1 Introduction

To improve productivity, many service companies have taken inspiration from ‘lean production’. Consequently, a growing number of service companies have launched lean initiatives (Allway and Corbett 2002; Netland and Powell 2016; Liker and Ross 2017; Fenner et al. 2022). This research is motivated by the real-world observation that many global service firms run consolidated corporate lean programmes across all their service units. Such programmes are unlikely to have the same fit for different types of service processes (Ansari et al. 2010; Netland and Aspelund 2014). We challenge the implicit assumption in many studies of universality of lean practices in service settings. We do so by exploring how employees in different service types perceive the relevance of standard lean service practices for their job contexts and how they respond to the mandate of adopting them. Hence, our research question is, ‘How do standard lean service practices fit different service types?’.

There is general agreement in the lean production literature that lean practices, when properly used, improve operational performance in manufacturing firms (Shah and Ward 2003; Dal Pont et al. 2008; Onofrei and Prester 2019). However, the effectiveness of lean practices in the service industry is more contested. On the one hand, many articles have suggested that lean improves performance in various service industries: examples include financial services (Holweg et al. 2018), information services (Staats et al. 2011), education (Alves et al. 2016), hospitality services (Rauch et al. 2020), call centres (Piercy and Rich 2009), public services (Bateman et al. 2018) and healthcare (Meredith et al. 2011). On the other hand, other empirical studies have shown that lean does not deliver its promise in service industries. For example, Poksinska et al. (2017) found that healthcare institutions that apply lean practices do not have more satisfied patients than those that do not focus on lean; Holweg et al. (2018) and Secchi and Camuffo (2019) reported failed lean transformations in the banking sector; and Radnor (2010) and McCann et al. (2015) found that public service organisations that adopt lean production did not realise the expected benefits. Clearly, there is mixed evidence of the effectiveness of ‘lean service’.

Rather than trying to understand how lean differs across service industries (banking, healthcare, telecom, education, etc.) or fits a particular service industry (cf. Langabeer et al. 2009; McDermott and Venditti 2015; Vanichchinchai 2021), we focus on different service types. This paper explores the fit between service types and standard lean service practices. Service companies comprise many different types of service operations (Johnston and Clark 2008). Consider the example of education services where tasks are as different as student enrolment, classroom teaching, facility management, examination and grading. Or in restaurants where service tasks include booking and table placement, order taking, cooking, serving, cleaning, billing etc. Furthermore, also manufacturing firms comprise various service operations, such as purchasing, maintenance, accounting, internal audits, sales, marketing and after-sales services. The key idea of this paper is that – because service operations differ in terms of customer contact, routine variability and technology usage, among other dimensions (e.g. Thomas 1978; Silvestro et al. 1992; Ponsignon et al. 2011) – the applicability of lean practices is likely to be more associated with the service type than with the service industry.

For this research we studied a multinational utility company that comprises many different types of services but deploys a consolidated corporate lean programme across all units. We conduct qualitative embedded case study research, which means studying different subunits within the same organisation and comparing them to identify common and conflicting patterns. The sample included four professional service teams, five service shop teams and six service factory teams. We conducted in total 47 interviews: 36 interviews with team members who adopted lean and 11 interviews with senior managers who oversee the lean programme. The interview data were analysed using a structured qualitative analysis. The findings are summarised in ten propositions that seek to explain what conditions reduce or increase the fit between standard lean practices and service-type settings.

This paper contributes a process-based perspective on lean services to the literature. It supports the idea that many standard lean service practices are applicable across various service types but emphasises that some require more adaptation to the context. We find that practice adaptation is more crucial for managerial lean practices than for technical lean practices and in a service context characterised by higher employee skills and discretion, less routine work, and in field and front office work. Overall, our research findings challenge the rationality of generalising lean to a whole industry, as in ‘lean healthcare’, ‘lean education’, ‘lean governments’ or ‘lean utility’. For practitioners, the findings can help companies overcome the challenge of context–practice fit, which is a critical success factor in corporate programmes. Managers can use this paper’s insights to tailor the lean programme expectations and actions to different service units within their firms.

2 Theoretical background

Over the past three decades, lean production practices have been introduced to many industries (Netland and Powell 2016; Cusumano et al. 2021). The industries classified as services have not been an exception, showing a growing interest in ‘lean service’ (Bowen and Youngdahl 1998; Allway and Corbett 2002; LaGanga 2011; Leite and Vieira 2015; Gupta et al. 2016; Hadid et al. 2016; Liker and Ross 2017; Seddon 2019; Fenner et al. 2022). In the next subsections, we first review the literature on lean service and summarise standard lean service practices; secondly, we discuss differences among service types; and thirdly, we derive the research gap concerning the fit between service types and standard lean service practices.

2.1 Lean service

Danese et al. (2018) reported that only 11% of the lean literature investigates the service sector—of which half concentrates on healthcare. This is not much considering the service industries’ contribution to GDP in most countries. For example, according to World Bank 2020 data, the service sector contributes an average of 65.6% to the GDP in the European Union, 54.5% in China and 80.1% in the USA. Despite its importance, it is repeatedly noted that the service sector has been lagging behind the manufacturing sector in terms of productivity (Allway and Corbett 2002; Liker and Ross 2017; Seddon 2019). We take this as a motivation to study lean services.

Several authors have conceptualised lean in service industries and specified how it differs from lean production (e.g. Bowen and Youngdahl 1998; Allway and Corbett 2002; Leite and Vieira 2015; Liker and Ross 2017). Although there is no general agreement on which specific practices constitute lean services, a review of the literature by Gupta et al. (2016) reveals a set of practices that are commonly included in lean services listed in Table 4 in Appendix A. These practices can be separated into three systems based on what they aim to transform (Found and Bicheno 2016):

  • The operation system

  • The management system

  • The culture system

A lean operations system is characterised by the efficient use of resources, which makes service production flow without disturbances. Practices that are commonly used to achieve this include process mapping, process standardisation, pull production, waste reduction, process capability (e.g., simplification, error prevention, built-in quality), workload balancing and workplace organisation (e.g., ‘5S’) (Gupta et al. 2016).

A lean management system refers to the managerial practices needed to establish and support a lean operations system. Frequently mentioned examples of such practices in the lean service literature are visualisation (‘visual management’), supporting organisational infrastructure (e.g., ‘lean teams’, ‘champions’), performance management (e.g., ‘daily huddle meetings’, ‘visual performance boards’), hands-on management (e.g., ‘gemba walks’, ‘early triage’ in hospitals) and training and skill management (e.g., ‘multiskilling’) (ibid.).

Lastly, a lean culture system is the foundation for sustained lean transformation (Liker and Ross 2017). Common practices that foster a lean culture include continuous improvement practices (e.g., ‘kaizen events’, ‘rapid improvement events’ and ‘plan-do-check-act’), customer orientation (e.g., ‘voice of the customer’ and ‘value for the customer’) and employee involvement (Gupta et al. 2016).

Because manufacturing and services differ, it is likely that lean practices should also differ in these two contexts. Whereas manufacturing delivers material products, services are activities that produce immediate, intangible and non-storable products. However, considering the practices mentioned above, it is not apparent how lean services differ from lean production. Except for ‘early triage’ (a lean practice in hospitals), all the mentioned practices are also common in manufacturing. It seems that lean service deviates from lean production in what is absent. In addition to the lean ‘service’ practices listed above, production companies tend to include a range of lean practices directly related to making physical products in factories. For example, practices related to total productivity maintenance, total quality management, takt time, supplier integration and supply chain, and inventory management are more prominent in production settings (e.g. Shah and Ward 2003; Netland 2013; Leite and Vieira 2015).

2.2 Service types

The service industries are highly heterogeneous (Johnston and Clark 2008; Victorino et al. 2018). Running a fast-food restaurant is entirely different from running a prison, utility company, hospital, university or consultancy. Even within a service firm, there are numerous different types of services. For example, there are fundamental differences in the job context of those who work in the internal audit department, human resources department or sales department. Hence, rather than looking at lean service at the firm level or across different service industries, we follow the advice of Shostack (1987), Wemmerlöv (1990) and Safizadeh et al. (2003), who suggest studying services at the process level.

Several operations management authors have suggested service process typologies (Chase 1978; Schmenner 1986; Wemmerlöv 1990; Silvestro et al. 1992; Ponsignon et al. 2011). As evidenced by Cook et al. (1999) comprehensive review of service typologies, there is considerable overlap between the typologies suggested in the literature. Typologies usually separate three or four types of services according to different service process characteristics. We adopt the typology proposed by Ponsignon et al. (2011) in their paper ‘Service delivery system design: Characteristics and contingencies’.

The service typology of Ponsignon et al. (2011) represents a meaningful classification of service process types. Ponsignon et al. (2011) draw on previous service classifications to propose three service process types: ‘professional service’, ‘service shop’ and ‘service factory’.Footnote 1 The typology has much in common with the classifications of Silvestro et al. (1992) and Kellogg and Nie (1995). However, unlike the many typologies that use their process typologies to categorise firms in service industries, Ponsignon et al. (2011) categorise service units within a firm. They sort services into three archetypes based on the role of people, the role of technology and equipment and the role of location and layout in the service process (Ponsignon et al. 2011). Each of the three ‘roles’ covers two dimensions, resulting in six dimensions of the service delivery mechanisms: level of skills and degree of employee discretion, degree of routineness and (potential) degree of automation and location and front-office versus back-office configuration. Figure 1 illustrates the dimensions of the typology and the resulting three service process types.

Fig. 1
figure 1

A typology of service processes (based on Ponsignon et al. 2011)

Full size image

Professional services tend to require specialised skills, provide a high level of discretion for the employee, be customised with a low automation potential and be delivered close to the customer in the field front office. On the opposite side of the spectrum, service factories require less specialised skills, provide a low level of discretion for the employee, follow standardised processes, have a high potential for automation and are delivered distantly to the customer in the back office. The service shop is characterised as something between professional services and a service factory. The typology is a stylised model that does not perfectly match all service processes. For example, there are also back-office jobs that require specialised expertise.

2.3 Fit between practices and service type

A lean transformation aims to instil a set of carefully selected lean practices throughout an organisation. Because service firms span a variety of internal and external environments, contingency theory is a useful theoretical perspective for studying lean practices (Donaldson 2001; Sousa and Voss 2008). According to Danese et al. (2018), contingency theory is the most frequently used theory in the lean literature (for examples, see Peng et al. 2011; Netland 2016).

Central to contingency theory is the concept of ‘fit’ (Drazin and Van de Ven 1985; Donaldson 1987, 2001; Sousa and Voss 2008). The concept of fit has also been discussed thoroughly in the strategy literature – explicitly or indirectly taking a contingency theory perspective (see Drazin and Van de Ven 1985; Donaldson 1987; Ansari et al. 2010; Volberda et al. 2012). Contingency theory suggests that organisations seek to achieve fit (or congruence) between their practices and their internal and external institutional environments. Hence, according to this theory, not all lean service practices should fit similarly well across different service types.

According to contingency theory, units that experience a high degree of fit between their environment and a specific lean service practice are expected to adopt the practice, whereas units that experience a low degree of fit are likely to reject or adapt the practice (Ansari et al. 2010). Here, adoption refers to an implementation of the practice that comes close to the original template, and adaptation ‘refers to the process by which an adopter strives to create a better fit between an external practice [by adjusting it] and the adopter’s particular needs to increase its “zone of acceptance” during implementation’ (Ansari et al. 2010, p. 71). It is worth noting that fit is not a binary variable but a continuous one. It is dynamic, multidimensional and dependent on the perceptions of humans (Ansari et al. 2010).

Against this theoretical background, we set out to conduct an explorative study into the fit of standard lean service practices across different service types.

3 Research method

To investigate our research question, we chose a qualitative embedded case study research design (Voss et al. 2002; Yin 2013). An embedded case design investigates several sub-units within the same case firm (Scholz and Tietje 2002). The phenomenon we study occurs in a context-rich environment with complex social structures. In this setting, case studies are recommended (DeHoratius and Rabinovich 2011). Case studies allow the exploration and formulation of propositions in order to generate new theoretical insight (Yin 2013).

We decided to conduct this research in one company to hold many contextual variables constant. We needed a firm that spans a range of service types and has deployed the same corporate lean programme across all units for a few years. We were lucky to team up with a leading European energy supplier, which we call ‘Energy Co.’, in the autumn of 2016. Energy Co. is a leading international company in its industry. We worked with Energy Co. until 2019. At the time of our research, it employed over 40,000 people, served over 30 million customers and had an annual revenue of approximately 40 billion euros. The company focuses on renewables, energy networks and energy-related customer solutions.

3.1 Energy Co.’s corporate lean programme

Facing increased competition and cost pressures, Energy Co. embarked on a corporate lean programme in 2014. Energy Co. selected and tailored commonly known lean practices into a programme aligned with its company culture and characteristics. The programme consisted of nine sets of lean practices clustered into three ‘blocks’: operating processes, management systems and mindset and behaviour. The blocks consist of nine sets of lean practices, which are summarised and exemplified in Table 1. The practices pursued by Energy Co. overlap considerably with the standard lean service practices reported in the literature (see Section 2.1). The practices were intended to fit all units and teams of Energy Co.

Table 1 The three blocks and nine sets of practices of Energy Co.’s lean programme, with descriptions and examples
Full size table

To track the status and incentivise the implementation of the lean programme, Energy Co. conducts regular lean maturity assessments on a team level. A maturity assessment consists of interviews with team members, observations of meetings and work areas, discussions and a feedback session. This results in a score of the team’s performance in each of the nine sets of practices. Two assessors spend one or two days onsite conducting an assessment. The unit receive an aggregated score ranging from 0 to 100%. The scores are represented as 1 to 3 ‘stars’, with a cut-off value > 25% for 1 star, > 50% for 2 stars and > 75% for 3 stars. In 2018, Energy Co. received external certification for its lean programme and won a prestigious international award for its efforts, indicating that the lean programme is a serious effort.

3.2 Data collection

We applied a purposeful group characteristics sampling strategy (Patton 2014, p. 267). According to Patton (2014), purposeful group characteristic sampling chooses cases that meet specific criteria to highlight relevant patterns within groups. In this study, we looked for teams representing the three service types of Ponsignon et al. (2011): professional service, service shop and service factory.

To avoid an undesirable selection bias, the teams should not have been recently established or reorganised and should have more than a year’s experience with the lean programme. We also wanted diversity in each of the service types. We sampled teams from three different countries: Germany, the United Kingdom and Romania. Energy Co. mandated that all teams implement the Energy Co. lean program around the same year (from 2014 onwards), which reduces concerns about bias due to prior differences in experience with lean services. In total, we examined 15 teams. Table 2 reports an analysis of the teams according to the six service supply mechanisms discussed in Ponsignon et al. (2011) and shown in Fig. 1. The analysis is based on observations from interviews during field visits and discussions with Energy Co. The sample included four professional service teams, five service shop teams and six service factory teams.

Table 2 The 15 Energy Co. teams were ranked and sorted according to the service typology of Ponsignon et al. (2011)
Full size table

We interviewed 36 employees directly supporting or leading the lean transformation projects in the teams (e.g. department manager, team leader, lean coach or lean auditor) at different locations in Energy Co. To understand the extent of local adoption or adaptation from the perspective of headquarters, we also interviewed 11 senior managers who were responsible for the entire lean programme at Energy Co. Figure 2 shows an overview of all interviews conducted at Energy Co. and Table 5 in Appendix B provides details of the interviewee profiles and the interview locations and durations.

Fig. 2
figure 2

Interviews in Energy Co. (number of interviewees in parentheses)

Full size image

More than half of the interviews were conducted face-to-face during onsite visits and the rest were conducted through online telecom conferences (cf. Appendix B). The interviews were semi-structured, allowing for the ‘reconstruction of the interviewee’s subjective view (Flick 2014, p. 207) about lean services. We used two interview guides: one for the teams based in Energy Co.’s subsidiaries (Appendix C) and one for the lean programme management team at headquarters (Appendix D). Both guides were pretested on employees at Energy Co. and carefully reviewed after the pre-test. Small adjustments were made as we progressed with the interviews and learned which questions triggered the most valuable information. The interviews ranged from 30 to 60 min in length and were conducted in German, English and Romanian. The German and Romanian interview guides were translated and back-translated to English by native-speaking staff in the research organisation. With the interviewees’ consent, we tape-recorded all interviews and offered them the opportunity to review the transcripts. We transcribed all interviews in full, resulting in about 700 pages of raw text. To increase internal validity, interviewees were invited to review their recorded statements.

We also collected a range of secondary data, including audit scorecard data and performance board data for the teams we interviewed. In addition, we attended two full-day lean maturity assessments and a three-day lean conference at Energy Co. and took field notes from numerous informal discussions. Shadowing the lean assessments and reviewing programme documentation allowed us to gain in-depth knowledge about Energy Co.’s lean programme and its application across different service types. We also collected data about the activities of the teams and the headquarters unit. All secondary data were stored and indexed in the research database.

3.3 Data analysis

For the data analysis, we followed the general three-phase process of Miles et al. (2014). In the first phase, we sorted and coded the data. In the second phase, we searched for patterns within and across cases. In the third phase, we sought to explain the emerging propositions. As is typical in qualitative research, our approach was iterative (Eisenhardt 1989).

In the first phase, we created a data inventory with transcribed interviews with team-level employees, company scorecard data (lean assessment scores) and observations and notes from team-level visits. We used Atlas.ti software to structure, codify and map emerging associations. Following Saldaña (2015), we chose descriptive coding. We first descriptively coded the text according to the three defined service types (professional service, service shop and service factory) and the three lean programme blocks of Energy Co. (operating processes, management system and mindset & behaviour).

We coded quotes reflecting either fit or incompatibility between Energy Co.’s lean practices and the team. These quotes typically referred to different reactions, emotions, actions and experiences associated with the fit of the lean practices to the interviewees’ working context. Practices that were positively perceived and reportedly implemented in line with corporate standards indicated fit and were coded as ‘adoption’. Practices that were negatively perceived and reported to be rejected, changed, or only superficially implemented indicated a low level of fit and were coded as ‘adaptation’. Practices that were not mentioned or mentioned without evoking any particularly positive or negative feelings, were not considered. In a subsequent step, we searched the interviews conducted with interviewees from headquarters for comments that would support or oppose the team-level findings. We uncovered only a few occurrences of discrepancy and decided to remove the affected quotes from further analysis.

Second, we searched for patterns in the data (Miles et al. 2014). We iteratively developed propositions (or ‘hypotheses’ in the language of Eisenhardt), which we sought to explain with the data and literature (Eisenhardt 1989; Yin 2013). During the data analysis process, different propositions emerged and were subsequently either discarded or refined. We triangulated between the two-level interview data, quantitative assessment scores and observation in Energy Co., and – as a result of this process – kept the propositions that could be backed with convincing evidence and that we found interesting in terms of practical and theoretical contributions.

Finally, we challenged the emerging propositions by looking for discrepancies or competing arguments and findings in the literature and through discussions with Energy Co. We also used post hoc discussions with Energy Co. to understand the underlying mechanisms in the suggested propositions.

4 Findings

By triangulating the quantitative and qualitative data, we sought patterns in lean service practice adoption and adaptation among the service types. Sorted by service type, Table 3 summarises the lean assessment scores of all teams at the end of the data collection. Sorted by the set of lean practices, Table 6 panels A, B and C in Appendix E illustrate the coding of exemplary quotes and their sorting into ‘adoption’ or ‘adaptation’. As explained in the methods section, we focused our analysis on practices that evoked clear responses—negative or positive. This reduced the number of lean practices explored in the analysis. The practices included are ‘standard working environment’ and ‘process optimisation’ from the operational lean practices set, ‘performance management’ from the managerial lean practices set and ‘leadership’ from the cultural lean practices set.

Table 3 Team-level lean assessment scores by summer 2019 (percentage points)
Full size table

We structure the within-case narratives by service type: professional service, service shop and service factory. For each service type, we briefly introduce the teams and draw on the quantitative and qualitative material to present their perspectives on the fit of operational, managerial and cultural lean service practices.

4.1 Professional services and fit with lean service practices

All the professional services we include are field services (cf. Silvestro et al. 1992). The four teams included one team doing technical field services of a power grid in Germany, one team working in network distribution in Romania, one team installing meters and another team maintaining them in the UK. All these teams employ educated electricians with a large degree of autonomy in their work tasks. They are used to being alone for long periods and have limited interactions with their peers. Overall, we found that the employees performing professional services felt that the lean service practices of Energy Co. did not cater to their specific contexts and needed to be adapted to fit.

As shown in Table 3, the lean maturity assessment scores are diverse in this group,Footnote 2 ranging from ‘1 star’ (53.1%) to ‘3 stars’ (94.2%), averaging 74.8%. Across all sets of lean practices, professional services scores were, on average, the lowest in Energy Co. As can be calculated from Table 3, professional services scored an average of 83% for operational lean practices, 64% for managerial lean practices and 77% for cultural lean practices. The exception is the Romanian field service team, which stands out due to its superior performance on the managerial lean practice set (100%). We learned that the manager in this Romanian team was fully devoted to the lean programme and received sustained support from senior managers in Romania.

The qualitative evidence explains the relatively higher levels of adaptation of lean service practices in professional services. Our data suggest that employees in this service type see themselves as experts and enjoy more freedom in how they conduct their work than do back-office employees. A team leader in the UK reported, ‘Field technicians are quite different [from office workers]. A seasoned person who has been out in the field spends days walking around in storms, blizzards and everything’. Generally, this autonomy makes field service technicians question and challenge any new practices coming from the office environment.

However, this evidence does not suggest that lean practices do not apply to professional services; rather, they must be implemented with greater caution and adapted to the context. The Romanian team provides evidence that professional service teams can achieve high levels of lean implementation. For example, one team leader on another professional service team in the UK gave a powerful testament in favour of the lean programme: ‘I am a massive fan of it [the lean programme]; [it is] the best thing that could have happened to me. Lean really put my interest back into work’. Other interviewees reported that there were clear benefits to having a standardised work environment, even in highly variable work environments. For example, one UK-based interviewee from a field services team explained that common process documentation ‘brought everyone to a [common] standard’, which improved the team’s know-how and performance. Other interviewees reported that the introduction of workplace organisation principles (e.g., ‘5S’) in the service cars helped them ‘become more organised’.

All field service teams, except the Romanian team, struggled mainly with managerial lean service practices. For example, field service teams struggled to implement the daily ‘performance dialogues’Footnote 3 that were mandated by the programme. As one team leader in a UK team lamented, ‘It takes me up to four hours to update the board every week’. We also heard accounts of negative experiences with lean training sessions from many of the field service teams. Apparently, it was hard to get all the teams together (due to the geographically spread activity of field services), and the standard training had a poor fit with professional services. ‘Long-term projects were not considered, and the applied examples were not relevant to field services’, according to team leaders from Germany and the UK.

Managerial lean practices appear as more difficult to implement and sustain in field services and it is also a challenge to build a common lean culture among field services colleagues. For example, the lean leadership practice of ‘go and see’ is experienced as impractical in field services. A lean manager in the UK reported, ‘We sort of fail [the lean programme], as we cannot fit four go-and-sees into our day jobs as expected’. When people work autonomously and remotely with limited network or office access, it can be harder to create a shared lean culture. It can, however, be argued that it is precisely under such conditions—when managers have limited possibilities to control and follow up—that a shared lean culture is most important for good performance.

We also noticed that there were benefits to the lean programme that the field staff were surprised to experience but that the headquarters staff had expected. The joint programme lowered the organisational walls that had prevailed in their relationships with back-office operations and headquarters. The lean programme helped forge a bond between field workers and office staff. Imposed meetings (for example, the mandatory and daily ‘performance dialogues’) improved transparency, team interaction and empowered people to solve inter-departmental problems.

4.2 Service shops and fit with lean service practices.

Our sample of service shop teams consisted of five teams: a planning team for new power grid connections and a field compliance team in the UK, a complaint management team and an operational support team for grid field services in Romania and a grid calculation team in Germany. These teams work in centralised back-office functions and handle work tasks with a medium level of variation and complexity.

As can be seen in Table 3, the lean maturity assessment scores range from ‘2 stars’ (73.0%) to ‘3 stars’ (99.1%). The average scores for operational practices were 92%, for managerial practices 81%, for cultural practices 86%, and overall, 86.3% (note that the Romanian operations team did not receive a lean maturity assessment at the end of the data collection). The scores and our interviews indicate that lean practices have a reasonable fit with service shop teams. A team leader in Germany suggested that the lean programme ‘applies to four of five of the processes in our context, while the remaining processes (typically large, long-term projects) require other methods and tools’.

As in all service types, we noticed resistance to work standards. One team member grumbled, ‘I am back office, so why does it matter what my desk looks like?’ Another said, ‘We lost our family atmosphere [after the introduction of standards]’. However, the thorough creation of user guides improved the teams’ clarity regarding the main work streams and promoted a uniform work style. The teams reported that they were more aligned. The team leader in the Romanian field service operational support team reported, ‘Lean standardises our responses and provides a clear working procedure for the different customer requests’. Other teams reported positive bottom-line effects. For example, one team in the UK reported that the lead time for a difficult process was reduced by three months.

Service shop teams were generally receptive to managerial lean practices. Several interviewees reported that team communication had improved due to the introduction of daily performance dialogues. ‘Meetings without a detailed agenda don’t take place here anymore’, said the team leader in the German grid calculation team. Visualising performance metrics accelerated the daily performance review and helped track work progress during staff absences. Sometimes the lean practice was adapted to fit the context of the teams. For example, one of the German teams was located in several cities and struggled to conduct daily performance dialogues. Rather than a physical meeting in front of a whiteboard, the team implemented a digital board that was accessible online and updated during daily conference calls. A challenge for all service shop teams was to select meaningful key performance indicators to track and follow up on. Several teams reported that they found the standard metrics irrelevant for order fulfilment. As for professional services, the interviewees reported that the lean training lacked a relevant business context for them.

Service shop teams mentioned that the prescribed leadership and coaching practices were generally difficult to implement and sustain. In particular, team leaders found more substantial hands-on leadership involvement to contradict the team’s desired work autonomy and employee empowerment. At the same time, team members in the UK asked ‘for a bit more motivation’ and felt that they lacked ‘lean buy-in from top management’. A grid planning team in the UK complained that information on efficiency savings from lean programme implementation was not shared.

4.3 Service factories and fit with lean service practices

Four of six of the service factory teams in our sample are customer care centres (two in Romania, one in Germany and one in the UK). These teams are commonly known as call centres. Modern call centres focusing on productivity have been equated with factories in several reports (e.g. Frenkel and Donoghue 1996; Gilmore 2001; Hudson 2011). They seek to operate based on strict standards and are subject to continuous performance measurement. The work tasks of ‘call agents’ include customer care, complaint management, error handling and customer retention initiatives. Our sample of service factories also includes a workforce planning team in Germany that supports the staffing of a call centre and a UK billing team that performs routine billing of electricity and gas usage.

Service factories report the highest implementation rates of the corporate lean programme (cf. Table 3). This is true for lean operational practices (an average score of 95%), lean managerial practices (84%) and lean cultural practices (91%). Our qualitative data suggest that the service factory teams adopted almost all lean operational practices and achieved high scores on managerial and cultural practices by adapting them slightly to their specific contexts.

Regarding operational practices, all call centre teams report that these practices are already part of their DNA. One team leader in Germany stated, ‘5S and working standards are clearly our strengths... we were already on a good standard... we are actually super at this. We are very well-organised people’. Another manager in one of the call centres explained, ‘We adopted the lean standardised work practices quickly, as we have always operated according to standards’. Creating additional process documentation and user guides provided the call centre teams with increased transparency, improved resource management and faster customer response times. ‘Customer guides are more [of] a consistent approach... everybody is doing the same thing when dealing with the customer’, according to a team leader in a call centre in the UK. The implementation of lean operational practices showed tangible results. For example, the UK-based billing team reported, ‘We had 66,000 exceptions... we completely reviewed our process, developed a customer guide, put timings behind to actually give us targets. In a matter of six months, we were down to 5,000 [quality exceptions]’.

The managerial lean practices are also at a high level but require a higher level of adaptation. In particular, performance measurement practices evoked negative responses. ‘Some of that is difficult because we don’t have the measures (…) sometimes, what worked for one area doesn’t work for another one—but you have to go with it’, reported a team leader in a call centre in the UK. A lean leader shared the following sentiment: ‘Team managers here didn’t really understand what metrics to track on their boards’. Call agents welcome team initiatives as an alternative to their usually isolated jobs because they provide a ‘platform for everyone’s voice to be heard’, as a team leader put it. Transparency of performance metrics and skills matrices enables smoother cross-team communication. We noticed that the daily performance dialogues did not always follow the corporate blueprint. For example, the reporting frequency requested by the lean programme did not align with the availability of metrics (daily meetings discussing week-old metrics are not very efficient).

Cultural lean practices also show high implementation levels, with a certain degree of adaptation. A general problem for teams working in the front office is the availability of quickly solving problems, as prescribed by the lean programme. For example, a team leader in the UK explained, ‘If we know we have a problem, we can’t walk out there right now and grab eight people and go, right, let’s problem solve, because then the customers would be waiting’. In addition, the long incubation time for process improvement conflicts with the service factory’s usual fast-paced work. We also noticed that Energy Co.’s increased focus on performance management in teams fuelled the teams’ competition and comparability, which does not necessarily align well with a lean culture where problems should be surfaced and problem-solving experiments should prevail. Rather, the result in Energy Co.’s service factories was greater individual performance pressure and increased conflict between call agents’ effectiveness and efficiency when answering call queries. As one agent in a team in the UK stated, ‘It felt at times as if we were working against each other, although we were simply taking care of the business’.

5 Discussion

According to one of the team leaders at Energy Co., ‘Lean application requires a lot of empathy... they [corporate] cannot preach the same standards to different types of businesses’. While this makes intuitive sense (cf. Ansari et al. 2010, 2014; Netland and Aspelund 2014), it is of little help to those who teach and support lean across different service types. Exactly how should the programme and practices allow unit-level adaptation? By comparing the within-case findings and drawing on the literature on practice adoption, we develop propositions regarding the fit between standard lean service practices and different service-type contexts.

A first observation is that all teams had reached some level of lean implementation, and none had actively opposed the lean programme. All teams we interviewed mentioned the general benefits of the lean programme, spanning from an improved collaboration climate and a more organised workplace to concrete performance improvement. Hence, we lend support to the branch of the lean service literature that suggests that lean can be implemented in a wide variety of service processes (e.g. Piercy and Rich 2009; Gupta et al. 2016; Liker and Ross 2017; Seddon 2019).

However, we find convincing evidence that several lean practices must be adapted in content or implementation to better fit different service contexts (e.g. Allway and Corbett 2002; Netland and Powell 2016; Tay et al. 2017; Seddon 2019; Rauch et al. 2020). This is as expected, according to contingency theory (Donaldson 1987; Sousa and Voss 2008; Lawrence and Lorsch 2015). At the highest abstraction level, we notice that lean practices have a particularly good fit with the service factory and the lowest fit for professional services. Service shops are somewhere in between professional services and service factories. Energy Co.’s Vice President for Lean Programme Development acknowledged this need: ‘We expect them to customise and tailor the lean practices to fit their context’. Confirmed by our findings, this is a useful advice, but it doesn’t tell when and how. We now turn to the specific propositions developed in this reseach.

5.1 Propositions for operational lean service practices

The set of operational lean service practices demonstrated the highest lean maturity score (91.3%) across all service types and was consistently the highest scoring set of practices within each service type. The scores and case narratives suggest that they have the best fit for service factories compared to service shops and professional services. Building on the service delivery mechanisms listed in the service typology of Ponsignon et al. (2011), we put forward the following three propositions:

Proposition 1a

The higher the employees’ skills and discretion needed in a service process, the lesser the fit of operational lean service practices.

Proposition 1b

The higher the degree of routine work and potential automation of a service process, the better the fit of operational lean service practices.

Proposition 1c

Centralised service processes have a better fit with operational lean service practices than field services.

To our surprise, we noticed that lean operational practices did not necessarily have a better fit with back-office operations than with front-office operations. A common argument is that back-office processes are shielded from the variability introduced by customer interaction and hence can run more like a factory (Safizadeh et al. 2003; Ponsignon et al. 2011). We could not find support for this argument. On the contrary, the back-office operations in our study seem to reject standards. ‘They claim it hinders their creativity’, explained a senior manager in the corporate lean department. We also found that front office operations, such as call centres, had a high degree of adoption and implementation of process optimisation and standardisation. Perhaps because front office services are subject to more significant variation introduced by customer interaction (Chase 1978; Shostack 1987; Johnston and Clark 2008), they benefit from practices that aim to reduce and control this variation (Shah et al. 2008). Standards are ‘a basis for managing uniformity and quality’ (Shostack 1982). Using a script-based, standard approach also controls customers’ expectations and standardises service consumption. A senior manager at headquarters confirmed this finding: ‘Call centres, for example, are more receptive to standards as they are very much process-driven’. Therefore, we propose the following:

Proposition 1d

Operational lean service practices are effective in repetitive front-office operations because they reduce variations in customer interaction processes.

An investigation of contingencies motivates a further proposition. Operational lean practices seem to be adopted more when the service is delivered by a team of agents dependent on each other. When an autonomous individual service agent delivers the service, there is more room for individual variation than when the service is delivered by a team that must be coordinated. To provide an everyday example, in a restaurant, it is more important that the team that serves the table is ‘lean’ than the barista who serves coffee at the bar. Thus, we propose the following:

Proposition 1e

Operational lean service practices fit less when employees have higher work discretion.

5.2 Propositions for managerial lean service practices

Managerial lean service practices generally have a lower implementation rate than operational and cultural lean service practices. The managerial practices discussed in this research relate to performance management. We found that managerial lean service practices seem most effective in services that are centralised, measurable and do not require specialised skills – that is, service factories and service shops. In these settings, managers can understand service delivery and tend to have more management tasks. In professional services, in contrast, the service operator tends to be given greater autonomy and the ability to ‘self-manage’. Especially in field services where service agents often work distantly from their managers, managerial lean practices fit less.

Proposition 2a

The higher the employees’ skills and discretion in a service process, the lesser the fit of managerial lean service practices.

It follows that a higher degree of adaptation of managerial lean practices is needed in professional services. We observed professional service teams that tried to establish online meetings, ‘gemba walks’, and performance reviews. However, these initiatives were hard to sustain for two reasons. First, in professional front-office services, service delivery always takes priority over internal meetings. Second, geographically distributed service delivery involves a host of technical problems, such as, for example, lack of internet connection (in some cases, even lack of telephone signal), no suitable workspace for meetings (often the service car) and scheduling problems. Some of the interviewed field service teams mentioned that the best they could do was telephone meetings in a car park on a bimonthly basis. With no access to the performance board and only a fraction of team members having sufficient telephone reception, it is clear that these performance review meetings are not very practical for professional field services. This discord leads to the deviation of communication style, level of discretion, format, frequency and communication channel of managerial lean service practices. We propose:

Proposition 2b

Managerial lean service practices fit less in distributed field services than in services delivered from a centralised office.

5.3 Propositions for cultural lean service practices

Lean practices aiming to build a culture of continuous improvement are more ambiguous than practices aiming to change operations and management systems. It is not easy for managers to know exactly what, for example, ‘coaching’ and ‘empowerment’ mean. Since these practices can take different forms, they generally fit all service types. With an overall lean maturity assessment score of 86.2%, cultural lean practices appear to fit any service setting.

However, a deeper look at the qualitative data reveals that creating a culture of continuous improvement is immensely difficult for all service types. For example, all teams struggle to institutionalise the leadership behaviours needed for effective problem-solving sessions (see also Tortorella et al. 2019; Vanichchinchai 2021) – but seemingly for different reasons. Because service factories are under high-performance pressure, there is little time for structured problem solving with root cause elimination. A lean coach in a German call centre observed, ‘They [the call centre] solve the problem, but do not see the need to record the solution and learnings’. In professional services, it is hard to get operators or teams together due to distributed fieldwork. In addition, autonomous service agents experience a reduced need for a coordinated and uniform lean culture. In service shops, we observed that leaders and employees often avoided problem-solving sessions because they did not see a need to improve. In back-office teams, we observed that team members were reluctant to give up autonomy and power related to tacit knowledge.

Proposition 3a

The higher the employees’ skills and discretion in a service process, the lesser the fit of cultural lean service practices.

Proposition 3b

The lower the degree of routine work and potential of automation of a service process, the lesser the fit of cultural lean service practices.

Whether or not the service is delivered in the field or in an office seemed to play a particular role for cultural lean service practices. When professional services are delivered in the field, the culture-building leadership practice of ‘go and see’ (‘gemba walks’) is perceived as impractical. One team member in a field service unit shared the following sentiment: ‘It is a bit tongue-in-cheek because I totally get the value of the go and see, but it takes three to four hours every time, and the set plan is to do a go and see three times a week’. We propose:

Proposition 3c

Cultural lean service practices fit better with centralised service processes than distributed (field) services.

6 Conclusions

We investigated the fit between lean service practices and three different service types: professional services, service shops and service factories. We concluded that standard lean service practices do not have the same fit across different service types and require different levels of adaptation. We summarised our findings as ten propositions.

6.1 Contributions to research

The results contribute new insights to the growing literature on lean services (e.g. Leite and Vieira 2015; Gupta et al. 2016; Hadid et al. 2016; Liker and Ross 2017; Seddon 2019; Rauch et al. 2020; Fenner et al. 2022). Instead of researching lean services at the industry level, which most studies do, we contribute a process-based perspective on lean services. In particular, this research demonstrates that studying a general lean service perspective across many different service types is an inaccurate simplification. Even when studying lean services in one industry or company, researchers must pay attention to the important nuances introduced by different service types. Overall, our findings draw into question the validity of concepts that generalise lean to an industry, such as ‘lean healthcare’, ‘lean education’, ‘lean policing’, ‘lean utility’ and so on.

Based on the service process typology by Ponsignon et al. (2011) and the concept of fit, we elaborate on how standard lean service practices must be adapted to three different service-type contexts: professional services, service shops and service factories. The propositions suggest how underlying mechanisms affect the fit level across these service types. We offer these propositions as promising avenues for future research on lean services.

6.2 Contributions to practice

Managers in service companies, as well as in manufacturing and governmental institutions, can use the findings to inform the lean journeys of their service units. We find that lean practices can be implemented effectively in a range of service types, but that the practices must be adapted to fit the service context. In particular, managerial lean practices require a higher level of adaptation in all service types.

More adaptation of lean is required when the service context is characterised by the following: higher employee skills and discretion, lower levels of routineness and potential for automation when the service occurs in the field and the front office. Conversely, standard lean service practices tend to be adopted when lower skills and discretion are needed in the service delivery process, when the routineness and potential of automation are higher and when the service takes place centrally in the back office. Hence, lean service practices have a higher fit with service factories than with professional services. Managers can use this insight to create different action plans and set different expectations for different service units.

We submit that a higher level of adaptation in some contexts should not discourage the implementation of lean in these service types. Because it is more ‘difficult’ to implement lean in these settings, it can provide a potent source of competitive advantage. For example, we found that front-office services that introduce standards reduce process variation in customer contact, which increases process cycle times and improves communication and coordination in service delivery teams.

6.3 Limitations and future research

This paper has the usual limitations of qualitative case-based research. We cannot claim general validity of the suggested propositions; therefore, we offer them only for future scrutiny and testing. Furthermore, we cannot rule out the possibility that some of the results were driven by contextual variables other than service type. Team size, managerial commitment, history and network integration are examples of other variables that affect lean adoption. Moreover, national and micro-organisational cultures can affect the implementation and fit of lean practices. We mitigate these limitations by conducting a deep case study in a single firm, which arguably keeps many confounding variables nearly constant: all units had the same lean programme, comparable senior management commitment and resource distribution, and were subject to the same industry dynamics. We encourage future research to use cross-sectional data or multiple case studies to investigate how contingencies affect lean implementation in different service processes.

Another limitation relates to the choice of service typology. The literature offers no commonly accepted typology, and no typology perfectly caters to all different service contexts. By sticking to the convention in the literature – specifically, building on the service typology of professional services, service shops and service factories (Ponsignon et al. 2011) – instead of describing the specificities of the service teams we study, we trade internal validity against greater external validity.

Relatedly, as we were working with one company, we were restricted to its available service types. For example, all our professional service teams were working in the field. These field services employ staff with higher seniority and experience than other service types. Moreover, they were organised into separate units with more autonomy from headquarters. It is possible that an office-based professional services service (e.g. specialised software development or consulting) with a younger staff would have led us to other propositions regarding professional services. Future research could draw on quantitative methods to explore such questions.