1 Introduction

Enterprise resource planning (ERP) is defined as “an integrated suite of business applications that share a common database and processes across the organisation” (Bond et al. 2000). The benefits of an ERP system may be realized through increased standardization of processes, which in turn enhance efficiency, leading thereby to cost savings. The ERP systems evolved from MRP I (Material Requirement Planning) and MRP II (Manufacturing Resource Planning). MRP I, launched in the late 1960s was a “technological system for inventory management and production planning”. MRP II has been defined as a “system for the effective planning of various activities in a manufacturing enterprise” (Alwabel et al. 2006). While both systems added a lot of value for the manufacturing industry at large, their limitations led to the emergence of the ERP system, having capability to integrate different internal business functions and processes. With central database of all entities as well as various transactions, ERP found numerous applications for external stakeholders like customers (CRM, E-Commerce) and suppliers (SCM, SRM) (Vrat 2014).

The ERP market is expected to reach USD 61.69 billion globally by 2025 (Wilson 2019). ERP software is available and accessible through two platforms known as in-premise ERP and cloud-based ERP. As the name suggests, major difference between the two includes its location and ownership, both of its software and data. Examples of in-premise ERP include SAP, Oracle, Tally etc.; they are deployed in the firm’s own computers and are largely managed internally. Cloud-based ERP systems like ZOHO, ODOO, etc., also known as software as a service, are hosted on the vendor’s servers, managed by experienced and trained professionals at data centres, which are owned by the vendors and accessed by end-users through a web browser (Duan et al. 2013; Peng and Gala 2014). The difference between the two (i.e., in-premise vs. cloud-based ERP systems) is that the cloud-based systems have reduced infrastructure needs, which lead to reduction total cost of ownership. Additionally, cloud-based ERPs are considered to be eco-friendly (Jain and Sharma 2016). For in-premise ERP systems like SAP, Sage, MS-Dynamics etc., customers can procure them in commercially-off-the-shelf (COTS) form or may even opt for customised systems to suit their particular needs. Additionally, for in-premise ERPs, the alternative could also be to choose from industry agnostic, general-purpose ERP (Horizontal) systems, which could be used for integrating generic processes like procure to pay (P2P), order to cash (O2C) or vertical-specific (i.e., steel, petrochemical, hospital, etc.) systems, specifically designed to cater to industry related processes (Rosa et al. 2013). In recent times, ERP products have been bundled with CRM, SCM and BI applications. Therefore, customers have the choice of procuring either a stand-alone ERP system or a complete bundled enterprise application (Gattiker and Goodhue 2004; Wieder et al. 2006).

Decision-makers in organizations, including those from the SMEs, have to do rigorous due-diligence to select the most appropriate ERP system, considering their context and constraints (Kahraman et al. 2010). Several internal and external factors influence this strategic decision (Illa et al. 2000). Secondary research by the researchers of this study, identified numerous factors that tend to affect ERP selection decision within a firm; however, most of these researches have been conducted with large companies (Ziaee et al. 2006). To constitute a final list of factors from a list identified through literature review, exploratory factor analysis (EFA) was used. Moreover, fuzzy analytical hierarchy approach (FAHP) was employed to calculate the weight of the factors and rank them accordingly. The findings of this study are expected to assist owners and managers of Indian SMEs in identifying and prioritizing factors that go on to influence an ERP selection process. It could also act as a point of reference for SMEs in other developing nations too. This paper aims at answering the research questions (RQs) given below:

  1. RQ1.

    What are the various factors that influence Indian SMEs in selecting ERP system?

  2. RQ2.

    Which of the factors are more important for Indian SMEs?

  3. RQ3.

    What is the relative importance and ranking of the factors critical to ERP selection decisions?

Remaining paper is organised as follows: section 2 provides the review of extant literature describing the importance of ERP for achieving competitive advantage, along with the factors that influence its selection. Section 3 presents the research methodology, application of the FAHP method and results along with validation. Section 4 provides the discussion. Section five presents the recommendations derived from the findings. The last section includes the conclusion of research and scope for future research.

2 Literature review

Information systems, strategically aligned with firms, have become indispensable for gaining competitive advantage (Zhang and Lado 2001; Piccoli and Ives 2005). Fierce competition has led firms to switch to more reliable information systems (Baki and Cakar 2005; Han et al. 2011), which possibly would help in smooth execution of various processes, such as planning, procurement, manufacturing and delivery (Pfeffer 2005). Importantly, all of this should be achievable without exceeding the budgets; in fact, the costs should be minimised in all the processes adopted. One way of maintaining low cost is by reducing waste and increasing efficiency (Zhu and Sarkis 2004). To this end, ERPs are an information or computer system that act as a decision tool, which can ensure the effectiveness and efficiency for firms adopting it (Shang and Seddon 2000; Gattiker and Goodhue 2004; Holsapple and Sena 2005). ERP is demarcated as the “integrated suite of software programmes that links back-end and front-office activities with the downstream and upstream supply chains” (Verville et al. 2007; Badewi et al. 2015). It is a “corporate strategy along with domain-specific applications which builds value for stakeholders by optimizing and enabling intra and inter-firm processes” (Bond et al. 2000).

Firms should select an appropriate ERP package, which aligns with their present needs and future ambitions (Al-Mashari et al. 2003; Presley 2006; Kilic et al. 2014). Upgrading to technological platforms is a crucial decision taken by top management, that affects the performance and competitive advantage of firms (Ahn and Choi 2008; Sood et al. 2019). Selection and successful implementation of ERP help firms to overcome the limitations of their legacy systems. For instance, the departments within firms that do not use ERPs, often tend to work in silos (Gardner 2016), which in turn hampers their productivity (Albrecht 2002). On the contrary, an ERP system not only collates information on a single platform, but it also provides solutions to the business problems. ERPs align all business-related activities, standardise business processes, ensure availability of real-time information, enable smooth communication, increase inter and intra firm collaboration and enhance the robustness of decision-making capabilities (Hwa Chung and Snyder 2000; Hwang and Min 2015). The other benefits of ERP include enhanced customer service, business growth, robust distribution systems and lower cost of operations (Dwivedi et al. 2009).

Despite the post-implementation benefits of ERP, its adoption is still a challenge (Pan et al. 2007; Helo et al. 2008). Therefore, selecting the most appropriate ERP involves various managerial contemplations like choosing from the options available, matching them as closely as possible to the needs of the organization, and finally selecting that within a specified budget. For example, in the past, wrong ERP implementation decisions have proved to be highly expensive even to renowned firms like Hershey foods, Nike, HP, etc. (Hwang and Min 2015). This ascertains thereof the criticality of the ERP selection decision, more so for firms, which work on a relatively smaller scale of operations and a limited budget, such as SMEs.

SMEs are firms, which are neither too small nor very large. In fact, globally, there is no unanimous definition of SMEs (Maduku et al. 2016). However, researchers and institutes working for SMEs, tend to adopt the contextual definition, depending mostly on the economic and infrastructural development of countries (Keskin and Sentürk 2010). In India, the Development Commissioner, Ministry of Micro, Small and Medium Enterprises (MSME) mostly make these classifications, through the MSME Development (MSMED) Act 2006. It defines SMEs as one with an investment in machinery and plan between INR 2.5 million to INR fifty million.Footnote 1 Similarly, a medium enterprise is classified as the one investing INR fifty million to 100 million in plant and machinery (dcmsme.gov.in 2019).

MSMEs current contribution in the Indian GDP is approximately thirty percent; this sector provides employment to 110 million people. The Indian government acknowledges that the contribution of this sector is very crucial in the economic development of the country (Dewan 2019). Mr. Nitin Gadkari, Minister for MSME, Govt. of India opined that, his ministry will work hard to raise the contribution of the MSME sector to half of country’s GDP from the current present 29 percent, in the coming years, and ensure higher employment to minimum 150 million compared to existing 111 million people (Press Information Bureau 2019).

An SME’s risk-taking appetite is much less than big corporates (Eniola and Entebang 2015). For SMEs, one decision gone wrong can swipe them off their business forever; and the ERP package selection decision is one such decision. A compatible ERP selection and implementation can improve the efficiency of a firm and similarly, improper ERP package selection would adversely affect the productivity of the firm. Researchers have proposed various criteria, which influence the decisions to select an ERP. For instance, decisions on “cost, technical and vendor specifications and ease of use” (Efe 2016); or “technical, corporate and financial criteria” (Kilic et al. 2014); or “vendor credibility, flexibility, product functionality, implementation methodology, support, customer focus and strategic plans” (Ünal and Güner 2009) and/or “functionality, quality, price, market leadership, implementation time, other interface, global focus” (Kahraman et al. 2010).

Similarly, researchers have also highlighted the role of sub-criteria in influencing adoption decisions; these include “price, consultant expenses, cost related to infrastructure and maintenance; security, function-fit, ease of operation, learning, integration and in-house development; upgrade ability; stability, recovery ability; scale and financial condition of vendor; capability related to R and D and technical support, implementation ability; warranties, training service, service speed” (Alanbay 2005; Wei et al. 2005).

It is clear that ERP package and vendor selection depend on number of criteria and sub-criteria as stated above. Extant literature regarding vendor selection for ERP packages have also emphasised the role of the multi-criteria decision-making techniques (Oztaysi 2014; Efe 2016). Many researchers in fact have emphasized on using techniques such as AHP, TOPSIS, PROMETHEE, Fuzzy techniques etc. for rational decision-making (Lien and Chan 2007; Demirtaş et al. 2011; Kilic et al. 2015). The basic premise for selecting an ERP hinges on the premise that such decisions should be integrated with the strategic as well as functional requirements of the adopting firm. In order to assist the SME decision-makers in selecting the appropriate ERP package, this research identifies and ranks the criteria and sub-criteria according to its importance. A brief overview of various factors identified have been described below:

2.1 Experience of ERP vendor

A richer experience of the vendor manifested in terms of their market share, the number of consultations done and installations performed, improves the probability of their selection (Huang and Palvia 2001). Moreover, the extent of infrastructural and other support provided, along with a detailed information about their previous projects in the proposal, are a few other critical factors that reflect the vendor’s commitment (Alanbay 2005). Wei et al. (2005) suggested that it is advisable to collect all the information available about the vendors and their systems. For instance, buyers can seek help of well-formulated questions in order to obtain holistic information about the vendors. This may also include the support services provided by the vendor’s post-implementation.

Client’s experience assists in identifying vendors aligned to their philosophy and domain (Ahn and Choi 2008). The vendors’ experience boosts client’s confidence about effective project management. Apart from the total experience, recent experience is considered as most critical, which tends to affect the robust financial status of a vendor. A customer on the other hand, should be concerned not only about the vendor’s past financial stability, but also needs to pay heed to glaring signs of crisis, which may loom large on the vendor’s business model, in order to ensure their uninterrupted support during the entire life cycle of ERP implementation within the firm (Zach 2011).

2.2 Client’s credentials

Client credentials as a factor has been one of the most significant considerations for vendor selection (Ahn and Choi 2008). In other words, client’s credentials include the feedback of ERP vendor’s previous as well as current client base. Profiles of their existing and past clients elucidate the ways and levels of the service, which the vendor can provide (Gefen 2004). In fact, the details of the vendor's clients help firms in determining some clear expectations from the vendors; for example, if the vendor only caters to large organizations, then, it may not be suitable for SMEs. Further, client credentials help in establishing whether the vendor is actually cooperative, compatible and successful.

2.3 Research and development

R and D capability along with an upgrade path (future releases) of an ERP package provided by the vendor, does play an important role in the decision-making process of adopting an ERP system. R and D capability may be judged in many ways (Unal and Guner 2008; Cebeci 2009), and is reflected in the investment being made, number of R&D professionals employed by the vendor, and their experience (Wei et al. 2005). Most importantly, the intention to upgrade vis-a-vis the actual upgradation of existing technologies by the vendor also explains his/her R and D capabilities, which the ERP adopting firm needs to consider (Çakır 2016).

2.4 Capability of manpower

The aspect of human expertise is another salient criterion for choosing the right vendor. Employees in any organization are considered to be ‘key assets’, more so, in service-intensive firms. They are central to the core functions and values that an ERP vendor bring on to the table (Ghodeswar and Vaidyanathan 2008). Factors related to employees, such as qualification, experience and their product and process knowledge reflect in their services that they provide. Hence, it is desirable that the staff of the vendor is selected with care in order to ensure that they’re well trained and are motivated (Alanbay 2005). While the technical ability of the employees is certainly one of the most important factors (Wei et al. 2005), a cooperative vendor is equally preferred because employees of cooperative vendors behave responsively and responsibly to the needs of customer (Wei et al. 2005; Cebeci 2009). Loyal and skilled employees who work well individually and in teams go on to provide a huge competitive advantage to firms, regardless of whether they are from the client/vendor.

2.5 Reputation

Vendor reputation is also regarded as a significant indicator in ERP selection (Efe 2016; Çakır 2016); this is reflected through vendor’s brand name, market positioning, financials, certifications and awards. It also determines the probability of long-term client satisfaction. Several researchers have proposed that a prompt way of assessing the reputation of a vendor is by scrutinising professional magazines, taking feedback in exhibitions, referring to yearbooks, browsing the internet, along with several other sources (Wei et al. 2005). The procurers should also be cautious of the financial health of the vendor firms, such as their debt-credit ratio, top and bottom line, possibility of takeover or merger etc. This information could then be used for shortlisting competent vendors (Kumar et al. 2002; Baki and Çakar 2005).

2.6 Functionality

Functionality of the ERP package plays a key role in evaluating the package (Ahn and Choi 2008); it relates to some core features and functions of the system, including the functional fitness of the software and even the security features of the package (Wei et al. 2005). Essentially, the functional fitness ensures that the functional capability of a software package per se fulfils the current as well as the future needs of adopting firms (Verville et al. 2007; Kahraman et al. 2010). Many researchers have emphasised on the need for understanding the security facets related to ERP (She and Thuraisingham 2007).

2.7 Solution to current concerns

With ERP, firms are able to take the right sourcing decisions, manage their inventories, optimize their production and enhance the level of customer service (Krumbholz and Maiden 2001). ERP packages provide many benefits to firms, which include integration and support for all the functional departments. With increased integration comes increased flexibility (Lee et al. 2003). The initial ERP systems used to integrate and bring the entire firm on one platform. However, contemporary ERP systems are not only integrated within firms, but also take the integration beyond the firms to their partners in the supply chain using modules like SCM and CRM (Pan et al. 2011). ERP systems today, reengineer an adopting firm’s current processes, and do have the potential to increase the efficiency and effectiveness of the firms through improved intra-firm and inter-firm business practices.

2.8 Facilitation for future strategy

Firms, which are at the verge of making technology adoption decisions always prefer systems, which align and help in sustaining their business goals (Velcu 2010; Sood and Jain 2020). In the case of selecting an ERP software, firms tend to prefer an ERP vendor that is capable of catering to changes even in their future business strategies (Al‐Mashari and Al‐Mudimigh 2003). These changes could be driven by both internal and external drivers like management vision, competitive pressures, customer expectations, changes in legal/ regulatory framework, along with other emerging disruptive business and technological needs and trends. A successful business and IT strategy alignment supports operations and makes firms more competitive (Bakås et al. 2007; Chen et al. 2008).

2.9 Employee’s comfort

Employees of ERP adopting firms become comfortable with the legacy and traditional systems that they use and do not wish to learn or move-to new ways of doing their job. Employee’s comfort with the existing processes does therefore hamper their adaptive ability to new-fangled technologies and processes (Torfi and Rashidi 2011). Hence, if the employees are made comfortable with a new technology, the transition process in adoption would smoothen and become easier. Thus, training the employees appropriately is extremely crucial, as that would help in building their skillsets as well as give them the confidence to adopt and work in a new technology thereof (AL-Ghamdi 2013; Sood and Jain 2013). Other than training, the adopting firm’s culture along with their rewards and recognition (R and R) structures also assist in enhancing the employees’ comfort levels (Motwani et al. 2005).

2.10 Management’s comfort

The top management has to be persuaded about the technology because unless the top management is convinced about it, the message cannot go down well at the bottom level. Since, ERP implementation leads to change in the entire process setup, it calls for micro-management, further justifying thereby their role in optimising the gains from this change (AL-Ghamdi 2013). Top management support and conviction help overcome employee resistance coupled with other challenges (Lien and Chan 2007). To this end, managers generally follow transactional or transformational styles (Ke and Wei 2008; Torfi and Rashidi 2011).

2.11 User-friendliness

User friendliness relates to the ease of learning and operations (Wei et al. 2005). It is a crucial success factor for ERP selection and implementation. For faster and hassle-free adoption, the software should be simple to understand and easy to use. A perfect ERP package is the one, which caters to the end-user requirements effectively, while maintaining its characteristics of reengineering and integration. User-friendliness shortens the adaptation phase and lessens the requirement of employee training (Karsak and Özogul 2009). It is preferred that the chosen software package provide reporting as well as self-help functions through multiple touch points i.e., user manuals, guidebooks, help feature in each module, help desk etc. Complicated user interface of the software may lead to inefficiency of the employees (Nah et al. 2005).

2.12 Support offered by ERP vendor

Post implementation service and support provided by the vendor has a direct and proportional influence on vendor reputation (Kilic et al. 2015). There is always an underlying expectation from the vendor to be able to extend help to their users; in other words, the system should always be capable of supporting client’s needs (Alanbay 2005; Ahn and Choi 2008). Thus, after-sales support becomes crucial because ERP applications are quite complex and expensive; hence, demand continuous support as and when required. Therefore, it becomes utmost important for adopting firms to ascertain the level and extent of the support services post implementation (Karande and Chakraborty 2012).

2.13 Time to deploy

Deployment time of the technology has been identified as an important criterion in information technology selection and adoption literature (Ünal and Güner 2009). Given that the decision-making and disposition of technology are time consuming, adopting firms prefer that the technologies be deployed in the least time possible without compromising the desired quality (Demirtaş et al. 2011). Here, the time of deployment includes both the package implementation time as well as the time taken to train the employees. Interestingly, time of deployment is influenced by the maintenance of the schedules, level of customisation expected by the firms, and the degree of cooperation provided by the employees of the adopting firm (Wei et al. 2005; Lien and Chan 2007).

2.14 Information security

ERP has the potential to integrate internal business processes of the adopting firm, but can also extend to their suppliers as well as customers. This end-to-end integration leads to a seamless interchange of information among all the connected entities. This flow of information in turn can only be smooth when all stakeholders are confident about the ‘security’ of their information. Security and privacy concerns often limit the adoption of technological platforms (Kiadehi and Mohammadi 2012; Hanine et al. 2016). Information security concerns include aspects such as confidentiality of the information shared, availability of the desired information, integrity, authentication and non-repudiation (Kahraman et al. 2010; Daneva 2006). Further, like for other new technologies and innovations, right information security policies should be in place to encourage the diffusion of ERP security too (McKnight and Kacmar 2006; She and Thuraisingham 2007).

2.15 Installation cost

It is well acknowledged and accepted that ERP implementation is not only complex but costly as well (Yaseen 2009). The implementation cost includes the expenses related to the trials conducted, fees paid to the consultants, software, hardware, accessories, databases, networking, skilled labour, licensing, etc. (Dowlatshahi 2005; Hellström 2009; Jain and Sharma, 2016). The cost and budget allocations for technologies depend on the total asset’s vis-a-vis the total sales of the adopting firm. Cost allocations for ERP implementation have been found to range between 0.5 to 3.5% of the annual revenues. Interestingly, adopting firms also take ERP implementation decisions to reduce their operating costs. Importantly, the installation costs are also impacted by the level of customisations availed (Kanellou and Spathis 2013).

2.16 Maintenance and up-gradation cost

An ERP package once deployed, in order to be effective, needs to be updated regularly to incorporate new features and technologies. While choosing the software and the implementation partner, firms need to evaluate them based on the tentative maintenance costs to be incurred in the future (Seethamraju and Seethamraju 2008). Maintenance and upgradation costs, also known as the running cost, primarily include expenses related to the replacement of damaged or outdated systems and hardware; it also includes upkeep of operating-systems and software (Hellström 2009). The annual maintenance costs generally vary, but can be up to thirty percent of the initial implementation costs. Maintenance cost is inevitable, as it helps adopting firms perform in full capacity of their ERP systems (Dowlatshahi 2005; Newman and Westrup 2005).

Following table summarises factor wise, industry wise literature review across countries Table 1.

Table 1 Factors for ERP package selection
Full size table

3 Research methodology and results

As depicted in Fig. 1, this section discusses the methodological steps followed in this study. Section 3.1 demonstrates the application of exploratory factor analysis (EFA) for factor reduction and constructs identification. The subsequent sub-sections explain the application of fuzzy AHP along with computation steps followed by the application of sensitivity analysis to ensure the robustness of the fuzzy AHP results.

Fig. 1
figure 1

Research methodology

Full size image

This mix-method research involved triangulation design by merging both quantitative and qualitative techniques. Initially, owners and senior managers of selected organizations, from SME sector, having authority and intention of taking decision about augmenting their firms’ competitiveness through ERP system were approached. They were asked to provide their preferences to sixteen factors identified through literature review on a seven points likert scale, affecting their decision regarding selection of ERP system in their firms. A total of 125 responses (Table 2) were collected to conduct EFA for finalizing factors for further analysis using FAHP approach. During second stage of the study, twelve experts from SMEs belonging to the manufacturing sector provided their responses on the relative importance of the finalised four criteria and ten sub-criteria in the form of linguistic variables on a nine point scale, suggested by Saaty (1990). In the third stage, research findings were validated through qualitative research, where four SME owners were contacted, who had already implemented ERP software in their firms in the last three years.

Table 2 Respondent’s profile for EFA
Full size table

3.1 Exploratory factor analysis (EFA)

EFA is a statistical tool deployed for data reduction (Lundqvist 2014). This tool helps in identifying the existence of significant patterns among the original variables (Lu and Shang 2005). The objective of using EFA is to decide the number of common factors impacting a set of measures and determine the strength of relationship among the factors leading to develop the groups of constructs. Researchers began the analysis with 16 constructs (Table 3), out of which, the latent variables were identified using EFA.

Table 3 Factors for EFA
Full size table

EFA was performed using Kaiser-Mayer Olkin test (KMO) and Bartlett’s test. KMO validates the sample adequacy of the data (Abdallah and Hilu 2015) and its value over 0.50 is acceptable for EFA (Marshall et al. 2007). IBM SPSS 20 was used to conduct EFA. The KMO value of the data here was found to be 0.721, confirming its suitability for EFA. Bartlett’s test is performed to check, whether a sample qualifies for a multivariate normal distribution (Abdallah and Hilu 2015) and its value should be less than 0.05. This value was found to be highly significant (χ2 = 504.267; df = 120; p = 0.000 < 0.05). Moreover, the application of confirmatory factors analysis is not suitable for this study as there is a lack of prior theory that can be employed to ten measurement items, as suggested by Dhochak and Sharma (2016).

The anti-image matrix was attained and items whose values were lesser than 0.5 were eliminated (Marshall et al. 2007). Thus, six factors out of 16 were dropped in this process. These six factors were research and development, functionality, employees comfort, management comfort, support offered by the vendor and time to deploy. This can be understood as the SMEs want an ERP package largely for integrating their existing processes that do not change often, therefore; research capability of vendor and functionality of package or richness of features are not much relevant to them. Further, there is not much process changes involved during ERP implementation at SMEs, therefore, comfort of employees and management do not matter much. Small or no-frills package preferred by the SMEs can be implemented relatively rapidly, therefore, deployment time is relatively less important. Finally, SMEs require only basic support from the vendor due to standard processes.

Varimax method of orthogonal rotation produced four constructs having eigenvalue over one. Table 4 provides the list of ten factors considered for further analysis (Refer Appendix 1 for details). Four constructs extracted from EFA were referred as “Vendor Credibility”, “Need Fulfilment”, “User Friendliness and Security” and “Cost of Deployment”. As shown below, the factor loadings values represent variance explained and Cronbach’s alpha (widely used measure of reliability) for the four items were: 0.737, 0.714, 0.624, 0.771, respectively. This measure shows good reliability as the values obtained are above 0.6 (Abdallah et al. 2015) and confirms that the survey items appropriately measured the underlying constructs.

Table 4 Construct’s loadings and reliability
Full size table

3.2 Application of fuzzy AHP

Selection of appropriate ERP package is one of the crucial strategic decisions for any organization for achieving competitiveness. Such decisions require consideration of many factors/ criteria at different hierarchical levels. Capability of multi-criteria decision-making (MCDM) techniques in dealing with multi-dimensional and complex decisions appealed many researchers to employ these methods for analysing several types of strategic problems including sustainable supplier selection (Mohammed et al. 2019), cloud computing technology selection (Büyüközkan et al. 2018), cloud service selection (Nawaz et al. 2018), third-party logistics provider (Raut et al. 2018), quality improvement program selection (Zhou et al. 2018) etc. There are variety of methods that fall under the umbrella of MCDM such as SAW (simple additive weightage), AHP (analytic hierarchy process), ANP (analytic network process), WPM (weighted product method), WSM (weighted sum method), TOPSIS (technique for order of preference by similarity to ideal solution), ELECTRE, PROMETHEE, and VIKOR. However, AHP (Saaty 1980) found wider acceptance and preference among MCDM methods, due to its potential to blend both qualitative and quantitative approaches in the evaluation and ranking of decision alternatives in complex decision-making scenario (Wong and Li 2008; Mathiyazhagan et al. 2014; Calabrese et al. 2019).

According to Saaty (1980), AHP, a powerful group decision making process, allows the comparative assessment of several criteria and alternatives at different hierarchical level through decomposition of the complex problem. AHP uses subjective judgments of the decision makers to develop pair-wise comparison matrices for each hierarchical level and provides the estimation of relative weightages of criteria and alternatives at each level. However, in the complex decision-making scenario, linguistic subjective assessments rendered by the decision makers are largely based on their experience and may not be precise (Zhou et al. 2011). To address the issues of uncertainty and vagueness in assigning the importance of criteria over each other, concept of fuzzy set theory (Zadeh 1965) needs to be incorporated with traditional AHP methodology. Fuzzy set theory allows the mapping of linguistic variables between values 0 (complete absence) and 1 (Complete presence) within fuzzy set.

Therefore, this study employs fuzzy AHP (FAHP) method to rank the criteria and alternatives for the selection of appropriate ERP package considering its ability to address issues of uncertainty and fuzziness in decision maker’s relative assessments (Awasthi et al. 2018). Though, various types of fuzzy numbers such as trapezoidal fuzzy number (Zheng et al. 2012), rectangle fuzzy number (Secundo et al. 2017), triangular fuzzy number (Lan et al., 2016), and continuous fuzzy number (Cevik et al. 2014) were used to describe the uncertainty and fuzziness in the group decision making process; fuzzy triangular numbers allow effective depiction and handling of linguistic variables for the purpose of analysis (Calabrese et al. 2019). Figure 2 demonstrates the hierarchical structure of criteria and sub-criteria selected for this study. Table 5 presents fuzzy linguistic terms used for FAHP analysis.

Fig. 2
figure 2

AHP model

Full size image
Table 5 Fuzzy relative importance scale used for making pair-wise comparison (Balusa and Gorai 2019; Ayhan 2013; Saaty 1980)
Full size table

Experts from various manufacturing SMEs having relevant knowledge and experience (Table 6) were initially requested to provide their feedback on various criteria and sub-criteria identified for ERP package selection problem. Based on their recommendations, final hierarchical structure was developed for further investigation. Furthermore, experts were asked to give relative ratings and weights for the selected criteria and sub-criteria to develop pairwise comparison matrices. Responses received from twelve experts in the form of linguistic variables ranging from “equally important” to “extremely important” were converted in to triangular fuzzy numbers ãij = (lij, mij, uij) as depicted in Table 5. Pairwise comparison matrices were developed by computing geometric mean of the received responses acknowledging its consistency and ability to uphold the reparability, unanimity and homogeneity in aggregating individual judgments (Mosadeghi et al. 2015; Thanki et al. 2016).

Table 6 Profile of experts for FAHP
Full size table

The steps for the Fuzzy AHP method are explained for a criterion to criteria pairwise comparison matrix as shown below (Tables 7, 8, 9, 10, 11, 12, 13, 14) (Balusa and Gorai 2019).

Table 7 Criteria to criteria (C to C) fuzzified pair-wise comparison matrix for ERP system selection model
Full size table
Table 8 Fuzzy geometric mean value for C to C matrix
Full size table
Table 9 Fuzzy weights for C to C matrix
Full size table
Table 10 de-fuzzified and normalized weights for criteria
Full size table
Table 11 Crisp comparison matrix
Full size table
Table 12 RI and recommended CR values
Full size table
Table 13 Consistency check
Full size table
Table 14 Local weights, global weights and ranking of criteria and sub-criteria
Full size table
  1. Step 1:

    formation of fuzzy-relative importance matrices

    The relative importance matrices were built by using scale given in Table 5 and were transformed into fuzzy matrices through following equation:

    $$ \overline{x}_{a} = [x{ - }\alpha ,x + \alpha ];\frac{1}{x} = \left[ {\frac{1}{{x{ - }\alpha }},\frac{1}{x + \alpha }} \right] $$
    (1)

    Here α ranges from 0 and 1 and higher values of α signify greater uncertainty. By employing Eq. (1), the relative importance of criteria and sub-criteria for ERP package were transformed into fuzzy matrices. This research has applied six different α values – 0, 0.2, 0.4, 0.6, 0.8 and 1 to perform sensitivity analysis. However, various steps for fuzzy AHP analysis (step 1–step 6) are demonstrated for α = 1 in the following discussion. The sensitivity analysis was performed considering six different values of α as mentioned above.

    The fuzzified pairwise comparison matrix can be developed as,

    $$ \tilde{A} = (\tilde{a}_{{{\text{ij}}}} )_{{{\text{nxn}}}} = \left[ {\begin{array}{*{20}c} {(1,1,1)} & {(l_{12} ,m_{12} ,u_{12} )} & \cdots & {(l_{{1{\text{n}}}} ,m_{{{\text{1n}}}} ,u_{{{\text{1n}}}} )} \\ {(l_{21} ,m_{21} ,u_{21} )} & {(1,1,1)} & \cdots & {(l_{{{\text{2n}}}} ,m_{{{\text{2n}}}} ,u_{{{\text{2n}}}} )} \\ \vdots & \vdots & \vdots & \vdots \\ {(l_{{{\text{n1}}}} ,m_{{{\text{n1}}}} ,u_{{{\text{n1}}}} )} & {(l_{{{\text{n2}}}} ,m_{{{\text{n2}}}} ,u_{{{\text{n2}}}} )} & \cdots & {(1,1,1)} \\ \end{array} } \right] $$

    where,

    $$ \tilde{a}_{{{\text{ij}}}} = (l_{{{\text{ij}}}} ,m_{{{\text{ij}}}} ,u_{{{\text{ij}}}} ) = (\tilde{a}_{{{\text{ij}}}} )^{ - 1} = \left( {\frac{1}{{u_{{{\text{ji}}}} }},\frac{1}{{m_{{{\text{ji}}}} }},\frac{1}{{l_{{{\text{ji}}}} }}} \right), \, i,j = 1,...{\text{n}};{\text{ i}} \ne {\text{j}} $$

    Table 7 presents the pairwise comparisons established by twelve experts for four criteria of ERP system selection. Each expert was requested to provide his/her response in linguistic term (see Table 5) by comparing each criterion against other criteria for its importance in ERP system selection decision making. The linguistic terms were converted into appropriate fuzzy scale and then, geometric mean values of twelve responses were used to develop Table 7.

  2. Step 2:

    calculating fuzzy geometric mean value \(\left( {{\tilde{\text{r}}}_{{\text{i}}} } \right)\)

    $$ \tilde{r}_{{\text{i}}} = \left[ {(l_{{{\text{i}}1}} * l_{{{\text{i}}2}} * \cdots * l_{{{\text{i}}n}} )^{{\frac{1}{{\text{n}}}}} ,(m_{{{\text{i}}1}} * m_{{{\text{i}}2}} * \cdots * m_{{{\text{in}}}} )^{{\frac{1}{{\text{n}}}}} ,(u_{{{\text{i}}1}} * u_{{{\text{i}}2}} * \cdots * u_{{{\text{in}}}} )^{{\frac{1}{{\text{n}}}}} } \right] $$

    Table 8 exhibits the fuzzy geometric mean value for each criterion. These values were further used to calculate fuzzy weights of the criteria in the next step.

  3. Step 3:

    calculating fuzzy weights (\(\tilde{w}_{{\text{i}}}\))

    $$ \tilde{w}_{{\text{i}}} = \tilde{r}_{{\text{i}}} \otimes \left( {\tilde{r}_{1} \oplus \tilde{r}_{2} \oplus \cdots \oplus \tilde{r}_{{\text{n}}} } \right)^{ - 1} $$

    Table 9 depicts the fuzzy weight for each criterion computed using step 3. These fuzzy weights were de-fuzzified to obtain crisp weight value for each criterion.

  4. Step 4:

    calculating de-fuzzified and normalized weights.

    For de-fuzzification: \(w_{{\text{i}}} = \left( {\frac{{l_{{\text{i}}} + m_{{\text{i}}} + u_{{\text{i}}} }}{3}} \right)\).

    The de-fuzzified weights (wi) obtained through step 4 were normalised to obtain the relative importance of each criterion (Table 10).

  5. Step 5:

    consistency check

    1. Step 5.1:

      Conversion of fuzzy matrices

      Using formula given below, a crisp comparison matrix (Table 11) was derived from the fuzzy comparison matrix (Table 7).

      $$ A = \left( {\tilde{a}_{{{\text{ij}}}} } \right) = \left( {\frac{{l_{{{\text{ij}}}} + m_{{{\text{ij}}}} + u_{{{\text{ij}}}} }}{3}} \right) $$

      Table 11 presents the crisp comparison matrix required to measure the inconsistency of the pairwise comparison matrix (Table 7), which was derived based on the expert’s judgments in criteria comparison process. As inconsistency is underlying in judgment process (Saaty 1980) which may lead to biased results, it is important to check the consistency of the matrix, once weights are computed to ascertain the quality of the decision.

    2. Step 5.2:

      calculation of λmax, consistency index and consistency ratio

      $$ \lambda_{{{\text{max}}}} = \frac{1}{{\text{n}}}\left( {\frac{{W^{^{\prime}}_{1} }}{{W_{1} }} + \frac{{W^{^{\prime}}_{2} }}{{W_{2} }} + ..... + \frac{{W^{^{\prime}}_{{\text{n}}} }}{{W_{{\text{n}}} }}} \right) $$

      where,

      $$ W^{^{\prime}} = {\text{AW}} = \left[ \begin{gathered} W^{^{\prime}}_{1} \hfill \\ W^{^{\prime}}_{2} \hfill \\ \, \vdots \hfill \\ W^{^{\prime}}_{{\text{n}}} \hfill \\ \end{gathered} \right] $$
      $$ {\text{Consistency index(CI)}} = \frac{{\lambda_{\max } - n}}{n - 1} $$
      $$ {\text{Consistency ratio(CR)}} = \frac{{{\text{CI}}}}{{\text{Random index(RI)}}} $$

      Table 12 exhibits the recommended consistency ratio (CR) values for matrices of size 1–10. The judgments provided by the experts are considered acceptable if CR is less than 0.10, else there is possibility that judgments are made arbitrarily.

      The CR value obtained with step 5.2 is shown in Table 13 for criteria to criteria pair-wise comparison matrix. As CR is less than limiting value indicated in Table 12, we can conclude that the judgments made by the experts are consistent.

      The local and global weights for sub-criteria were also computed using the steps discussed above. Table 14 presents the ranking of criteria and sub-criteria based on the weights obtained. This ranking explicates the importance of respective criteria and sub-criteria in ERP system selection process for Indian SMEs.

3.3 Sensitivity analysis

By varying the fuzzification factor α in Eq. (1) and decision–making attitude (λ) in Eq. (2), sensitivity analysis was conducted for the proposed model. The λ was considered for following conditions—optimistic (λ = 1), pessimistic (λ = 0) and neutral (λ = 0.5), respectively. Subsequently, the decision-making model output was analysed for each combination of α (six α sets: 0, 0.2, 0.4, 0.6, 0.8 and 1) and λ (three λ sets: 0, 0.5, 1). The results highlighted that the ranking were not altered by either changing the decision-making attitude or changing the fuzzification factor from 0 to 1. The sensitivity of the decision-making results are given below:

Thus, results presented from Tables 15, 16, 17, 18, 19 and Figs. 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17 prove therobustness of the FAHP results for selection of ERP package.

Table 15 Global weights of four criteria for different decision-making attitudes (λ) and fuzzification factors (α) for ERP package selection
Full size table
Table 16 Global weights of sub-criteria (vendor selection) for different λ and α for ERP package selection
Full size table
Table 17 Global weights of sub-criteria (Need Fulfilment) for different λ and α for ERP package selection
Full size table
Table 18 Global weights of sub-criteria (user friendliness and security) for different λ and α for ERP package selection
Full size table
Table 19 Global weights of sub-criteria (cost of deployment) for different λ and α for ERP package selection
Full size table
Fig. 3
figure 3

Sensitivity analysis for criteria (λ = 0)

Full size image
Fig. 4
figure 4

Sensitivity analysis for criteria (λ = 0.5)

Full size image
Fig. 5
figure 5

Sensitivity analysis for criteria (λ = 1)

Full size image
Fig. 6
figure 6

Sensitivity analysis for sub-criteria (vendor selection) (λ = 0)

Full size image
Fig. 7
figure 7

Sensitivity analysis for sub-criteria (vendor selection) (λ = 0.5)

Full size image
Fig. 8
figure 8

Sensitivity analysis for sub-criteria (vendor selection) (λ = 1)

Full size image
Fig. 9
figure 9

Sensitivity analysis for sub-criteria (need fulfilment) (λ = 0)

Full size image
Fig. 10
figure 10

Sensitivity analysis for sub-criteria (need fulfilment) (λ = 0.5)

Full size image
Fig. 11
figure 11

Sensitivity analysis for sub-criteria (need fulfilment) (λ = 1)

Full size image
Fig. 12
figure 12

Sensitivity analysis for sub-criteria (user friendliness and security) (λ = 0)

Full size image
Fig. 13
figure 13

Sensitivity analysis for sub-criteria (user friendliness and security) (λ = 0.5)

Full size image
Fig. 14
figure 14

Sensitivity analysis for sub-criteria (user friendliness and security) (λ = 1)

Full size image
Fig. 15
figure 15

Sensitivity analysis for sub-criteria (cost of deployment) (λ = 0)

Full size image
Fig. 16
figure 16

Sensitivity analysis for sub-criteria (cost of deployment) (λ = 0.5)

Full size image
Fig. 17
figure 17

Sensitivity analysis for sub-criteria (cost of deployment) (λ = 1)

Full size image

4 Discussion

This section deals with the results of both our quantitative (data collected from 12 SMEs owners who intended to implement ERP in their firms) and qualitative research, conducted in order to validate the results. For later, researchers interviewed four SME owners, who had already implemented ERP in their firms in the last three years. Insights shared by them are given along with research findings:

Table 14 suggests that the criterion ‘Cost of Deployment’ obtained maximum weightage (0.355) while ‘Vendor Credibility’ received minimum weightage of 0.125 in the process of selecting an ERP system. Criterion ‘User Friendliness and Security’ and ‘Need Fulfilment’ with weightage of 0.265 and 0.255 respectively, were ranked as second and third. Our qualitative research also revealed almost similar results; hereby, the cost of the software or an ERP system drives decision-making when it comes to the selection and purchase among the Indian SMEs. This is because most SMEs are integral part of value-chain to bigger companies, wherein capital investment expenses are critical to operate in lean fashion to survive and grow. Moreover, ‘Maintenance and Upgradation Cost’ is ranked first in the sub-criteria, which indicates that owners/decision-makers of SMEs prefer ERP systems, which can help them manage their limited operational requirements with minimum running cost. During our qualitative research, a promoter of a small automobile ancillary manufacturer based at Ahmedabad, Gujarat shared his concerns for cost in the following way: “We are a small firm with an annual turnover of around INR 50 Million. For us, cost is the most important parameter for choosing any software including an ERP system. We kept the ceiling of INR 30,000 for complete package installation. Even then, every month, we need to spend around INR 10,000, which is not a small amount for us”.

The ‘User Friendliness and Security’ criterion was ranked second in preference. Herein, the sub-criteria of Security (0.607) assumed more importance than another sub-criteria User friendliness (0.393); this may be because the SMEs neither have a dedicated IT department/IT employee nor do they have formally qualified and trained employees in different nuances of IT including security. Due to the lack of in-house competencies, they want to ensure the security of the chosen package, in turn, ensuring the security of network as well as the data. An SME owner with a Foundry based at Rajkot endorsed similar views about this criterion: “No doubt, cost is a predominant factor for all sorts of procurements for us; but I consider user-friendliness of ERP software almost at par with the cost. As our employees are not highly educated and well-versed with technology, software should be easy to understand and manage. Unless they adopt it and learn it fast, software is of no use to us. I always prefer a robust software with limited features but higher acceptance among employees than another software having too many features but which is too complex for employees to understand. Our philosophy is simple, once employees become comfortable with a system, we can always upgrade it for higher value. As we do not understand technicality like security, we rely on a software that is commonly used.”

Criterion of ‘Need Fulfilment’ (Current and Future) also influences SMEs’ decision-making process to a great extent. In this, the sub-criterion Solution to Current Business Problems got higher weightage (0.744), than another sub-criterion Facilitation for Future Strategy (0.256). This reveals that the SMEs prefer to invest in IT systems, largely keeping current problems in mind; they do not want to consider issues emanating after 5–10 years. Two co-owners of a small packaging firm narrated their viewpoint in following manner: “We had procured an ERP system three years ago. Prior to that, we had multiple systems catering to different functional requirements. Only problem faced by us was the lack of integration. Therefore, we took a conscious decision to adopt an ERP. As we did not have much requirements like large companies, we settled for a basic ERP system, which could provide us limited reports that we needed. Our customers, processes and products are largely same from last 22 years, so there is near certainty in terms of data processing requirements. As and when critical need arose ex. GST, ERP vendor provided us add-ons. When we bought an additional software, vendor provided an interface with our existing ERP software at a very marginal cost. However, many of reports which are still not available, we rely on Ms-Excel.”

Vendor credibility plays least important role out of the four factors; this may be because ERP packages chosen are generally relatively low in terms of features and price, and need little time for implementation. SMEs prefer to opt for vendors who is known to them or some of their acquaintances in order to reduce this risk. The opinion of an owner of a steel pipe manufacturer based at Jamnagar with around INR 80 million turnover about the role of vendor credibility in ERP selection process is: “ERP system has become important in last one decade to fulfil our day-to-day functional and regulatory requirements but it remains in the backend only. Unlike large companies, any ERP vendor with decent experience of implementation and reasonable competence of technical manpower suits us as our requirements are limited, need for frequent change does not exist and our data and processes are not very confidential. We believe that, a known vendor of moderate quality is better than an excellent vendor unknown to us. For us, cost including the operational expenses drives our purchase decision.”

Study of recent literature highlights some additional insights about the factors affecting ERP selection decision. Czekster et al. (2019) evaluated and mined nine crucial criteria from the several criteria present in recent literature related to selection of ERPs. These are ERP package reputation and references, acquisition and monthly costs, ERP package’s feature set, level of support and training, deployment experience, efficiency, easiness of use, reliability, and maintainability. It has been posited in many research work that vendor’s reputation, their willingness to provide support to the customers throughout the life of the software and their engagement for customer value creation are the important deciding factors in implementing ERP systems (Seethamraju 2015; Garg and Khurana 2017; Claybaugh et al. 2019). Through his research work, Seethamraju (2015) asserted that external factors such as data security, competition, and performance of the ERP system have no influence on adoption decision. Moreover, change management and enhancing the usage effectiveness are challenging, however, the intent of the ERP vendor to accommodate client’s needs and the regular value creation bolster the clients during the support phase.

Wie (2005) proposed an AHP based framework for selection of most suitable ERP system considering two criteria, appropriateness of ERP system and best ERP vendor. The application of the framework was demonstrated for a Taiwan based electronic company. The AHP analysis positioned sub-criteria ‘functionality’ at the first rank, while ‘total cost’ and ‘user-friendliness’ were ranked at 4th and 5th positions, respectively. These results indicate that in a developed country like Taiwan, industries are looking for ERP systems which are providing higher level of functionality in terms of richness of features, broad scope of modules and parameter settings etc. along with flexibility in upgradation, integration and development. This is not the case in the developing country like India, and specifically in SME environment where cost of the ERP package and its user-friendliness are more crucial. This is because of the fact that Indian SMEs are operating with limited funds and lower level of competency pool.

5 Recommendations

The findings of this research paper have many implications, specifically for ERP vendors. Vendors can use these findings in developing appropriate marketing mix (product, price, place, promotion) strategies. They need to develop/modify their ERP product with a modular approach, a core package with necessary functionality but several add-ons, for the SME segment at large. On one end, this would help in reducing the cost of the package, but on the other, it would ensure incremental revenue through add-ons at regular intervals. Keeping end-user’s comfort and convenience in mind, design and navigation should be kept simple. Simple user manual (in user’s preferred languages in addition to English), online FAQs and tutorials can also add value for the vendors. Cloud based ERP (software as a service) can also be proposed for lowering capital expenditure of SMEs (Lenart 2011). A product can be bundled with service support for different time-periods (e.g., one year, three years etc.) for lowering total cost of ownership (TCO) including capital and operational costs. Moreover, in order to reduce operational costs, support (upgrade, maintenance etc.) may be offered online.

SME owners can also make use of these findings. Two or more SMEs from the same industry can explore possibility of collaboration for buying an ERP together, for economies of scale in initial due-diligence, training and other initiatives. Further, they can consider the total cost of ownership (TCO) or lifetime cost instead of an upfront price of the ERP package. Ultimately, for realising the desired benefits from an ERP system, it may be worthwhile for SMEs to invest in people and process along with technology implementation.

6 Conclusion and scope for further research

SMEs are the growth engine for developing countries like India. Information technology including the ERPs facilitate integration of their internal business processes and enable real-time information for smooth functioning of business. Considering various challenges, selection of an appropriate ERP system becomes very important for SMEs. Previous researchers have identified various factors affecting the ERP selection. However, this research fills the gap by ranking these factors, specifically for the SME sector of India. ‘Cost of deployment’ was identified as the most significant criteria affecting the adoption decision of an ERP product by the SMEs. ‘Vendor credibility’ was found to be the least significant factor. Criterion ‘User friendliness and Security’ and ‘Need fulfilment’ were ranked as second and third. ERP vendors can use these findings in developing appropriate marketing mix strategies. SME owners can also make a leaf out of this study for decision-making.

The first limitation of this work is that the technique used i.e., FAHP does not consider interrelations among the variables (criteria and sub-criteria), which can be checked using some contemporary Operations Research techniques like Decision-making trial and evaluation laboratory (DEMATEL) and analytic network process (ANP). Second, while researchers through literature review tried to list major factors affecting the selection of ERP packages, there may be some important factors, which might not have been covered. Third limitation is that all the respondents (in the second and third stages) involved in ranking the four criteria and ten sub-criteria belonged to the manufacturing sector only.

Future researches may be conducted for understanding the interrelations among the variables by applying ANP and DEMATEL. Second, the scope of research can also be extended by comparing small enterprises with medium scale companies and/or larger companies. Third, researchers may also undertake similar studies for SMEs operating in the service sector. Lastly, future research can also consider alternatives for selecting ERP vendors.