Servitization: Disentangling the impact of service business model innovation on manufacturing firm performance
Summary (3 min read)
1. Introduction
- Increasingly, durable goods manufacturers choose to innovate their offerings by providing services to accompany their existing products throughout the life cycle.
- The firm under study, Atlas Copco Compressor Technique (referred to in the remainder of the text as Atlas Copco), is a global manufacturer of durable industrial equipment with a worldwide network of country sales-and-service subsidiaries.
- With more than 130 years of experience in product innovation, Atlas Copco has, in recent decades, extended its innovation trajectory into services.
- In revealing the nature of value creation and appropriation, their study not only yields one of the first theoretical underpinnings of servitization but also demonstrates how product firms can achieve revenue growth and profitability by engaging in services.
2. Literature review
- Service research has been a growing area of operations management research (Chase and Apte, 2007; Heineke and Davis, 2007).
- While the evidence regarding a manufacturer’s ability to appropriate value from servitization is inconclusive, the evidence on the ability of servitization to create value on the level of the product and the customer is more positive.
- The addition of services such as maintenance, upgrading and remanufacturing prolongs product life and so reduces product turnover (Mont, 2004b).
- This set of arguments points to high potential for services to act as a complement to products, offsetting the substitution effect; subsequently, higher service revenues will result in higher product revenues.
- Interplay between the arguments presented above is likely to result in a curvilinear relationship between service scale and profitability where investments paired with a low scale of services initially result in a decrease in margin but then convert to a positive impact as the investment is internalized and higher scale is reached (Fang et al., 2008; Suarez et al., 2011).
4.1 Research design
- To test the hypotheses, the authors collected data from the national sales-and-service subsidiaries of a large multinational equipment manufacturer, Atlas Copco.
- Over time, a given country subsidiary may have sold tens of thousands of equipment units to thousands of clients.
- In recent years, Atlas Copco has concentrated on promoting a service offering that covers related machinery, aiming to improve reliability and reduce energy costs for the entire functional group of products.
- Regarding the level of decentralization and subsequent diversity of subsidiaries, subsidiaries offer homogeneity with respect to the product and service portfolio, brands and pricing, on the one hand, while each subsidiary differs significantly in terms of its organizational structure, local practices, management style, and deployment of its business model, on the other.
- “The ability to get closer to theoretical constructs is particularly important in the context of longitudinal research that tries to unravel the underlying dynamics of phenomena that play out over time” (Siggelkow, 2007).
4.2. Dependent, Independent and Instrumental variables
- As per H1a, the installed product base enables the sale of services in the following year, after expiry of the obligatory warranty.
- In turn, services imply greater interaction with customers, which may result in additional sales of products, almost simultaneously with the provision of services (H1b).
- The instrumental variables approach has been used as a popular resolution of the endogeneity problem in operations and strategic management, particularly when coupled with estimators such as GMM and two-stage least square (Guajardo et al., 2011; Novak and Stern, 2009; Suarez et al., 2011).
- As Atlas Copco allows cross-border sales of products but not services (e.g. the German subsidiary can sell products in France despite the presence of a French subsidiary), the authors identified the macroeconomic indicator of country exports as an adequate product sales instrument (H1a).
- Firstly, the authors complemented the existing instruments for service sales (country imports, service coverage, lagged product installed base and manufacturing capex) with country population density, given that labor-intensive services thrive in densely populated areas.
4.3 Control variables
- Due to the scarcity of data and complex nature of interdependencies between products and services, the choice of control variables represented a particularly important part of the model design.
- As all subsidiaries have been operating for 15 years or more prior to the observation period, the subsidiary age was deemed irrelevant.
- Subsidiaries have also exhibited homogeneity with respect to the product portfolio, brand and high-level pricing strategy (allowing a modest degree of discretion for individual client negotiations).
- Furthermore, the authors capture differences in market development by GNP per capita (e.g. subsidiaries operate in countries ranging from China to Switzerland), while a year dummy variable is used to capture the expected yearly effects of price increases.
- Table 1 contains further information on the variables, while Models M1 (H1a), M2 (H1b & H1c) and M3 (H2) are formally represented below.
4.4. Estimators, model corrections and diagnostic checks
- In all econometric models, the authors use panel data analysis with fixed effects.
- The authors introduce fixed effects to control for time-invariant, unobserved heterogeneity among subsidiaries, given their expectation that time-constant differences may determine the effectiveness of the service strategy (Greene, 2003).
- For service staff and consequently non-service staff, data was missing for nine subsidiaries, and on four years in the case of one additional subsidiary.
- This choice was also in line with Suarez et al.’s (2012) choice of GMM to test the impact of servitization on performance.
- In addition, instruments are jointly exogenous for each of the models, given that the P-value of Hansen’s J statistic exceeds 10%, thereby rejecting endogeneity in instruments (Hansen, 1982).
5. Results
- Table 3 below summarizes the results obtained in relation to the testing of the hypotheses.
- Interestingly, service sales has a greater impact on product sales too.
- Further to the relationships hypothesized, their models demonstrate a strong and significant influence of country development, as measured by the GNP/capita.
6.1. Study results and the contribution to theory and practice
- The results of Model 1 (H1a) and Model 2 (H1b and H1c) indicate that product sales and service sales complement each other and that the customer proximity of service offerings reinforces the positive feedback from services to product sales.
- While positive effects are anticipated in progressing from products to related service activities, the reverse relationship – whereby service sales positively influence product sales – is far less obvious, since the impact of servicing may be negative when services act as substitutes for products.
- In practice, this phase is marked by a high proportion of spare-parts activity and a handful of service contracts for high-paying customers, where attractive margins can be achieved without substantial investment in staff and organization.
- The authors study also contributes to the literature on servitization by disentangling the complex relationship between servitization and profitability.
6.2 Limitations and future research
- The authors are fully aware of the limitations of the research reported in this paper.
- Firstly, the authors have focused only on one mother firm with one business model – the integrated product-service business model – whereas servitization can equally imply the deployment of less related service offerings.
- This research should be complemented by similar efforts in different industries to assess the broader validity of the findings obtained.
- Furthermore, it should be noted that, while this study focuses on sales complementarities, engaging in servicing can also yield spillovers by generating insights that have a considerable impact on product development activities.
- Assessing the presence of complementarities and substitution effects over longer time periods would add significant value to the results reported in this paper.
Did you find this useful? Give us your feedback
Citations
1,201 citations
Additional excerpts
...BMI and organizational implications/ performance Survey data Zott and Amit (2007), Giesen et al....
[...]
1,168 citations
688 citations
490 citations
479 citations
References
46,648 citations
"Servitization: Disentangling the im..." refers background in this paper
...The necessary investments range from ‘operational’ service capabilities and resources such as service delivery, service sales skills (Barney, 1991; Peteraf, 1993), and service information systems and tools (Penttinen and Palmer, 2007), to more dynamic capabilities enabling service deployment (Teece…...
[...]
28,298 citations
27,902 citations
"Servitization: Disentangling the im..." refers background in this paper
...…Peteraf, 1993), and service information systems and tools (Penttinen and Palmer, 2007), to more dynamic capabilities enabling service deployment (Teece et al., 1997) such as service management and top management capabilities to reorganize a manufacturing firm from pure product provider to…...
[...]
...The ecessary investments range from ‘operational’ service capabilties and resources such as service delivery, service sales skills Barney, 1991; Peteraf, 1993), and service information systems nd tools (Penttinen and Palmer, 2007), to more dynamic capailities enabling service deployment (Teece et al., 1997) such as ervice management and top management capabilities to reoranize a manufacturing firm from pure product provider to roduct-service provider....
[...]
26,580 citations
"Servitization: Disentangling the im..." refers background or methods in this paper
...In the model specification, we also use ‘heteroskedasticity andautocorrelation-consistent’ (HEC) standard errors (Bascle, 2008) to avoid concern over invalid inferences caused by these two violations of standard OLS assumptions (Arellano and Bond, 1991)....
[...]
...In the model specification, we also use ‘heteroskedasticity and-autocorrelationconsistent’ (HEC) standard errors (Bascle, 2008) to avoid concern over invalid inferences caused by these two violations of standard OLS assumptions (Arellano and Bond, 1991)....
[...]
...A number of contributions have examined effective ways of (commercially) engaging in extended warranty and after-sales service (Balachander, 2001; Balachandran and Radhakrishnan, 2005; Jack and Murthy, 2001, 2007; Patankar and Mitra, 1995). For example, Cohen et al. (2006) developed a product life-cycle model that studies a set of strategic choices manufacturers face as they design their joint product/service bundle – requiring, in all likelihood, after-sales maintenance and repair support....
[...]