DTD 5
Journal of Operations Management xxx (2005) xxx–xxx www.elsevier.com/locate/dsw
Achieving competitive advantage through implementing a replicable management standard: Installing and using ISO 9000 Eitan Naveh a,*, Alfred Marcus b,1 a
Technion–Israel Institute of Technology, Faculty of Industrial Engineering and Management, Haifa 32000, Israel b University of Minnesota, Carlson School of Management, Minneapolis, MN 55455, USA Received 1 December 2002; received in revised form 1 January 2005; accepted 28 January 2005
Abstract This paper investigates the effects of implementing the international standard ISO 9000 on measures of business and operating performance. Based on the literature and a case study we identified two stages in implementing ISO 9000—(1) Installation, which has two dimensions: (a) external coordination and (b) integration; and (2) Usage, which also has two dimensions: (a) in daily practice and (b) as a catalyst for change. The hypotheses were that installation of ISO 9000 is positively related to use of ISO 9000, and use of ISO 9000 is positively related to operating performance. In addition, use of ISO 9000 is positively related to business performance since operating performance is positively related to business performance. We used hierarchical linear models (HLM) to test our hypotheses and validated the results by comparing the longitudinal performance of ISO 9000 certified companies with four matched samples of companies that were not ISO 9000 certified. Our analysis indicated that while the installation stage was necessary to successfully implement ISO 9000, organizations achieved a distinct operating advantage from this replicable standard when they used it in daily practice and as a catalyst for change. These findings were based on responses to a survey of 1150 quality managers in 924 organizations, which was supplemented for about one-third of the organizations with longitudinal information from the Compustat database on the organizations’ business and operating performance. The validation indicated that implementing the ISO 9000 standard led to improved operating performance, but that this outcome did not necessarily or automatically yield better business performance. # 2005 Elsevier B.V. All rights reserved. Keywords: ISO 9000 standard; Operating and business performance; Usage; Installation
1. Introduction * Corresponding author. Tel.: +972 4 829 4527. E-mail addresses:
[email protected] (E. Naveh),
[email protected] (A. Marcus). 1 Tel.: +612 624 2812.
The International Standards Organizations in Geneva established the ISO 9000 standard in 1987. Its origins, however, rest in earlier American (1960s— MIL-Q-9858A and MIL-I 45208A) and British
0272-6963/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jom.2005.01.004
DTD 5
2
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
standards (BS 5750). By the year 2000, more than 100,000 organizations worldwide were ISO 9000 certified and over 100 countries had adopted this standard (The International Standards Organization, 2000). Management standards, such as ISO 9000, are a subset of a broader category of voluntary management practices, often referred to as administrative innovations (Brunsson et al., 2000). The ISO 9000 standard requires that firms have in-house standardized and replicable routines and procedures for product design, manufacture, delivery, service and support, but whether implementation of such a standard is helpful to organizations is not known with certainty (Anderson et al., 1999; Greve, 1999). Organizational analysts have identified implementation failures as a main reason that organizations have not benefited from practices such as ISO 9000 (Nutt, 1986; Klien and Sorra, 1996). Abrahamson (1991, 1996, 1997) sees the process of implementation as akin to fads and fashions; managers implement such practices in the same ways that people make decisions about the length of a skirt or the width of a tie—as a consequence of being persuaded by glitzy rhetoric that exploits their vulnerabilities and psychological needs. According to Abrahamson (1991, 1996, 1997), implementation is a social construction that does not narrow real performance gaps. Nevertheless, Collis and Montgomery (1997) suggest that the implementation of practices such as ISO 9000 can raise organizational performance and result in real competitive advantage. Quality standards, such as ISO 9000, have continually evolved and been refined over time and are supposed to offer a competitive factor in many industries (Cole, 1999). However, since knowledge of ISO 9000 is public (Matusik and Hill, 1998) and implementation of the practice aims to create greater homogeneity among organizations (Brunsson et al., 2000), it is debatable if ISO 9000 constitutes a competitive factor. As Barney (1986) maintains, competitive advantage arises from the differences among firms (heterogeneity), but if all companies implement ISO 9000 in the same way, can a particular company derive a special benefit? We argue that competitive advantage can be gained from implementing a replicable management standard, such as ISO 9000, if implementation is understood not as a discrete and homogenous industry-wide phenomenon, but if variations in this
process are considered (Lewis and Seibold, 1993; Westphal et al., 1997; Meyer and Goes, 1988). For our study, we built a model for implementing ISO 9000, which takes into account company-specific variations. We used information from a survey of 1150 managers as well as data collected from the Compustat database about organizations that are ISO 9000 certified and four matching samples of non-ISO 9000 certified organizations whose characteristics otherwise match the ISO 9000 certified organizations. We then tested this model to determine the extent to which variations in implementation have an effect on operating and business performance.
2. ISO 9000s effect on organizational performance The issue with which we deal in this paper is whether ISO 9000 can have a positive effect on an organization’s competitive performance. We deal with this question because there is a considerable debate in the literature as to whether ISO 9000 has a positive impact. In 1996, Dun and Bradstreet conducted a comprehensive survey (ISO 9000 survey, 1996) of all ISO 9000 certified U.S. organizations and found that the average company benefited from implementation. Nonetheless, Terziovski et al. (1997) rejected the proposition that a significant relationship between ISO 9000 implementation and organizational performance exists (also see Curkovic and Handfield, 1996). Batchelor (1992) found that the benefits of implementation were procedural efficiency and error rate reduction and not market share, staff motivation, or cost reduction. Allan (1993) and Brown (1994) found that managers went back to ‘‘fire fighting’’ after they implemented ISO 9000 and their commitment to planning, prevention and continuous improvement was short-lived. Robin and Dennis (1994) found that ISO 9000 had a positive effect because of the introduction of statistical process control, though the standard itself did not explicitly require it, and Askey and Dala (1994) claimed that managers saw certification as end in itself, not as a means to an end, and that after certification they reverted to old practices. Implementation of ISO 9000 should make an organization’s operating performance go up, but consistent results that could verify this proposition
DTD 5
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
3
Fig. 1. Implementation of a replicable management standard.
are not found in the literature.1 Superior operating performance denotes lower defect rates, reduced cost of quality, higher productivity, on-time delivery and customer satisfaction (Flynn et al., 1994). With better operational performance, the products or services the organization offers should become more attractive to customers and the firm should have better business performance. Sales and profitability should increase (Heskett et al., 1997). Among studies that support the benefit of implementing ISO 9000 are: Basil (1994), Lawrence (1996), Eugenia (1996), Edmunds (1996), Wade (1996), Dov (1998), Smith (1994), Houten (1994), Donna (1995) and Avery (1995). Elmuti and Kathawala (1997), for instance, claimed support for the argument that productivity, quality of product, and quality of work life improved due to certification. Brown and Loughton (1998) found benefits of greater quality awareness, improved awareness of problems and better product quality, but did not see improvements in productivity, costs, wastage rates, staff motivation and staff retention. However, overall the literature has not conclusively demonstrated that ISO 9000 produces these benefits (see Hunt, 1997, which
concludes that ISO 9000 did not necessarily guarantee product quality).
3. Modeling the effects of implementing ISO 9000 on organizational performance Our model of ISO 9000s effects on an organization’s competitive performance (see Fig. 1) is based on the literature and a case study. The model has two stages—installation and usage, with each stage having two dimensions. For installation, the dimensions are (i) internal integration and (ii) external coordination, and for usage they are (i) use in daily practice and (ii) use as a catalyst for change. We argue that installation is related to usage (Venkatraman, 1989). Usage in turn, is related to operating and business performance, with the relationship to business performance being partially mediated by operating performance. In what follows, we discuss the stages in this model and illustrate them with a case study. 3.1. Installation: external coordination and internal integration
1 Keynes (1994) found that small firms gained no advantage from ISO 9000 certification over their non-certified competitors, the International Quality Study concluded that certification did not have a significant effect performance (Powell, 1995), and Highlands (1995) questioned whether ISO 9000 was worthwhile based on the relationship between cost, time, effort, paper work and customer needs. Larson (1997) questioned top management’s commitment and wondered if certification was necessary.
Installation consists of concept development and preparation, which must occur before a system such as ISO 9000 can be effectively used. By installation (Nutt, 1986, 1989), we mean establishing rules that allow an organization to effectively adhere to ISO 9000 standards. These rules reside in quality policies
DTD 5
4
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
and manuals. They include definitions of responsibilities and involve documentation, inspection, calibration, testing, data collection and analysis that show how the organization takes corrective action/s and what its internal auditing plans are. While there are differences between implementation of a management standard and adoption of new technologies, the planning–doing tension (Argote, 1999; Crossan et al., 1999; also see Lewin, 1952; Zmud and Apple, 1989; Cooper and Zmud, 1990) applies to both. Installation is both the attempt to set a management standard in place and the form of planning that occurs before the standard is used. It deals with introducing the standard, establishing a system for carrying it out and developing a set of rules for how it will be applied. Installation has many aspects (Meyer and Goes, 1988; Campbell, 1989; Attewell, 1992; Crossan et al., 1999; Cooper and Zmud, 1990; Fichman and Kemerer, 1997; Daft, 1982; Ettlie, 1980; Ettlie and Vallenga, 1979; Beyer and Trice, 1978). They can be categorized along two dimensions: external coordination with customers and suppliers so that the requirements of the standard fit the needs of critical stakeholders; and internal integration, or customizing and integrating the standard with the organization’s current stock of assets, that is, its existing practices (Dierickx and Cool, 1989). ISO 9000 is generally externally induced, in that for many companies it is a condition for their doing business. Many customers require that their suppliers be certified, thus the pressure that customers exert is a reason for certification (Anderson et al., 1999). By external coordination, we mean that ISO 9000s design and development is harmonized with the needs and expectations of customers and suppliers. With supply chains increasingly critical to an organization’s success, external coordination takes on growing importance (Eisenhardt and Tabrizi, 1995; Lengnick-Hall, 1996). ISO 9000 certification provides assurances to customers that suppliers have a system of quality assurance in place. Internal integration of a standard, such as ISO 9000, means designing and developing new systems to conform to the standard based on an analysis of a company’s existing internal processes. It requires integrating the standard with practices already in place. When a new practice such as ISO 9000 is introduced, the organization must find a fit between
the ISO 9000 rules and its old ways of operating (Meyer and Goes, 1988). The old ways represent past knowledge that was used to meet prior needs (Matusik and Hill, 1998). Dierickx and Cool (1989) compared the introduction of the new practice to a ‘‘flow’’ of a new liquid into a ‘‘tub’’ already full of liquid. Consistency must prevail between the new practice and existing practices (Brunsson et al., 2000; 127). Links have to be established between the organization’s old policies, rules and procedures, and the new ones (March et al., 2000; 64). For example, a full-scale organizational system, including procedures for training and employee education, has to be added and made consistent with the organization’ previous practices (Peter, 1995). A standard such as ISO 9000 must be customized to the company’s needs, an activity best led by employees trained and nurtured in the company, as opposed to one led by outside consultants (Dierickx and Cool, 1989; Westphal et al., 1997; Attewell, 1992). 3.2. Usage in daily practice and as a catalyst for change Installation is not enough, however, for a quality system such as ISO 9000 to have a long-term effect on an organizations performance (Nutt, 1986). The system has to be used, on the one hand, in daily practice, and, on the other, as a catalyst for change. We derive these dimensions from the literature on rules and routines (Nelson and Winter, 1982; March et al., 2000) and on organizational strategy, especially the literature that discusses the complementary nature of structure and flexibility (Brown and Eisenhardt, 1998). Daily practice is regular application of a standard (Fichman and Kemerer, 1997; Ettlie, 1980) to the point that it becomes part of an organization’s everyday routines (Cooper and Zmud, 1990). Employees must consider it necessary and pay attention to it. An indicator of their daily use of the standard is that they do not prepare for external audits at the last minute. ISO 9000 requires that twice yearly independent, third party auditors evaluate a company’s procedures and carry out site visits to determine the extent to which it is being used. In the U.S., the Registrar Accreditation Board (RAB) accredits third party auditors and the American National Standards Institute (ANSI) and the American Society of Quality
DTD 5
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
(ASQ) administer the audits. If employees just put on a show for the auditors and revert back to earlier noncompliance, it suggests that they are not really using the standard. Daily use, however, is just one step. By itself, it is not sufficient for improvement. Braley (1994), Hoyle (1994), Frehr (1997) and Herb (1998), for example, suggest that daily use is a beginning stage in the effort to systematically enhance an organization’s quality. Juran (1993) also asserted that daily use was not equivalent to quality achievement; an organization has to go beyond this. The results of a comprehensive survey of manufacturing firms (U.S. Industry Week, 1998) concluded that certification could be an excellent means to track and assess quality processes, build a framework for quality and impress customers, but by itself did not result in improved quality. Though ISO 9000 has extensive formal requirements (e.g., see Bagchi, 1996) and routine use is demanding, there is another way it can be used. ISO 9000 can and should become a springboard for rethinking the way a company does business and a point of departure for additional innovation. Being a catalyst for change means that ISO 9000 is used as a launching pad for new understanding about how a company does business. ISO 9000s rules become the basis for branching out, expanding and moving in new directions (Brown and Eisenhardt, 1995; Van de Ven et al., 1999; Mirvis, 1998). Employees extemporaneously paraphrase, embellish and reassemble what the rules require (Sutcliffe and Sitkin, 1996). Successful implementation of a management practice, then, depends on more than its routinization. As Follett (1995) argued, without independent thinking and initiative, there is likely to be ongoing resistance to implementation (also see Guth and Macmillan, 1986). Van Meter and Van Horn (1975) and Edwards (1980) also maintained that independent thinking and initiative at the individual and organizational levels positively affect implementers’ dispositions and overcomes ‘‘routine’’ and ‘‘mechanical operationalization’’ (Fidler and Johnson, 1984; 704), which sabotage implementation. 3.3. A case study To illustrate these concepts, we draw on a case study we conducted. To preserve confidentiality, we gave the company a fictitious name—Telecommuni-
5
cations International Company (TIC). TIC is a leading global company with products installed in many countries and customers among the world’s largest network operators. Prior to doing the case study, we gathered and read publicly available information about it. During the site visit at headquarters, we heard a general presentation about the company from a member of the top management team. Then, we interviewed 12 people serving in different positions, such as corporate quality management, purchasing, production, marketing and software development. The interview guide had open-ended questions about why the company became ISO 9000 certified, how the certification process evolved, and how ISO 9000 fit in with the company’s other quality programs. While the interviews were taped, we also took handwritten notes. On-site we collected additional documents and discussed our impressions with company officials. We had a number of off-site debriefings with members of the project team. The published material, interview transcripts and notes, documents, discussions and debriefings provided the basis for a case study we wrote. We searched for consensus views among those with whom we had been in contact, some of whom had been at TIC for a long time and some who had not, but nevertheless, had a good understanding of TIC’s history and culture (Marcus and Nichols, 1999). The case study deals with the company’s business strategy and quality programs other than ISO 9000. The parts excerpted here show how ISO 9000 was installed and used. 3.3.1. Installation of ISO 9000: external coordination and internal integration Respondents at TIC said that the push for certification was externally driven. Since the decision to apply for ISO 9000 certification originated in customer demands, TIC consulted with customers to make certain that the system would satisfy their needs. The effort to become certified took about a year, and TIC became accredited only after coordinating its system with suppliers and customers. Employees wrote numerous procedures to govern relationships with suppliers and customers. Coordination involved the collection of performance data, documenting methods, calibrating equipment and developing common packaging. TIC collected and used data its customers and suppliers provided. Customers and suppliers participated in equipment tests and helped
DTD 5
6
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
establish test specifications. TIC considered the future plans of customers and suppliers in its planning. Before implementation, TIC had a number of other quality initiatives in place (for example, TQM). These had to be made consistent with ISO 9000 rules. The company had to redefine and rework procedures to harmonize them with ISO 9000. Its managers had to integrate old procedures with ISO 9000 mandates. Prior to certification, it had started to build a quality assurance system where reliability was the main goal. Using field data, employees calculated and made predictions about reliability. Before certification too, employees had operated on the basis of many tacit processes and procedures. Much of what they did they ‘‘knew by heart.’’ To become certified, TIC had to put old procedures down into writing and develop new ones. To develop these procedures, interdisciplinary teams from many parts of the company analyzed and learned about the processes that employees used. The purpose of recording the practices was to understand them better. The company not only had to make ISO 9000 consistent with its existing practices but also it used the opportunity to try to improve them. 3.3.2. Usage in daily practice and as a catalyst for change TIC’s first attempt to become certified led to a 25page auditors’ report that showed deficiencies in implementation. As a result, TIC had to take many corrective actions to satisfy the auditors. The auditors required that TIC’s standard operating procedures be incorporated into training, with top management held accountable. After certification, employees had to use the new procedures ISO 9000 mandated to help organize the work they did. Employees found the procedures helpful because they were readable, clear and organized. An additional benefit from implementation was that the procedures helped intra-company communication. TIC employees thereby saved time and money. Nonetheless, they were critical of some aspects of ISO 9000. For instance, they criticized the method for discovering and eliminating defects, which they felt was passive, because it meant eliminating problems after-the-fact. The auditors told them that they had to use this system regardless of what they thought. Employees had to identify and eliminate defects, an awareness of which was an impetus for introducing improvements.
Nevertheless, just being certified and using the procedures that ISO 9000 mandated, TIC employees reported, was not enough if the company was going to acquire the reputation for innovation and quality it was seeking. They had to go beyond what ISO 9000 required and TIC employees told us that ISO 9000 was the springboard for the development of many additional quality initiatives. After the company became certified, TIC introduced innovations built on ISO 9000s foundation. TIC managers introduced 6 quality initiatives before ISO 9000 and 25 after certification. After certification, the company developed a work plan and method for testing and control in product development, procedures on how to start a project, who should be involved, what steps should be taken to meet milestones and how the company should work with subcontractors. It moved into the area of statistical process control (SPC), introduced customer satisfaction surveys, paid more attention to customer feedback and created customer call centers. It reduced the number of people devoted solely to quality control and assigned more of its quality staff to support marketing and development. The company concluded that 20–25 key metrics influenced its operations and business success, and accordingly, developed a system for quarterly management reviews based on these metrics. TIC also became registered to the international environmental standard ISO 14000, started an initiative in knowledge management, and won a quality award similar to the Baldrige. TIC, thus, did not stop with ISO 9000. In the words of various TIC managers, ISO 9000 was the ‘‘trigger’’ or ‘‘springboard’’ that launched these developments. Without it, the company would not have advanced as far in building a system for quality. TIC kept developing and adding layers to its overall quality assurance system so that, in the end, it had a very complex system of which ISO 9000 was only a part. Without the influence of ISO 9000, these managers believed that TIC would not have advanced as far as it did in the area of quality. 3.4. Hypotheses When installed and used as it was at TIC, we argue that ISO 9000 should lead to better operating and business performance. Such company-specific variations and extensions in implementation can potentially
DTD 5
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
create a competitive advantage. A competitive advantage will be revealed in better operating performance as expressed in such measures as lower defect rates and a reduced cost of quality. It will be demonstrated by on-time delivery and customer satisfaction, and involve productivity improvements, for as Garvin (1988) maintains, quality and productivity are related. More time would be devoted to manufacturing acceptable products; there would be less rework, less scrap and fewer wasted materials. Quality management as well as operations theories suggest that better operating performance should lead to another component of competitive advantage, better business performance (Hackman and Wageman, 1995). Less rework and higher productivity should result in lower expenses, which would translate into higher gross profit margins. With lower defect rates and on-time delivery, sales should grow because new markets are created and customer retention is high (Baldrige, 2003). However, if ISO 9000 is not installed and used in the way described above, better operating and business performance will not come about. In other words, the potential for gaining competitive advantage from the implementation of a replicable standard does not exist. Thus, we hypothesize that under the following conditions ISO 9000 has the potential to yield competitive advantage: Hypothesis 1. Use of ISO 9000 (in daily practice and as a catalyst for change) is positively related to operating performance. Hypothesis 2. Installation of ISO 9000 (external coordination and internal integration) is positively related to use of ISO 9000 (in daily practice and as a catalyst for change). Hypothesis 3. Use of ISO 9000 (in daily practice and as a catalyst for change) is positively related to business performance. Hypothesis 4. Operating performance is positively related to business performance. In our model, operating performance is an intervening mechanism or mediator between usage and business performance. Mediation means that it is not usage that leads to better business performance,
7
but usage that leads to better operating performance and better operating performance leads to better business performance. Our argument is that the mediation of operating performance is partial, as usage is likely to have some direct effects as well on business performance.
4. A test of the model To test the model, we conducted a survey of ISO 9000 certified companies and analyzed the way variations in implementations affected performance. To validate the ISO 9000 survey results, we matched the ISO 9000 certified companies, for which we had Compustat data, with four control samples of companies that were not ISO 9000 certified. 4.1. Sample selection The newsletter Quality System Update published by McGraw Hill routinely collects lists of all facilities in the U.S. that have demonstrated through external auditing that they have implemented ISO 9000. We sent a postcard to the ISO 9000 management representatives, who appeared in this list, asking this person to participate in the study. The management representatives were given a code that gave them entry to an Internet site where our questionnaire could be completed (full confidentiality was promised). For the typical respondent, the questionnaire took about 25 min to complete. About a month after the first mailed postcard, we sent a second postcard to nonresponders. At the time of the survey there was a total of 4233 ISO 9000 certified organizations in North America. Depending on an organization’s decision, registration could be made at the organizational level or at the facility level. In cases where it was made at the facility level, two or three registrations per organization, each with a different management representative, were common. In organizations with more than one registration, we sent a separate postcard to each of the facility level management representatives. They were asked to reply to the questionnaire with regard to the certification of which they were in charge, and not to the organization as a whole, since we assumed that they were more familiar with the implementation of
DTD 5
8
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
Table 1 Response rate ISO 9000 certified population (number of organizations)
Survey respondents (number of organizations)
Response rate (percent)
Organizations
4233
924
21.8
Number of registrations—different management representative 1 2 3 4 and more
3612 1020 (510) 225 (75) 214 (36)
728 322 (171) 48 (17) 52 (8)
20.2 31.6 21.3 24.3
Total
5071
1150
22.7
Organizations with publicly available (Compustat) data
1454
313
21.5
Number of registrations—different management representative 1 2 3 4 and more
1154 394 (197) 210 (70) 200 (33)
249 76 (41) 42 (16) 48 (7)
21.6 19.3 20 24
Total
1958
415
21.2
which they were in charge. For this reason, we received two or more responses from some organizations. In sum, out of the total number of organizations in North America with ISO 9000 registrations, 1150 managers from 924 organizations completed the survey, which represents a response rate of greater than 20 percent (see Table 1). Among surveys of quality managers, this response rate is typical. Published empirical work by Flynn et al. (1994), Powell (1995) and Douglas and Judge (2001) draws conclusions from surveys with similar response rates. About one-third or 313 of the organizations that responded to the questionnaire had publicly available data in the financial dataset Compustat, which we used to augment the results. Organizations that responded to the survey were matched with the Compustat dataset using the responders’ Internet site entry code. The responding organizations’ industry breakdown includes six industry types. These are: (1) agriculture, forestry and mining (SIC codes 1–14), (2) services, trades and construction (SIC codes 15–17 and >40), (3) food, tobacco, textile and wood products (SIC codes 22–26), (4) petrochemicals and plastics (SIC codes 28–30), (5) manufacturing, metals and machinery (SIC codes 32–35, and 37 and 39) and (6) electronics and instruments (SIC codes 36 and 38). Seventy-five percent of the respondents were manu-
facturing organizations and 25 percent were service organizations. The largest two sectors in terms of number of respondents were manufacturing, metals and machinery, and electronics and instruments, from which 60 percent responded. Because of a concern that survey respondents might have a different profile than non-respondents, we divided respondents by industry (using SIC codes), time since certification, and size, and compared the percentage of those who responded in these groups with their actual registrants. The numbers were close enough for us to believe that those who responded came from similar groups as the population as a whole. In addition, we sent mailed questionnaires to a sample of 50 non-responding organizations, and followed up with telephone calls until we obtained responses from nearly all 50. As the ability to generalize the findings of this study rests in part on the extent to which the sample was free from non-response bias, a time trend extrapolation test (Armstrong and Overton, 1977) was also performed as a check on non-response bias. The assumption behind this test was that late respondents (those whose responses came from the second postcard, and telephone follow-up) would be very similar to non-respondents, given that they would have fallen into that category had a second set of postcards (or telephone follow-up) not been mailed. We used the multivariate general linear model (GLM) procedure to
DTD 5
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
test the null hypothesis of no difference; the procedure simultaneously compared the three survey groups (respondents after first and second postcards and for the mailed questionnaires) with respect to our study variables. This analysis indicated no difference (insignificant Wilke Lambda) (Guthrie, 2001). Although the threat of non-response bias cannot be ruled out, the results of the comparison testing of the second mailed postcard and mailed questionnaire increased our confidence in the representativeness of the sample (Guthrie, 2001). A similar comparison test for the respondents’ subgroup, for which we had publicly available data in the Compustat database, also
9
showed similar proportions to the ISO 9000 certified population. To ensure a high degree of validity, we carefully designed the questionnaire in an iterative process with 19 leading practitioners in the field. Worldwide leading experts on the ISO 9000 standard such as members of the TC 176 group (the international committee that is in charge of the standard), U.S. registrars, consultants and professional quality managers participated. We also conducted a pre-study pilot study to which several dozen respondents gave their comments. They were interviewed, and as a consequence, the questions used in the survey were
Table 2 Result of confirmatory factor analysis—item loadings: independent variables S. no. 1.
2.
3. 4. 5.
6. 7. 8. 9. 10.
11. 12. 13. 14.
15. 16. 17.
Independent variables Catalyst for change To what extent was design and development of your ISO 9000 system a springboard to introduce new practices? To what extent has ISO 9000 led to the discovery of improvement opportunities? To what extent was your investment of time and resources in ISO 9000 A starting point for other more advances practices? A catalyst for rethinking the way you do business? Understood as an opportunity to innovate? Daily practice To what extent Are the documents created for the purpose of ISO 9000 registration used in daily practice? Are preparations for external audits made at the last minute? Is the system regularly ignored? Is the system an unnecessary burden? Has it become part of your regular routine? Internally integrated To what extent was design and development of your ISO 9000 system Coordinated and led by employees who were trained and developed inside the organization? Integrated with practices already in place? Based on an analysis of internal processes and performance? Customized to the needs of your company? Externally coordinated To what extent was design and development of your ISO 9000 system Coordinated with suppliers? Coordinated with customers? Based on learning from other companies that already were registered?
Factor 1
Factor 2
Factor 3
Factor 4
.74
.78
.78 .80 .87
.72 .69 .71 .75 .81
.81 .87 .89 .88
.82 .85 .88
DTD 5
10
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
improved to the point where we had great confidence in them. Three iterations were carried out to refine the questionnaire. 4.2. Independent variables The items we used for assessing internally the Integrated variable drew on Dierickx and Cool (1989), the Externally Coordinated variable was based on Eisenhardt and Tabrizi (1995), the Daily practice variable drew on Cooper and Zmud, (1990) and the Catalyst for Change variable was derived from Argote
(1999). The results of a confirmatory factor analysis (CFA) using EQS (Bentler, 1995), which we carried out to identify the independent variables, done on all 16 items representing the four dimensions, indicate that the four dimensions can be constructed, two related to installation and two relating to usage; they had a good fit to the data (x2 = 293.45, d.f. = 1112, goodness-of-fit index [GFI] = .97, comparative fit index [CFI] = .93, root mean square error of approximation [RMSEA] = .059). Item loadings (see Table 2) were significant ( p < .01; Hurley et al., 1997).
Table 3 Comparison of the characteristics of the ISO 9000 firms and the industry-matched, industry-size-matched, industry-size-ROA-matched, and Industry-size-ROA-BM-matched controls Industry-matched
Industry-sizematched
Usage
Number of ISO 9000 certified firms to be matched Number of ISO 9000 certified firms matched Total ISO 9000 certified firms matched (%) Average book-to-market ratio ISO 9000 certified firms Average book-to-market Non-ISO 9000 certified firms t-Statistic for the paired differences in book-tomarket ratio Average market value of equity of ISO 9000 certified firms ($ millions) Average market value of equity of Non-ISO 9000 certified firms ($ millions) t-Statistic for the paired differences in market value of equity Industry matching statistics Firms matched at four-digit SIC code (%) Firms matched at three-digit SIC codes (%) Firms matched at two-digit SIC code (%)
Industry-size-ROAmatched
Usage
Industry-size-ROABM-matched
Usage
Usage
All
Low
High
All
Low
High
All
Low
High
All
Low
High
313
156
156
313
156
156
313
156
156
313
156
156
309
154
155
287
145
142
283
144
139
281
143
138
99
99
99
91
93
91
90
92
89
90
92
88
.70
.70
.71
.69
.69
.69
.70
.72
.68
.74
.74
.74
.58
.56
.61
.57
.56
.58
.69
.69
.69
.72
.72
.72
.43
.44
.43
.52
.56
.54
.89
.92
.87
1.03
1.05
1.10
1959
1978
1952
1772
1999
1556
1530
1425
1645
1769
1897
1652
1292
1185
1399
1754
1850
1655
1139
1125
1154
2016
2213
1811
1.49
1.63
1.85
.89
.95
1.12
1.15
1.23
.77
.68
.95
.58
38.6
38
39
33
34
32
19
20
18
21.5
22
21
47.5
48
47
30.5
33
28
26.5
28
25
18
19
17
13.9
14
14
36.5
33
40
54.5
52
57
60.5
59
62
DTD 5
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
The Cronbach alphas for each of the four dimensions were high enough to justify combining these items. For Integrated the Cronbach’s alpha was .92; for Externally Coordinated, it was .91; for Daily practice, it was .92 and for Catalyst for Change .92. Each of the dimensions was a composite measure of these items (see Table 2), taking the mean of the answers to the items.
11
4.3. Dependent variables For our dependent variable, the aspects of operating performance about which we asked were: defect rate as a percentage of production rates, cost of quality (scrap, rework and inspection as percentage of total sales), productivity, on-time delivery and customer satisfaction with the product and/or service. For each aspect of
Table 4 Means, standard deviations and correlations S. no.
Variable
N
Mean
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1140 1120 1123 1099 1070 1091 1078 1088 401 1076 398 402 405 401
3.5 2.1 3.8 3.1 3.7 33.4 1.2 .25 .23 726 .06 .2 1.15 85
.95 4209 .1 .22 1.4 7.5
15
Integrated Externally coordinated Used in daily practice Catalyst for change Operating performance Time since registration (months) External audits to registration (#) Manufacturing/services Gross profit margin A,B Annual sales volume A R&D expensesA,B LeverageA,B Capital intensityA,B Percent of your companys sales volume that is generated by registered facilities Number of employees
1120
184
7258
S. no.
Variable
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Integrated Externally coordinated Used in daily practice Catalyst for change Operating performance Time since registration (months) External audits to registration (#) Manufacturing/services Gross profit marginA,B Annual sales volumeA R&D expensesA,B LeverageA,B Capital intensityA,B Percent of your company’s sales volume that is generated by registered facilities Number of employees
15 A B * ** ***
On year of survey. Companies with public available data. p < .05. p < .01. p < .001.
S.D.
Minimum
0.8 0.7 0.7 0.8 0.5 22.9 .6
Maximum
1.5 1 1.4 1.2 2 4 0
5 5 5 5 5 108 5
1
2
3
4
.04 .26*** .15*** .04 .02 .1 .1 .02 .12 .08 .03 .03
.47*** .34*** .01 .04 .18* .25* .02 .34** .33** .1 .05
.46*** .07 .01 .22* .04 .03 .19 .19 .008 .05
.04
.05
.06
10.3 10 0 0 .22 71
.97 10000000 .27 .78 4.8 100
.12*** .2*** .3*** .15*** .01 .07* .17* .06 .03 .02 .1 .06 .05
25
19562
.07
5
6
7
8
9
10
11
12
13
.12*** .002 .01 .16 0.03 .12 .01 .18 .08
.03 .07 .23* .1*** .0005 .07 .005 .05
.04 .06 .02 .12 .07 .001 .06
.08 .06 .003 .1 .08 .04
.06 .3*** .003 .24** .05
.1 .14 .06 .06
.22* .5*** .05
.05 .04
.05
.06
.2*
.05
.04
.14*
.24***
.12*
.11*
.24***
14
.04
Variable
Operating performance
Model 1.a
DCost of goods sold normalized by salesa Model 1.b
Daily practice
Catalyst for change
Model 2.a
Model 2.b
DAnnual sales normalized by salesa Model 3.a
2.1 (.2)***
6988 (6479)
3.2 (.1)***
.1.7 (.1)***
227 (590)
Installed Integrated Externally coordinated
.005 (.03) .06 (.03)*
186 (1208) 2147 (1325)
.16 (.03)*** .04 (.03)
.25 (.03)*** .25 (.03)***
36 (104) 132 (128)
Usage Daily practice Catalyst for change
.1 (.03)*** .2 (.03)***
2520 (1190)* 2273(1469)y
DCost of goods sold (normalized by salesa)
220 (107)* 291 (119)**
Mode 3.b
DAnnual sales normalized by salesa Model 4.a
DGross profit margin Model 4.b
.02 (.2)
271 (686)
.04 (.2)
.025 (.03) .015 (.04)
33 (104) 122 (131)
.025 (.03) .02 (.04)
.055 (.03)y .07 (.03)*
219 (110)* 267 (120)* 227 (103)*
.05 (.03) .07 (.03)* .08 (.08)
Control variables .2 (4.4) .0005 (.001) .0007 (.0009) .15 (.35) .00002 (.0001) 1.6 (3.1) .0002 (.0001) Time since registration .003 (.0008)*** (months) Number of external audits .009 (.05) 2146 (2660) .06 (.05) .01 (.06) 230 (226) .08 (.07) 191 (229) .09 (.07) to attain registration Manufacturing/services .03 (.05) 3299 (2513) .1 (.05)y .04 (.06) 324 (242) .02 (.07) 330 (243) .007 (.07) Percent of your company’s .01 (.04) .05 (.03) .01 (.04) .05 (.02) .01 (.03) .04 (.04) .04 (.01) .04 (.03) sales volume that is generated by registered facilities R&D expenses 248 (595) 253 (669) .02 (.07) 325 (674) .02 (.07) Leverage 401 (516) 234 (420) .09 (.07) 274 (423) .09 (.07) Capital intensity 1.5 (10) 2.1 (10) .015 (.001)** 3.5 (10) .015 (.001)** Observations 1014 387 1030 1032 370 387 364 387 2 Res log likelihood 1744 6594 2018 2059 6147 252 6032 230 Adjusted R-square .25 .19 .33 .35 .18 .21 .24 .25 Standard errors in parentheses. a Normalized by sales (changes in annual sales divided by annual sales). y p < .1. * p < .05. ** p < .01. *** p < .001.
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
Intercept
DGross profit margin
DTD 5
12
Table 5 Results of hierarchical linear model
DTD 5
E. Naveh, A. Marcus / Journal of Operations Management xxx (2005) xxx–xxx
13
operating performance, the person who answered the survey was asked to indicate on a Likert Scale (1–5) a company’s level of performance improvement since gaining ISO 9000 registration and until the time of the survey (or up to 5 years after certification, if more than 5 years had passed since being certified, at the time the survey). We added a comment to the questionnaire to indicate that by performance improvement since being certified, we meant performance after certification in comparison to the average performance in the 5 years before certification. The operating performance measure was a composite measure of these aspects, taking the mean of the answers to the items, since the Cronbach alpha (.89) was high enough to justify combining them. For companies with Compustat data, following the classical work of Barber and Lyon (1996), we used the measure change in annual cost of goods sold as the measure that operationalized operating performance (see, for their discussion on measuring operating performance, Barber and Lyon, 1996; 361 and 364. On ISO 9000 see, as well Wayhan et al., 2002). Cost of goods sold reflects the ability of the organization to reduce the direct costs associated with producing their products. In studies of quality, performance typically has two dimensions—operating and business performance, which are about long-term returns (Benson et al., 1991; Everett and Adam, 1994; Flynn et al., 1994). The aspects of business performance that we used were changes in annual sales (changes in this variable may also be affected by organizational size so we ran the analysis normalized by sales—that is, changes in annual sales divided by annual sales (see Section 4.5 below and Table 5) and gross profit margin (defined as sales minus the cost of goods sold divided by sales) (see Heskett et al., 1997). These measures came from the Compustat dataset. We took the longitudinal measures for cost of goods sold, sales and gross profit margin for 5 years prior to gaining ISO 9000 certification, the year of certification, and 5 years after certification (or from certification until the survey was administered, if it was less than 5 years).2 We followed Hendricks and Singhal (2001) in
choosing this time frame. They argue that the quality management literature does not provide much theoretical guidance regarding the appropriate time period for analysis of implementation of quality improvement programs. However, they suggest that 4-to-5 years before and after implementation capture the necessary preparation time before an effective quality program can be put into place and make its impact felt.3 Accordingly, the measures we used were annual cost of goods sold, sales and gross profit margin in year t + 1 minus annual cost of goods sold, sales and gross profit margin in year t (normalized by sales; see Table 5) (Makridakis et al., 1998). As other researchers who have used this measure, we corrected for inflation. Many factors are related to changes in cost of goods sold, sales and gross profit margin. The 11-year longitudinal movement in these variables partially takes into account the fact that companies gained certification in different years (some of the companies obtained certification in the early 1990s and some in the late 1990s) and under different economic conditions. They were also from different sectors. However, these differences do not diminish the effects of changes in economic conditions and market trends, which might contribute to variations. Thus, an analysis that takes into consideration these aspects is also required. Further validation of the analysis (see Section 4.5 below and Table 5) with samples of four matching non-ISO 9000 organizations is described below.
2 Mean time since certification for respondents with public data was about 4 years (for all survey respondents it was about 3 years– see Table 3). About a quarter of the respondents with public data had been certified for 5 years or more, a quarter had been certified between 2.5 and 5 years, a quarter between 1 and 2.5 years and a quarter 10 and the variance of the other variables explained by the variable with high condition index is >.5. Mild collinearity exists when the condition index is 10–30; it is moderate when the condition index is 30–100, and severe when the condition index is >100. Except in Model 3, the condition indices were 10 but
