improvements. Since customer experience refers to customers' nondeliberate, spontaneous responses and reactions to offering-related stimuli along the customer journey (Becker and Jaakkola, 2020), high-quality service journeys should logically be connected to positive customer experiences. Due to the complexity of the customer experience phenomenon, its measurement is challenging. Advances in this area can offer the means to study the suggested experience link in more detail (e.g. the EXQ measure by Kuppelwieser and Klaus, 2021), in order to clarify whether and how service journey quality relates to customer experiences. Finally, we call for research attention on the key moderators – that is, those conditions that either strengthen or weaken SJQ's antecedent and outcome sides in different contexts.