The Impact of Size on Enterprise Resource Planning (ERP) Implementation in the US Manufacturing Sector

The Impact of Size on Enterprise Resource

Planning (ERP) Implementation in the US

Manufacturing Sector

ARTICLE

in

OMEGA · JUNE 2003

Impact Factor: 4.38 · DOI: 10.1016/S0305-0483(03)00022-7

CITATIONS

131

557

3 AUTHORS:

Vincent A. Mabert

Indiana University Bloomington

SEE PROFILE

Ashok Soni

Indiana University Bloomington

SEE PROFILE

Munirpallam Venkataramanan

Indiana University Bloomington

SEE PROFILE

Available from: Ashok Soni Retrieved on: 20 October 2015

Omega 31 (2003) 235–246

www.elsevier.com/locate/dsw

The impact of organizationsize onenterprise resource

planning (ERP) implementations in the US manufacturing

sector

Vincent A. Mabert, Ashok Soni

_{, M.A. Venkataramanan}

Operations and Decision Technologies Department, Kelley School of Business, Indiana University, Bloomington, IN 47405, USA Received 19 February 2001; accepted 29 January 2003

Abstract

Enterprise Resource Planning (ERP) systems have experienced a phenomenal growth in the last 5 years and at present they are pervasive in the US manufacturing sector. This paper describes an attempt to chronicle this phenomenon through a series of case studies and an extensive survey. Manufacturing companies ranging in size from a few million dollars in annual revenues to over a hundred billion dollars are included in this study. The key 4nding from this study is that companies of di5erent sizes approach ERP implementations di5erently across a range of issues. Also, the bene4ts di5er by company size. Larger companies report improvements in 4nancial measures whereas smaller companies report better performance in manufacturing and logistics.

?2003 Elsevier Science Ltd. All rights reserved.

Keywords:Enterprise resource planning; Organizational size; Survey methodology; Operations systems

1. Introduction

Today’s global business environments are characterized by unprecedented competitive pressures and sophisticated customers who demand innovative and speedy solutions. Understanding and optimizing business processes is a cor-nerstone of success in these fast-changing environments. Global distribution channels, numerous international plant sites, and closely integrated sourcing arrangements have changed the way hundreds of companies do business. A key component of managing these organizations is Information Technology (IT). Over the past few years, many companies have embraced a new class of planning and resource man-agement software systems to integrate processes, enforce data integrity, and better manage resources. These package systems are broadly classi4ed as Enterprise Resource Plan-ning (ERP) systems.

∗_{Corresponding author. Tel.: 855-3423; fax:} +1-812-856-5222.

E-mail address:[email protected](A. Soni).

ERP systems, which evolved from Materials Require-ments Planning (MRP) and Manufacturing Resource Plan-ning (MRP II) systems, are expected to provide, at least in theory, seamless integration of processes across functional areas with improved workAow, standardization of various business practices, improved order management, accurate accounting of inventory, and better supply chain manage-ment. The Gartner Group coined the term ERP in the early 1990s to describe these systems and stipulated that such software should include integrated modules for accounting, 4nance, sales and distribution, human resources, materials management and other business functions based on a com-mon architecture that links the enterprise to both customers and suppliers.

According to industry reports at least 30,000 com-panies worldwide have implemented ERP systems. The vast majority of these implementations have taken place inthe mid-1990s to 2000. Thus, these systems are rela-tively new with very little research available concerning their implementation, their operations or their impact. Like many emerging areas, the initial research consisted 0305-0483/03/$ - see front matter?2003 Elsevier Science Ltd. All rights reserved.

primarily of case studies and articles in the business press or practitioner-oriented journals. Many of these articles tended to be in terms of anecdotal information based on a few successes or failures. As a result, initial information on ERP often tended to be contradictory and skewed to 4t certain points of view. Also, these articles tended to be general whereas our interest was to gather data in a more systematic manner concerning the implementation of such systems in manufacturing companies.

To better understand and guide this process, the authors adopted a two-phased approach to analyze ERP adoption and implementation experiences in the manufacturing sec-tor. Inthe 4rst phase, a case study approach was used to study 12 di5erent manufacturing ERP implementations. These implementations were studied using structured in-terviews of key managers, IT professionals, and users associated with each company’s implementation. In addi-tion, senior consultants at six consulting 4rms specializing in ERP implementations were interviewed at length to get their perspectives onERP [1,2]. The primary objective of the case studies and the consultant interviews was to obtain reliable and detailed information on the current status of ERP practice and implementations in the manufacturing sector.

Two key issues emerged from this phase of the project. First, companies of di5erent sizes tended to do di5erent things in their implementations. In particular, there were dis-tinct di5erences between small and large companies over a range of issues. These di5erences included: (1) the moti-vationto go with anERP system; (2) the di5erent systems adopted; (3) the implementation strategies; and (4) the de-gree of reengineering and customization of the base system. And second, there were di5erences in the outcomes and ben-e4ts attained. While the case studies proved useful in under-standing the general nature of these di5erences, this part of the study was based on a small sample. To con4rm our initial 4ndings, a survey of a larger sample of companies was un-dertakeninthe second phase of the project inorder to obtain a broader perspective of ERP practice and experiences relat-ing to adoption, selection of systems, customization, costs, and performance, and success factors across di5erent sized companies. More speci4cally, the primary objective of this project is to study the impact of the organization size on ERP adoption and implementation.

The relationship between structural variables of an organization such as size, industry type and organizational structure, and their impact on various operations has been studied for a long time [3–5]. Organizational size is the most frequently examined structural variable and has been used to study issues relating to innovation, R & D expen-ditures and market power [6–8]. The impact of company size on adoption, implementation and use of information technologies has received increasing attention in the recent academic literature as well [9–12]. Gremillionconjectures a lack of relationship between the size and usage. Addition-ally, several papers [8,11,13] suggest that larger companies

are more likely to be early adopters of information technol-ogy innovations.

In the Operations area, several studies of manufacturing 4rms indicate that organization size plays a critical role in terms of the level of adoption and use of technologies [14–

18]. These 4ndings show that in general small manufactur-ers tend to lag behind large manufacturmanufactur-ers in implement-ing new technologies, plus employ di5erent practices. ERP implementations have followed similar trends. While larger companies were the 4rst movers to ERP systems in the mid-1990s, today smaller companies view this approach as an important management tool. Increasingly, many mid-size and smaller companies are either implementing or planning to implement ERP systems [1].

The research reported here provides an insight into some of these critical issues outlined above. In the next section, we discuss the relevant research germane to this evolving area and the research methodology employed to conduct the investigation. Section3describes the initial data collection 4eld study, initial observations and 4ve research proposi-tions. Section4outlines in detail the mail survey steps used for a more extensive and systematic data collection e5ort, and testing of the propositions. Section5presents other ob-servations and insights gained from both the case studies and the survey, with the 4nal section highlighting our con-clusions.

2. Research issues and research framework

Despite the implementation of ERP systems since the mid-1990s, academic research inthis area is relatively new. It is only recently that researchers have dealt with various aspects of ERP in a more systematic manner. The initial thrust of many of these articles has beeninthe implemen-tationarea. Davenport [19] inanearly ERP article looked at the reasons for implementing ERP systems and the chal-lenges of the implementation project itself. Van Everdigen et al. [20] surveyed 2647 European companies across all in-dustry types to determine adoption and penetration of ERP by functionality. Mabert et al. [1] used a hybrid approach with a series of case studies followed by a survey to study penetration of ERP systems, motivation, implementation strategies, modules and functionalities implemented, and op-erational bene4ts as they apply to the US manufacturing sector. Adam and O’Doherty [21] used case studies to study ERP implementations in small and medium enterprises in Ireland.

More recently, several researchers have dealt with orga-nizational issues either regarding implementation of ERP systems or performance and bene4ts of ERP systems. Abdinnour-Helm et al. [22] look at pre-implementation attitudes and organizational readiness for implementing an ERP system. They conclude that extensive organiza-tional investments in shaping pre-implementation attitudes do not always achieve the desired e5ects. Stratman and Roth [23] develop and operationalize eight organizational

ERP competence constructs. They de4ne ERP competence as a portfolio of managerial, technical and organizational skills and expertise hypothesized as antecedents to im-proved business performance after an ERP system is op-erational and functionally stable. Sarkis and Sundrraj [24] present a bene4ts-evaluation framework for assessing of-tenunder-emphasized resources created as a result of ERP implementations. Bendoly and Jacobs [25] discuss the im-portance of aligning ERP system implementation strategies with operational requirements, and how operational require-ments drive the appropriateness and bene4ts accrued from these strategies. Another issue that has started to emerge is the operational impact of ERP systems on various en-terprise applications such as e-business and supply chain management systems. One such paper is by Bendoly and Kaefer [26] where they study the impact of ERP systems on transactional eMciencies of e-commerce.

While these studies provide insights into some ERP im-plementation issues and into bene4ts, the published research to date does not provide a systematic approach to analyze and evaluate ERP implementations for manufacturing com-panies. In recent years, a growing number of researchers have proposed exploratory research designs using method-ologies such as case studies, the Delphi technique and sur-veys for either emerging areas or where the knowledge base is small [27,28]. The case study approach is increasingly becoming popular because of the detailed information that canbe obtained [29,30]. Several studies inthe IT area have used this approach [31]. Surveys have beenextensively used for research. In the Operations Management area, there is a strong tradition for using this methodology. For example, studies of MRP and MRP II, two areas closely related to ERP, have used this approach [32–34]. For this project, we chose to use the case study and survey methodologies in a phased approach. The case study approach was selected to collect detailed information while the survey approach was selected to provide a broader perspective.

In deploying ERP systems, companies have found that the software installation complexity is just the tip of the iceberg. A successful ERP implementation involves more than hav-ing sophisticated software and advanced computhav-ing tech-nologies. For example, each system is “generic” in that it is a standard representation of how a typical company does business. Within this standard representation, ERP systems provide di5erent options on how to set up various processes. These options, usually referred to as “Best Practices”, pro-vide Aexibility insetting up the ERP system. Evenwith this Aexibility, it may not be possible for a company to 4t an ERP system exactly to 4t their organization. Thus, to implement an ERP system, a company can either change their processes to 4t the package or customize the system to 4t their processes. Just how a company approaches this decisioncanhave major implications onhow the imple-mentation will be carried out and what the eventual out-comes are. Thus, it is critical to study a range of di5erent implementations.

3. The case studies

For the case study phase, 18 manufacturing companies and six consulting 4rms were contacted to be part of the project. Of these, twelve of the manufacturing companies and all six of the consulting 4rms agreed to be in our study. The manufacturing companies are located in Indiana, Illi-nois, Minnesota and Ohio. The size of these companies ranged from $30 million in annual revenues to over $35 bil-lion. A detailed questionnaire was sent to the companies in advance of the interviews. All interviews were conducted at a company site by at least two interviewers. These inter-views were all done during the summer of 1999. At least one key executive, one member of the implementation team and one key user were interviewed. All sessions were taped and the transcriptions of these tapes were reviewed and authen-ticated by all interviewers before any information or data was used. In the case of two of the manufacturing compa-nies, a follow-up visit was conducted to clarify some of the initial 4ndings. The primary objective of the case studies was to compare ERP initiatives and experiences as imple-mented by manufacturing. The interviews were exploratory in nature and designed to provide insight into the following set of research questions:

1. What motivates a manufacturing company to imple-ment an ERP package?

2. Which ERP packages do manufacturing companies im-plement and how are they selected?

3. What is the con4guration of the systems implemented? 4. Which implementation strategies do manufacturing companies utilize?

5. What degree of customizationoccurs inmanufacturing ERP implementations and are there some modules/processes that are customized more thanothers?

6. What does it cost to implement an ERP system and are the major cost categories inAuenced by type of implemen-tationor implementationstrategy?

7. Does returnoninvestment play a major role inthe decisionto implement?

8. What are the bene4ts the companies expect as a result of implementing an ERP system?

The most striking part of the case studies was the amount and detail of information companies were willing to share about their implementations. These interviews provided the following insights into these companies’ approach to their ERP implementations:

• Companies adopted ERP systems for a variety of rea-sons. These included theY2K problem, replacing legacy systems, system simpli4cation and improvement, process and operations improvement, reducing costs of informa-tionsystems, and competitive pressures.

• Most companies performed some type of ROI analysis to justify adopting ERP systems. A few even approached

the decisionsimply as a strategic initiative or as a cost of doing business.

• The con4gurations of systems implemented varied. Some companies implemented a single ERP package while others selected modules from di5erent ERP packages (Best-of-Breed approach). One even developed a totally homegrownsystem.

• Several implementation strategies were used. These included implementing all key modules at once (The Big-Bang approach), phasing in modules one or a few at a time (The Phased-In-By-Module approach), imple-mented a sub-set of modules all at one time (The Mini Big-Bang), and modules phased in by divisions, plants, business units or geographies (The Phased-In-By-Site approach).

• All companies stressed the importance of planning an implementation. However, the degree of planning varied across companies.

• Most companies customized the base system but the de-gree of customizationvaried from very minor modi4ca-tions to major rewrites of code for certain functionalities. Any major modi4cations added signi4cantly to both costs and implementation time.

• Most companies undertook some reengineering of pro-cesses. A few did major reengineering upfront while most deferred it to after the system had beenimplemented. • Most companies expected positive returns from their ERP

implementation. The bene4ts and returns expected varied signi4cantly across companies.

• The cost estimates and percentage breakdowns by various categories such as the costs for the software, hardware, consulting and training also showed interesting variations across the companies.

While many of the above issues, activities, approaches and strategies were commonto all our case study companies, there are many di5erences across these implementations. One key di5erence is that companies of di5erent sizes tend to do di5erent things in their implementations across a range of issues. For example, smaller companies are more likely to change their processes to 4t the system whereas larger companies are more likely to customize the system. Any changes to the system can have major implications. Generally, modi4cations lead to higher costs, longer im-plementation time and more complicated imim-plementations. Other di5erences across smaller and larger include the motivationto go with anERP system, the implementation strategies, type of systems adopted, the extent of modi4ca-tions to the base system, and the bene4ts the companies get from ERP.

Organizationsize canbe de4ned intwo ways—by number of employees or by revenues. Several studies have used num-ber of employees as a measure of company size [14,15,17] while others have used annual revenues [9,11,35]. We chose to use annual revenues as a convenient measure of organiza-tion size classifying the companies in both the case studies

and the survey as follows: companies with annual revenues of less than$200 millionare classi4ed as small, those be-tween$200 millionand $1 billionare classi4ed as medium, and those with more than$1 billionare classi4ed as large. These breakdowns were developed in consultation with the case study companies and the consultants. With this clas-si4cation scheme, the key di5erences across companies of di5erent sizes inthe case studies are stated inthe form of the following propositions:

Proposition 1. Adoption of ERP systems by large compa-nies is motivated more by strategic needs whereas tactical considerations are more important for smaller companies.

Proposition 2. Larger companies employ more ERP func-tionality than small companies.

Proposition 3. Large companies customize ERP software more while small companies adopt business processes within ERP systems more.

Proposition 4. Large companies use an incremental im-plementation approach by phasing in the systems while smaller companies adopt more radical implementation ap-proaches such as implementing the entire system or sev-eral major modules at the same time(The Big-Bang or the Mini Big-Bang approach).

Proposition 5. Large companies report greater bene8ts in the 8nancial areas, while small companies report more ben-e8ts from their ERP implementations in manufacturing and logistics.

These di5erences were also reAected in such measures as the cost to implement, the cost breakdowns by category, and the implementation time. The cost of implementation among the case study companies ranged between 1.5% and 6% of annual revenues. Measuring the implementation cost as a percent of revenue provides a convenient way to de-termine if size of the organization has any impact on ERP costs. Our data show that small companies spent between 3.5% and 6% of revenues to implement ERP, with an aver-age of 5.53%. Medium-sized companies reported an averaver-age of 3.08% while large companies had an average cost to rev-enue ratio of 2.23%. The breakdowns of costs into various categories also showed key di5erences. Smaller companies spent a higher proportionof their budgets onthe cost of soft-ware. Larger companies, on the other hand, spent a higher percentage on their ERP implementation teams. The imple-mentation times also varied by the size of the companies. Some of the key di5erences are showninTable1.

The information from the case studies was used to de-velop the survey questionnaire for the broader study of ERP practice and experiences. The methodology used and the 4ndings from the survey are reported in the next section.

Table 1

Case study 4ndings of cost, time, and implementation strategies

DescriptionSmall Medium Large

4rms 4rms 4rms

Number of 4rms in

case study sample 4 3 5

Average cost to

an-nual revenues ratio 5.53% 3.08% 2.23% Average

implementa-tiontime (inyears) 2.125 3 4.17

Numbers using the

big-bang approach 3 1 0

Numbers using the

Phased-inapproach 1/4 2/3 5/5

4. Surveyobjectives, methodologyand results

The survey questionnaire was four pages long and had a total 24 questions. It included questions on company and respondent demographics, adoption and selection of a sys-tem, implementation, customization, costs and bene4ts, and post-implementation plans. This questionnaire was designed as an exploratory instrument to collect information about these phases of an ERP project. This instrument was not de-signed to probe into the rationale for why companies chose to do things incertainways.

The responses were encoded using a mix of check boxes, open-ended answers, and a Likert scale with measures from 1 to 5. The case studies provided the guidance for the en-coding scheme in terms of what type of questions required what responses. For example, the total cost was segmented into buckets because the interviews showed that respon-dents were more comfortable with providing approximate 4gures instead of exact values. The motivational and bene4ts responses were encoded using the Likert scale because re-spondents were generally good at determining relative mea-surements for these kinds of questions. After the initial development of the survey questionnaire, it was thoroughly tested by two ERP project leaders from our case study com-panies and two consultants, and 4ne-tuned.

The survey and cover letter were mailed to a randomly selected sample of 5000 APICS members employed in manufacturing companies in the US in August 1999. No follow-ups were done. By September 1999, 482 usable re-sponses had beenreceived for anoverall response rate of 9.6%. Given the length and comprehensive nature of the survey, this response rate was concluded to be reasonable. Respondents were not asked to provide company-identifying information, and postage-paid return envelopes were pro-vided to maintain con4dentiality.

Of the 482 responses, 44.6% of the companies had im-plemented an ERP system, another 18.5% were in the pro-cess of implementing, 10.4% were planning to implement one within the next 18 months and 26.5% had no plans for

anERP system for the foreseeable future. The companies spanned a wide range as measured by revenue and employ-ment. Approximately 45% have annual revenues of less than $200 million. About the same percentage of companies em-ploy less than1000 people. The largest company has over $100 billion in annual revenues and 240,000 employees. The smallest has $2 millioninrevenues and 10 employees. The distribution of make-to-stock and make-to-order was evenly balanced across this set of companies. The respondents were a mix of managerial and sta5 personnel. Since the question-naire was sent to only APICS members working for man-ufacturing companies, a large number of respondents were employed in the materials and production planning areas. Just over 22% were executives, 17.0% were materials or supply chain managers, 9.0% were plant managers, 11.6% were purchasing managers or buyers, 19.8% worked in the production or inventory control areas, and 8.4% were in the systems area. Another 12.1% either did not respond to this question or indicated some other position.

For the analysis reported here, only responses for entire 4rms or divisions who have already implemented or are in the process of implementing an ERP system were selected. The companies that met these criteria were separated into large, medium, and small, based onthe classi4cationscheme discussed above. This resulted in the following breakdowns: large companies—65, medium companies—52 and small companies—76. From here on, we refer to these 193 com-panies as the ERP sample. This sample of ERP comcom-panies represents a total of 193 out of the 482 4rms who responded to the original survey. Within the ERP sample, the num-ber of small 4rms is 39.4% compared with 41.3% of small companies in the overall sample, the number of medium companies accounts for 26.9% compared with 25.9% in the overall sample while the number of large companies in the ERP sample is 33.7% compared with 32.8% inthe overall sample of 482. Thus, the breakdownof small, medium and large companies implementing ERP systems is very similar to the companies who responded to the survey.

Establishing whether the companies in the ERP sample are representative of the population of companies adopting ERP systems is harder to do since there was no demographic information available on ERP implementations in the man-ufacturing sector. One methodology we used was to com-pare our sample with the information provided to us by two of the consulting 4rms in our case studies. There are sev-eral statistics that indicate that this sample comes close to the populationof companies who had adopted ERP systems at the time of the survey. First, the analysis on package adoptionshows that package adoptionrates of the sample companies are very similar to the market shares of those packages (Table6). Second, the consulting 4rms had esti-mated that about 50% of the manufacturing 4rms had imple-menting an ERP system while our sample put this number at 44.6%. Third, the consulting data showed that about a third of the implementations used the “Big-Bang” approach. Our sample puts this number at about 36% (Table3).

A variety of methods have beenused to evaluate the data. The rest of the section summarizes our 4ndings.

4.1. Proposition 1: motivational factors

The questions relating to motivational factors employed a 4ve point Likert scale (4ve being very important and one being unimportant) to measure their importance. These mo-tivational factors are a mix of IT factors and business fac-tors. The responses were analyzed as follows: a response of 4 or 5 was considered positive, a 1 or 2 as negative and a 3 as neutral. The neutral responses were not included in the analysis. Table2presents the summary of the respondents’ answers to these questions. The percentages reAect only the positive responses. That is, the percentages indicate the per-centage of 4rms who listed that motivational factor as either important or very important. For example, 85.9% of all 4rms listed “Replace Legacy Systems” as either important or very important. Similarly, 86.8% of all the small 4rms listed this motivational factor to be either important or very important. When analyzed across di5erent sizes of companies, the results show several similarities and di5erences. The impor-tance of these factors by company size was tested using a One-Factor ANOVA on the data from the survey questions. Thep-values inTable 2are from this ANOVA test. The factors “Replace Legacy Systems” and “Simplify and Stan-dardize Systems” are important to all companies. These two reasons were important to all of the case study companies as well. All of them had beenoperating with a patchwork of legacy systems that were becoming harder to maintain and upgrade. Additionally, the competitive pressures on them in-creasingly required more responsive systems with real-time integrated information that the legacy systems could not provide easily. Thus, it is not surprising that these two fac-tors are important to all companies. However, there is a signi4cant di5erence (p-value = 0:001) onthe importance of the factor “Simplify and Standardize Systems” by size of company. Almost all large companies considered this important. Inadditionto the ANOVA test, pair-wise com-parisons were analyzed using categorical analysis. The sig-ni4cantp-values reported are from the resulting Chi-square test. These pair-wise comparisons show signi4cant di5er-ences between large and small companies (p-value=0:001), and large and medium companies (p-value = 0:04) as well. One possible explanation is that large companies are likely to have a number of legacy systems in place so simpli4ca-tion and standardizasimpli4ca-tion becomes a more important issue. For example, Owens Corning replaced 211 legacy systems with their ERP system [19].

The factor, “GainStrategic Advantage”, also shows a signi4cant di5erence by company size. The pair-wise com-parisons show a signi4cant di5erence (p-value = 0:008) be-tween large and small companies. This is surprising because having an ERP system by itself is not likely to give a com-pany any signi4cant strategic advantage since their compe-titionis likely to have implemented similar systems. For

example, over 300 pharmaceutical companies around the world have implemented ERP systems (www.sap.com Au-gust 2000). Thus, the ERP system by itself is not likely to provide any signi4cant competitive or strategic advantages. Our case study companies expected their strategic advantage to come from how they leveraged the vast amounts of oper-ational data generated by their ERP systems. One key use of this operational data is in 4nancial analysis. All 4ve of the large case study companies reported increased eMciencies in budgeting, 4nancial controls and 4nancial close cycles as a result of the information availability from their ERP sys-tem. Another area of strategic importance is the use decision support models and data mining tools. These systems are generally not part of an ERP system but separate systems known as “Bolt-Ons”. Finally, strategic and competitive ad-vantages can also come from how companies integrate and manage their specialized strategic systems, such as supply chain management, customer management and e-business systems, with their ERP system.

ERP systems are also important to large companies for linking their global activities. There is a signi4cant di5er-ence (p-value¡0:0001) betweenlarge and small compa-nies. That is to be expected since large companies are more likely to have global activities. Linking to suppliers and cus-tomers was a key objective of the both medium and large case study companies. It was not emphasized as much by the small companies in the case study sample. However, the survey results show that this is of equal importance to all companies. A surprising result from the survey is that the Y2K issue was ranked very low among motivational fac-tors. Much of the literature cites Y2K as the reasoninthe upsurge of ERP inthe mid-1990s. Solving the Y2K prob-lem was more important to small- and medium-sized com-panies. One key di5erence between the larger and smaller case study companies was that managers in the larger com-panies expressed more con4dence in solving the Y2K prob-lem since they had large IT sta5s whereas the small- and medium-sized companies did not have such dedicated re-sources and looked more at the ERP system to solve that problem.

4.2. Proposition 2: implementation strategies

The strategy used for the implementation is one of the most important factors in assessing the impact of an ERP system on an organization. Strategies can range from a sin-gle go-live date for all modules (Big-Bang) to sinsin-gle go-live date for a subset of modules (Mini Big-Bang) to phasing in by module and/or site. While the Big-Bang approach usu-ally results inthe shortest implementationtime, it is also the riskiest approach because it canexpose the entire stability of a company incase of any problems. Clearly, inaninte-grated environment, problems in one part of the system can seriously impact the entire enterprise.

The decisiononwhich strategy to deploy depends on a range of issues such as complexities of size, processes

Table 2

Summary responses for motivational factors

Motivationfactors All 4rms Small 4rms Medium 4rms Large 4rms Statistical

(%) (%) (%) (%) signi4cance

Replace legacy systems 85.9 86.8 78.9 89.5 No di5erence

Solve the Y2K Problem 56.5 63.1 63.6 42.3 0.04

Ease of upgrading systems 44.5 35.3 45.2 54.3 0.09

Simplify and standardize systems 83.3 72.4 82.9 94.7 0.001

Pressure to keep up with competitors 49.2 41.7 45.2 59.6 No di5erence

Improve interactions and communications

with suppliers and customers 75.2 70.6 81.3 76.1 No di5erence

Restructure company organization 32.0 32.8 27.0 34.6 No di5erence

Gainstrategic advantage 79.6 70.0 75.8 91.8 0.03

Link to global activities 55.5 35.6 61.8 73.6 0.0001

Table 3

Implementation strategies

Strategy All 4rms Small 4rms Medium 4rms Large 4rms

(%) (%) (%) (%) Big-Bang 36.32 47.37 48.00 14.06 Mini Big-Bang 17.37 23.68 18.00 9.38 Phased-Inby Module 17.37 19.74 10.00 20.31 Phased-Inby Site 25.79 7.89 24.00 48.44 Other 3.16 1.32 0.00 7.81

and operations. For both the case study and the survey, ap-proximately half the implementations used one of the two Big-Bang approaches and half used one of the Phased-In approaches. However, as Table 3 shows, there are very clear di5erences in the implementation strategies by size of company. Over two-thirds (69%) of implementations inlarge companies were phased ineither by module or by site whereas over 70% of small companies used one of the Big-Bang approaches. Di5erences in strategies be-tween both large and small companies, and large and medium companies were statistically signi4cant (Chi-square p-values¡0:0001). There was no statistical di5erence between small and medium companies.

Manufacturing companies seem to be selective in which modules/functionalities to implement. The survey results show that 4ve modules/functionalities (Financial Account-ing/Control, Materials Management, Order Entry, Produc-tion Planning, and Purchasing) have been installed in just over 87% of the reporting companies. Four of these modules are generally implemented during the 4rst or early phase of the project, with the Production Planning module typically delayed until later in the implementation. The Financial Ac-counting module, at 92% by reporting companies, was most frequently implemented. Of the “manufacturing/logistical” modules, the Materials Management module was the most frequently implemented at 90% of the companies. All com-panies reported having at least implemented two or more of

these 4ve modules. Incontrast, VanEverdingenet al. [20] in a survey of 2647 small and mid-sized companies (across all industry types) in Europe found that 13% of them used ERP software in just one functional area and 70% used it in more than three functional areas.

4.3. Proposition 3: customization of packages

Customizationrefers to modifying the package through code re-writes, changes or additions. Because of the inte-grative architecture of ERP systems, customizations can be prohibitively expensive. A highly customized system also becomes harder to upgrade because all the changes have to be accounted for inthe upgrade. The commonhypothesis is that companies are generally more willing to change their operating processes than customizing the package. Our sur-vey results, however, indicate that almost all companies went through some form of customization, as showninTable4.

The degree of customization varies signi4cantly across size of company. Larger companies customize more. There are signi4cant di5erences between small and large compa-nies (Chi-squarep-value¡0:001), and between medium and large companies (Chi-squarep-value = 0:06). The sur-vey results show that over 50% of the large companies did either signi4cant or major modi4cations whereas most small companies only made minor modi4cations. For the large companies, it may not be possible to avoid customization.

Table 4

Degree of customization

Overall All 4rms Small 4rms Medium 4rms Large 4rms

customization(%) (%) (%) (%) Minor 61.11 72.86 62.00 46.67 Signi4cant 29.44 22.86 24.00 41.67 Major 7.78 2.86 12.00 10.00 Other 1.67 1.43 2.00 1.67 Table 5 Module customization

CustomizationAll 4rms Small 4rms Medium 4rms Large 4rms

by module (%) (%) (%) (%)

Order entry 33.86 31.58 41.18 33.85

Production planning 21.36 11.84 19.61 23.08

Materials management 16.15 5.26 19.61 21.54

Their complex operations and organizational structure tends to increase the pressure for more custom-build processes and reports. An interesting observation from the case studies was that companies who started their implementations ear-lier tended to customize more. Some of the managers have hypothesized that the evolutionof ERP systems through the late 1990s had “improved” the systems to a point where it was no longer necessary to customize as much, while others think that the knowledge base among consultants and ven-dors had improved signi4cantly over time, minimizing the need for customization.

Order entry is the most customized module (Table5). A third of all companies made some changes to their or-der entry module. Many of the case study companies had also changed the order entry module, primarily for two rea-sons. First, they felt this was such a critical process that they just could not a5ord to make any errors. Second, most felt their company order entry processes were unique because of pricing, product lines, product bundling and customer ser-vice/interface. Production planning modules had the second highest degree of customization, followed by materials man-agement modules. One out of every 4ve implementations is customized to some degree for medium and large compa-nies in these two categories. Multiple plants, geographical dispersion (at times globally), and di5erent production pro-cesses for multiple product lines could be attributed to this customization.

4.4. Proposition 4: package adoption and con8guration

The issue of which ERP package to implement is an im-portant decision for any company not only for functionality and ease of implementation but also for future upgrades and for using other specialized packages with the ERP system. Tables6and7 present company based data for adoption

of di5erent ERP packages. Table6summarizes the adop-tion by package breakdowns across all companies and then by size of company, with Table7providing data on how the packages are implemented. Table6 also includes the global market shares of each package. Overall, the penetra-tionof ERP packages inUS manufacturing 4rms appears to be very similar to overall global market shares reported (a Chi-Square test was insigni4cant). The only exception is Peoplesoft, which has only a 2.6% share in manufactur-ing versus an overall market share of 9%. That is to be ex-pected since its traditional strengths are in human resources and not in manufacturing. There are clear di5erences across the di5erent sized companies on the packages they adopt. Chi-Square tests between all three combinations of compa-nies (small vs. large, small vs. medium, and medium vs. large) were all signi4cant (p-values¡0:001).

As expected, large companies favor SAP more than small companies (41.5% vs. 10.5%). Over 72% of the large com-panies use just 4ve di5erent packages (SAP, Oracle, Baan, JD Edwards and SSA) as compared to 47% for the small companies. Van Everdingen et al. [20] intheir survey found that “best 4t” with “current business practices” and pack-age Aexibility were the key criteria inthe packpack-age adoption decision. Since many of the “smaller” ERP systems, such as MAPICS and QAD, have evolved directly from MRP II packages, companies looking for a good 4t with their cur-rent business practices are more likely to adopt ERP systems that have evolved from their MRP II systems. For example, two of the small companies in our case study sample chose to go with ERP systems from the vendors of their original MRP II system for precisely that reason.

There are also key di5erences (Table7) among compa-nies on the con4guration of the ERP system implemented. Over 56% of small companies use only a single ERP pack-age whereas only 33% of the medium-sized and 28% of the

Table 6

Summary responses for package adoption

ERP package Overall Small 4rms Medium 4rms Large 4rms Market sharesa

(%) (%) (%) (%) (%) SAP 25.0 10.5 25.5 41.5 32 Oracle 14.6 11.8 19.6 13.8 13 Baan9.4 14.5 5.9 6.2 7 JDE 6.8 10.5 2.0 6.2 7 SSA/BPCS 3.6 0.0 7.8 4.6 3 Peoplesoft 2.6 1.3 3.9 3.1 9 QAD 2.1 2.6 2.0 1.5 2 MAPICS 1.6 3.9 0.0 0.0 1 Others/multiple 34.4 44.6 25.5 23.1 26

a_{1998 Forecasted market shares, AMR research [36].}

Table 7

Summary responses for package implementation

Approach All 4rms Small 4rms Medium 4rms Large 4rms

(%) (%) (%) (%)

Single ERP package 40.6 56.6 33.3 27.7

Best-of-Breed from di5erent packages 4.2 1.3 2.0 9.2

Single ERP package with other systems 48.4 36.8 60.8 52.3

Multiple ERP packages with other systems 5.2 3.9 2.0 9.2

Others 1.5 1.3 2.0 1.5

Table 8

Implementation cost breakdowns

Cost breakdownSmall 4rms Medium 4rms Large 4rms Statistical

(%) (%) (%) signi4cance Software 35.14 28.70 23.38 0.001 Hardware 20.56 18.75 14.40 0.028 Consulting 23.51 29.48 25.00 No di5erence Training 9.61 9.78 12.34 No di5erence Implementation team 10.78 12.43 22.95 0.000 Other 0.49 0.83 0.86 No di5erence

large companies use this approach. Almost two-thirds of the medium-sized and 71% of the large companies use multi-ple systems compared to only 42% of the small companies. One clear distinction driving this is the complexity of the organization. Large companies are more likely to have more global operations, more sites and generally more complex operations. EventhenERP systems by themselves may not be able to provide the functionality required to manage these complex enterprises. To remedy such shortcomings, compa-nies are increasingly using either self-contained add-on ERP modules or extension systems, called Bolt-Ons, for such functions as demand planning, order tracking, warehouse management, supply chain management, customer relation-ship management, on-line collaboration, e-procurement and online business-to-business transactions. Not every ERP

sys-tem cansupport these specialized add-ons. Thus, the use of these specialized packages thenbecomes a key decisionfac-tor for not only which system is adopted, but also for how the package is implemented, and future enhancements and upgrades.

4.5. Proposition 5: costs and bene8ts

The implementationcosts reported inthe survey part were very similar to the case studies. As expected, implementa-tions at larger companies generally cost much more than at smaller companies. The cost breakdowns, however, show di5erences across companies of di5erent sizes. These are summarized inTable8. The cost of the software at smaller companies was higher as a percentage of overall cost than at

Table 9

Summary responses for performance measures

Outcomes All 4rms Small 4rms Medium 4rms Large 4rms

(%) (%) (%) (%)

Reduced direct operating costs 20.8 23.1 18.8 20.0

Quickened information response time 75.5 76.9 70.6 79.2

Improved order management/order cycle 66.3 75.0 57.6 61.9

Lowered inventory levels 35.8 35.7 34.4 38.1

Increased interaction across the enterprise 79.0 75.0 77.4 87.0

Decreased 4nancial close cycle 59.6 47.7 60.6 77.8

Improved on-time delivery 50.5 60.0 48.4 31.6

Improved cash management 26.5 26.3 23.1 31.6

Improved interaction with suppliers 44.8 40.0 55.2 38.9

Improved interaction with customers 53.6 59.0 56.7 33.3

Table 10

Summary responses for areas bene4ting

Bene4t areas All 4rms Small 4rms Medium 4rms Large 4rms

(%) (%) (%) (%)

Integration of business operations/processes 80.4 76.7 82.4 84.0

Availability of information 82.8 86.8 71.1 92.0

Quality of information 75.5 80.4 62.9 84.0

Customer responsiveness/Aexibility 36.5 41.0 37.0 26.3

Financial management 59.8 55.6 48.0 81.0

Personnel management 15.4 8.9 18.4 23.8

Decreased information technology costs 12.8 11.6 8.8 20.8

Inventory management 62.6 71.4 59.4 50.0

Supplier management/procurement 49.4 52.3 54.2 35.3

medium or large companies. The importance of these costs by company size was tested using a One-Way ANOVA test onthe data from the survey questions. Thep-values in Table8are from this ANOVA test. There are signi4cant di5erences across the three groups on the costs associated with software, hardware and the implementation teams, as indicated in the table. The implementation costs are di5erent because all of the large case study companies had created special ERP implementation teams. Team members were given special incentives to participate, increasing this com-ponent of cost. Surprisingly, the consulting and the training costs are very similar across all 4rms. All indications from the case studies were that large companies spend more on both consulting and on training. Another di5erence from ac-cepted norms is that the overall consulting costs of between 1 and 1.5 times the software costs in the survey are lower thanthe 2 to 5 times the software costs reported ina number of publications [37].

Getting a measure of success and contribution for an ERP implementationis diMcult, giventhe scope, complex-ity and timing of this type of project. Many of these systems have been implemented only recently so it may be too early to judge the full impact of anERP package at this stage.

Table 9 summarizes the impact of ERP systems onthe performance measures of key operating areas. The most improvements are in “Increased Interaction across the En-terprise”, and “Quicker Response Times for Information”. There are also improvements in order management, on-time deliveries, customer interaction and 4nancial close cycles. The least improvements are in traditional cost measures such as direct operating costs, inventory levels and cash man-agement. The ANOVA test across all three company sizes showed no signi4cant di5erences. However, there are key di5erences between large and small companies on several metrics. These pair-wise comparisons were analyzed us-ing categorical analysis. Larger companies report better im-provements in the 4nancial close cycle (p-value = 0:023). On the other hand, smaller companies have more improve-ments in order management (p-value=0:08), on-time deliv-eries (p-value=0:026) and customer interactions (p-value= 0:057).

Table10summarizes the areas bene4ting the most from ERP systems. As expected, integration of business pro-cesses, availability of information and quality of information are the areas most positively impacted. The areas bene4ting the least are the costs of information technology and

personnel management. There are also several di5erences here between companies of di5erent sizes. More large com-panies report bene4ts in 4nancial management (p-value = 0:06) and personnel management (p-value = 0:08) than small companies. On the other hand, small companies re-port higher bene4ts than large in inventory management (p-value = 0:04) and procurement (p-value = 0:08). Over-all, both the case study and the surveyed companies re-ported very similar trends on the impact of ERP systems on performance.

All of the case study companies had done some form of ROI or economic value added analysis to justify their ERP system. Surprisingly, approximately 30% of the sur-vey companies reported not doing an ROI or any form of capital investment analysis. An approximate weighted ROI for those companies who responded is approximately 20%. There were no statistical di5erences across the three groups on this dimension. This ROI is di5erent from what has beenreported inseveral publications. For example, the Meta Group [38] reported the NPV of animplementationis a neg-ative $1.5 million.

Most companies view their ERP system to be a long-term investment. The expected life of an ERP system is just over 8 years, a value considered very high for a software system. Over 80% of all companies indicted and expected life of over 5 years. Approximately a third expects the life of their ERP system to be greater than10 years.

5. Conclusions

This study provides some key insights into the implemen-tation and use of ERP systems in the manufacturing sector. Our initial case studies suggested that enterprise size played an important role in ERP implementations on several key dimensions. This was later con4rmed through an extensive survey. While there have beennumerous studies onthe im-pact of organizational size on various issues, very little has beendone to study impact of size onthe implementationand utilizationof large-scale enterprise-wide systems. This re-search shows that size is againa key factor inthe implemen-tationapproach for company-wide systems. This may have implications for manufacturing companies as they move to implement the next wave of enterprise systems such as sup-ply chain management and customer relationship manage-ment systems.

While this study covered all aspects of anERP imple-mentation, it was not designed to study such issues as the rationale for doing things in certain ways or to determine exact outcome relationships. For example, one key ques-tion that our study could not answer de4nitively is the cost and bene4t relationship. Another issue that needs to be studied is whether early adopters or late adopters received the better returns. While early adopters may have received some competitive advantages, late adopters generally ben-e4ted from upgraded systems and a better implementation

knowledge base. This raises the issue of the optimal time to start animplementationof a large system. This study is an initial 4rst step in answering such questions. ERP systems are here for the long haul and will need to be studied more thoroughly.

References

[1] Mabert VM, Soni AK, Venkataramanan MA. Enterprise resource planning survey of US manufacturing 4rms. Pro-duction and Inventory Management Journal 2000;41(20): 52–8.

[2] Mabert VM, Soni AK, Venkataramanan MA. Enterprise resource planning: common myths versus evolving reality. Business Horizons 2001;44(3):67–76.

[3] Schumpeter JA. Business cycles. New York: McGraw-Hill, 1939.

[4] Galbraith JK. Americancapitalism. Boston: HoughtonMiWin, 1952.

[5] Kamien MI, Schwartz NL. Market structure and innovation. Cambridge: Cambridge University Press, 1982.

[6] Mans4eld E. Industrial research and technological innovation: an econometric analysis. New York: Norton, 1968. [7] Rothwell R, Dodgson M. Innovation and size of 4rm. In:

Dodgson M, Rothwell R, editors. The handbook of industrial innovation. Aldershot Hants: Edward Elger, 1994.

[8] Swanson EB. Information systems innovation among organizations. Management Science 1994;40(9):1069–92. [9] Eir-Dor P, Segev E. Organizational context and the success

of management information systems. Management Science 1978;24(10):1064–77.

[10] Gremillion LL. Organization size and information technology use: an empirical study. Journal of Management Information Systems 1984;1(2):4–17.

[11] Raymond L. Organization context and information systems success: a contingency approach. Journal of Management Information Systems 1990;6(4):5–20.

[12] Harris SE, Katz JL. Firm size and the information technology investment intensity of life insurers. MIS Quarterly 1991;15(3):333–52.

[13] Yap CS. Distinguishing characteristics of organizations using computers. Information and Management 1990;18(2):97–107. [14] Kimberly JR. Organizational size and the structuralist perspective: a review, critique, and proposal. Administrative Science Quarterly 1976;21:571–97.

[15] Yasai-Ardekani M. E5ects of environmental scarcity and muni4cence on the relationship of context to organizational structure. Academy Management Journal 1989;32:131–56. [16] Mans4eld E. The di5usion of Aexible manufacturing systems

in Japan, Europe and the United States. Management Science 1993;39(2):149–59.

[17] Swamidass P, Kotha S. Explaining manufacturing technology use, 4rm size and performance using a multidimensional view of technology. Journal of Operations Management 1998;17(1):23–35.

[18] White RE, PearsonJN, WilsonJR. JIT manufacturing: a survey of implementations in small and large US manufacturers. Management Science 1999;45(1):1–15. [19] Davenport T. Putting the enterprise into the enterprise system.

[20] VanEverdingenY, VanHillegerberg J, Waarts E. ERP adoption by European midsize companies. Communications of the ACM 2000;43(4):27–31.

[21] Adam F, O’Doherty P. Lessons from enterprise resource planning implementations in Ireland—towards smaller and shorter ERP projects. Journal of Information Technology 2000;15:305–16.

[22] Abdinnour-Helm S, Lengnick-Hall ML, Lengnick-Hall CA. Pre-implementation attitudes and organizational readiness for implementing an Enterprise Resource Planning (ERP) system. Working Paper, Wichita State University, 2002.

[23] Stratman JK, Roth AV. Enterprise resource planning (ERP) competence constructs: Two-stage multi-item scale development and validation. Georgia Institute of Technology, Atlanta, Working Paper, 2002.

[24] Sarkis J, Sundarraj RP. Factors for strategic evaluation of enterprise information technologies. International Journal of Physical Distribution and Logistics Management 2000;30(3/4):196–220.

[25] Bendoly E, Jacobs FR. Alignment in operational and IT solution strategies: performance implications of within-and between-context mismatch. Working Paper, Emory University, 2002.

[26] Bendoly E, Kaefer F. The impact of enterprise systems on business-to-business and business-to-commerce transactional eMciency. INFORMS Annual Meeting, Miami Beach, 2001. [27] Meredith JR, Raturi A, Amoako-Gyampah K, KaplanB. Alternative research paradigms in operations. Journal of Operations Management 1989;8(4):297–326.

[28] Klassen RD, Whybark DC. Barriers to the management of international operations. Journal of Operations Management 1994;11(4):385–96.

[29] YinRK. Applications of case study research. Newbury Park, CA: Sage Publications, 1993.

[30] YinRK. Case study research: designand methods. Thousand Oaks, CA: Sage Publications, 1994.

[31] Orlikowski WJ. The duality of technology: rethinking the concept of technology in organizations. Organization Science 1992;3(3):398–427.

[32] Schroeder RG, Anderson JC, Tupy SE, White EM. A study of MRP bene4ts and costs. Journal of Operations Management 1981;2(1):1–9.

[33] White EM, Anderson JC, Schroeder RG, Tupy SE. A study of the MRP implementation process. Journal of Operations Management 1982;2(3):145–53.

[34] Yusuf Y, Little D. Anempirical investigationof enterprise-wide integration of MRPII. International Journal of Operations and Production Management 1998;18(1):66–86. [35] Choe J. The relationships among performance of accounting

information systems, inAuence factors, and evolution level of information systems. Journal of Management Information Systems 1996;12(4):215–30.

[36] Laughlin SP. An ERP game plan. Journal of Business Strategy 1999;20(1):32–7.

[37] Markus ML, Tanis C. The enterprise system experience— from adoption to success. In: Zmud RW, editor. Framing the domains of IT management. Cincinnati: PinnaAex, 2000. [38] Wheatley M. ERP training stinks. CIO Magazine, June 2000.