An Integrated PLS-SEM-TOPSIS-Sort Approach for Assessing ERP Solutions Acceptance Across Various Industries

Abstract

In the context of accelerated digitalization, enterprises are increasingly adopting information-driven solutions to support managerial decision-making, with Enterprise Resource Planning (ERP) systems playing a crucial role in organizational development. Despite its importance, ERP adoption varies significantly across industries, particularly between developed and developing economies, where technological and structural differences persist. This paper proposes and validates a classification framework for assessing industry readiness for ERP adoption, based on an integrated PLS-SEM-MCDA methodological approach. PLS-SEM identified statistically significant factors and transformed them into weights to compare ERP user attitudes across eleven industries in Serbia and Slovenia. In addition, the TOPSIS-Sort method classified industries into high, moderate, and low readiness as predefined order classes. Finally, sensitivity analysis and comparative analysis are performed with AHP expert weights and the PROMETHEE-FlowSort method to determine the robustness of the PLS-SEM-TOPSIS-Sort results. The results show that the IT industry is the most consistent in adopting ERP systems. In contrast, other industries exhibit varying levels of readiness, depending on their degree of digital maturity and organizational preparedness. The proposed framework’s methodological flexibility allows it to be adapted to various contexts, making it suitable for future academic research and comparative studies. Additionally, the practical implications of the research are twofold. For ERP suppliers, the findings provide guidance on how to approach market segmentation and strategic positioning tailored to the specific needs of individual industries. For ERP users, their success in ERP adoption can be amplified by using the research insights as a benchmarking model.

1. Introduction

One of the main drivers of digitalization is an increased engagement of businesses in electronic information-sharing platforms [1]. The rapid advancement of internet technologies has led to the widespread adoption of software solutions, such as Enterprise Resource Planning (ERP) systems, which have become an indispensable part of business practice [2]. ERP systems are large information systems platforms that companies use to automate and optimize their business processes, enabling cross-functional transactions [3]. Today, these systems serve as a means of standardization, rationalization, and automation of processes [4].

Worldwide, a growing number of companies are implementing ERP systems. As of 2023, Eurostat reports that 43.3% of companies in the European Union used ERP software solutions. This indicates an increasing trend compared to previous years. Although the application of ERP systems is widespread across various industries [5], the manner, intensity, and specifics of their use vary depending on the sector’s needs, organizational complexity, and the company’s digital maturity [6]. In 2024, a comparison across various industry sectors revealed that more than half of the companies in the information and communication sector, as well as those in manufacturing and the electricity, gas, steam, and air conditioning sectors, utilized ERP systems. Following these sectors, those involved in scientific, professional, and technical activities, as well as retail trade and wholesale, also showed significant ERP adoption [7].

This diversity raises an important question: How to classify industries based on ERP system acceptance patterns, and what methodological approaches provide reliable and applicable criteria for such classification? Classification of industries in the context of ERP system adoption enables a deeper understanding of the distinct needs, barriers, and success factors associated with each sector [8,9].

To validate theoretical models that considered a linear relationship between influence factor and attitude to use ERP systems, many research studies [10,11,12] use Structural Equation Modeling (SEM). However, while SEM provides valuable insights into causal relationships, it does not fully capture the comparative and evaluative dimensions necessary for classifying industries by readiness levels. Recognizing this gap, this study proposes an integrated PLS-SEM-TOPSIS-Sort methodological framework to identify key factors influencing ERP acceptance and classify industries by their readiness for ERP adoption. This approach combines the strengths of SEM in modeling complex relationships with the advantages of multi-criteria decision analysis (MCDA). PLS-SEM is first used to identify and quantitatively validate the importance of latent determinants of ERP implementation. Then the results of this structural analysis are directly integrated into the TOPSIS-Sort multi-criteria classification model. Therefore, upgrading SEM with MCDA methods can provide a more holistic and practically relevant evaluation framework [13].

The study [14] has highlighted a significant technological gap between the developing and developed countries and the European average. In this context, examining ERP adoption in emerging economies provides valuable insights into both the drivers and barriers of digital transformation, contributing to a better understanding of the processes that shape technological advancement [15]. Such research is particularly important for the regional context, where accelerating digital integration remains essential for achieving sustainable business transformation and reducing the existing technological lag. To test the proposed approach, empirical research was conducted across eleven industries in Serbia and Slovenia—two countries that represent transitional, but differently developed economic contexts [1].

The paper is divided into several sections. After the Introduction, the Literature Review explains the previous studies on ERP systems, followed by a section on Hypothesis Development. The Research Methodology section describes the methodological approach and research method. The subsequent sections cover the Research Results and Discussion. The final part of the paper, the Conclusion, summarizes the study’s novelty, theoretical and practical contributions, outlines limitations, and suggests directions for future research.

2. Literature Review

Examining ERP system acceptance is an important objective of many research studies. In the context of the ERP system research topics, ERP has been considered in terms of selection criteria [16], ERP utilization [17,18], and its implementation process [19].

Within the mentioned areas of ERP system examination, numerous studies have been conducted in various industries. For example, in the manufacturing industry, the primary task is to transform physical raw materials into tangible products [20]. It is important to track material flows and production costs. By implementing an ERP system, manufacturing companies can enhance the efficiency of their operations and ensure the timely delivery of products, while also reducing costs and streamlining document flow [21]. Therefore, managing production automation is one of the main advantages of ERP adoption.

To integrate all activities in the food production industry, companies in this sector must coordinate manufacturing, inventory, and assembly functions [22]. Therefore, such companies need to implement an ERP system solution, which will, in turn, minimize costs, maximize efficiency, help track the production process, and ensure product quality [23,24].

The automobile industry is characterized by numerous processes that involve the ongoing flow of materials, semi-finished goods, or finished goods. Therefore, the ERP system basically makes it easier to follow all transactions. The study conducted by [25] identifies the key factors influencing ERP system acceptance in the automotive sector as organizational, technological, environmental, and individual. Additionally, a study by [26] proved that the efficiency of operations increases significantly more after implementing the ERP system than before ERP implementation.

Lagging behind the digitalization process and low-intensity acceptance of new software solutions are key attributes in the wood industry. Ref. [27] stated that in overcoming these challenges, ERP systems are essential for improving the efficiency and effectiveness of production processes, enhancing communication, and enhancing supply chain integration in the wood processing industry. Furthermore, even the construction industry, due to its project-oriented nature, is embracing digitalization trends. This is one of the industries with the highest failure rates during ERP implementation, primarily due to functional issues [28].

When it comes to the information technology (IT) industry, a study by [29] emphasized the importance of a modular approach for ERP utilization in this sector. During the software development phase, the modules dedicated to project management and planning activities are critical for ensuring that the project stays on track, within budget, and meets the specified requirements. During the software distribution phase, the emphasis shifts to customer relationship management modules. These modules help in tracking customer interactions and preferences, while sales management tools focus on optimizing the sales pipeline, forecasting revenue, and analyzing market trends.

Recent studies have focused on explaining the factors that affect ERP system acceptance. For this purpose, most studies employ the PLS approach. A study by [30] employed the PLS technique to investigate the impact of external factors on the perceived usefulness and perceived ease of use of the ERP system. Reference [31] found a positive and statistically significant influence of external factors on ERP implementation. Research by [32] utilized the PLS technique to examine the impact of three factors on ERP system adoption in a sales and distribution company. A study by [33] examined the influence of external variables, including user training, user adaptability, non-technological complexity, and trainer support, on the ERP acceptance research model.

Furthermore, in light of current studies in the field of ERP systems, the application of MCDA methods is represented. In this regard, the paper by [34] utilized the VIsekriterijumsko KOmpromisno Rangiranje (VIKOR) method for selecting an adequate ERP software, employing four groups of criteria. For the evaluation of ERP software, ref. [35] employed a hybrid fuzzy approach that combined the application of the Analytic Hierarchy Process (AHP) and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method. Also, in their study, ref. [36] used the AHP-TOPSIS method. Ref. [37] used the fuzzy Combinative Distance-based Assessment (CODAS) method for ERP system selection. Ref. [38] used the intuitive fuzzy and interval gray number-based Measuring Attractiveness by a Categorical Based Evaluation Technique (MACBETH) method for the selection of ERP systems. Fuzzy AHP is also used for ERP system selection in the study by [39]. The systematized literature review is presented in

Table 1. Systematized literature review.

Year	Author(s)	Aim of the Study	Methodology
2013	Sternad & Bobek [30]	Examination of external factors on ERP system acceptance	SEM
2018	Bhattacharya et al. [31]	Examination of determinants of the intention to adopt ERP system	SEM
2020	Putri et al. [32]	Analysis of critical success factors for ERP acceptance	SEM
2021	Limantara et al. [33]	Analysis of factors for ERP usage	SEM
2021	Ayağ & Samanlioglu [35]	ERP software packages selection	MCDA (hybrid fuzzy AHP–TOPSIS)
2022	Uddin et al. [36]	ERP system selection	MCDA (hybrid AHP–TOPSIS)
2021	Aydoğmuş et al. [37]	ERP selection	MCDA (fuzzy CODAS)
2022	Jin [34]	Extension of VIKOR method for ERP system selection	MCDA (VIKOR)
2022	Yurtyapan & Aydemir [38]	ERP system selection	MCDA (fuzzy and interval gray number-based MACBETH)
2024	Dağci Yüksel & Ersöz [39]	ERP software selection	MCDA (fuzzy AHP)

.

A review of the relevant literature reveals that the degree of acceptance of ERP systems varies significantly across industries. ERP system implementation involves multiple dimensions that contribute to its overall success. According to Table 1, to assess the factors affecting ERP system adoption [30,31,32,40], previous research has predominantly used the SEM technique, while methods of multi-criteria decision analysis have mainly been applied to select appropriate ERP solutions from the perspective of individual companies [38]. A review of the literature revealed significant complexity in determining the impact of individual factors and analyzing the readiness of industries to accept the ERP system. Consequently, the nature of the problem necessitates the development of a holistic approach that integrates and considers both problems simultaneously. Responding to this research challenge, this research introduces an integrated PLS-SEM-TOPSIS-Sort approach, designed to identify and evaluate key factors influencing attitudes towards the use of ERP systems, as well as to classify industries according to their readiness to adopt ERP solutions. The proposed research model was verified in a case of companies that use ERP systems and operate in Serbia and Slovenia. By combining the advantages of the SEM method in validating complex cause-and-effect relationships with the ability of the MCDA approach to systematically rank and classify alternatives, this integrated framework enables a comprehensive and multidimensional insight into the acceptance of ERP systems, providing a deeper understanding of the reasons why certain industries show a higher or lower degree of readiness. To further ensure the robustness and validity of the proposed methodological framework, the results obtained through the PLS-SEM-TOPSIS-Sort approach will be cross-verified using the PROMETHEE-FlowSort method. This additional verification not only strengthens the credibility of the findings but also highlights the potential of combining multiple decision-support techniques in ERP system research. Therefore, the proposed integrated framework effectively addresses the gap previously considered in the literature.

3. Hypothesis Development

To identify the most important factors that impact the attitude of ERP systems, the concept of the ERP systems-based model is analyzed, and several hypotheses are proposed.

3.1. Work Compatibility (WC)

Work compatibility refers to the degree to which the technology is aligned with the user’s needs. It also refers to the ability to align technology with the way tasks are performed in a company [41]. From the perspective of organizational culture, it can be defined as the degree to which innovation is perceived as consistent with existing values, past experiences, and employee needs [42]. If users of the ERP system see the possibility of harmonizing this system with the company’s needs and business philosophy, their attitude towards its use will be positive [43]. Compatibility is one of the key factors that influence the attitude towards use [44]. Therefore, the following hypothesis is proposed:

H1. 

Work Compatibility of the ERP system positively impacts Attitude to Use.

3.2. Perceived Usefulness (PU)

Numerous studies have consistently demonstrated a statistically significant positive correlation between perceived usefulness and attitude toward use [17,32,45]. The perception of system usefulness is a highly influential determinant of end-user satisfaction, leading to a greater willingness to use the ERP system [46]. Accordingly, the following hypothesis is proposed:

H2. 

Perceived usefulness of the ERP system positively impacts Attitudes to Use.

3.3. Perceived Ease of Use (PEoU)

Perceived ease of use (PEoU) is the degree to which a person believes that using a particular system would be effortless [47]. PEoU refers to employees’ belief that using an ERP system will make it easier for them to perform a task [32], and thus, the system is considered very useful. If the user of the ERP system perceives this system as easy to use, then the perceived usefulness of the system will be higher [2]. The ease of performing transactions in the business system can increase the level of user comfort at work [48]. Additionally, previous research has confirmed that PEoU has a positive and statistically significant impact on attitude towards using ERP systems [17,18]. If the ERP system is considered easy to use and masters the functions it offers, it increases the degree of acceptance of such an ERP solution [49]. The simpler the ERP systems, the higher the level of acceptance will be [50]. Accordingly, the following hypothesis is proposed:

H3. 

Perceived ease of use of the ERP system positively impacts attitude to use.

3.4. External Factors (EF)

This paper examines the impact of several external variables, including system complexity, user manuals, system performance, social influence, and alignment with business processes. The concept of considering EF as a second-order construct with five lower-order constructs stems from the use of the MCDA hierarchical decision model.

System Complexity (SC): An ERP system is often viewed as a complex IS to understand and further use to perform business tasks in an organization. SC can be defined as the level of complexity of technology from the perspective of employees or as the degree of difficulty and volume of transactions processed in the IS [51]. In the literature, it is noted that SC can negatively impact employees’ attitudes towards the use of ERP systems in companies [52]. Poor understanding and navigation in the ERP system can negatively affect the working atmosphere and productivity of employees [53].

System Performance (SP): If the ERP system has enviable performance, i.e., when it responds quickly to business actions and ensures efficient process execution, users will adopt the same system more quickly [54]. Thanks to the performance of the ERP system, which enables fast, real-time transactions, previous research has confirmed an improvement in the company’s business indicators following its implementation [55]. In a research study by [56], a significant relationship between the successful use of ERP systems and performance was proven.

User Manuals (UM): A critical phase following ERP implementation involves ongoing support from the ERP provider, which is essential for receiving guidance during the integration of new versions or when updating existing features and modules [57]. If there is good support in the form of user instructions, it is expected that the degree of ERP system acceptance will be higher.

Social Influence (SI): The concept of SI refers to the impact of social pressure or the opinions of influential individuals on an individual’s acceptance of a technology [58]. Managers’ attitudes toward ERP system implementation are the critical factor in IS acceptance. Manager’s support contributes to a higher degree of ERP acceptance [55]. The implementation of new technologies into the company’s operations is driven by both stakeholder requirements and competitive pressure [59].

Business Process Fit (BPF): Business processes are critically important as they represent how organizations and individuals communicate [60]. The successful acceptance of an ERP system depends on its technological compatibility with the organization’s existing processes. BPF is a critical factor in ERP system selection and implementation, as the opposite can lead to additional costs [61]. The more compatible the ERP system is with the organization’s processes, the greater the readiness.

Considering EF as a higher-order construct that consists of lower-order constructs (SC, SP, UM, SI, BPF), the following hypothesis is proposed:

H4. 

External factors positively impact the Attitude to use ERP system.

Figure 1 presents the research model. WC, PU, and PEoU represent first-order constructs, while the variable EF was derived based on a reflective second-order construct using a repeated indicators approach to capture its higher-order structure. The repeated indicator approach’s advantage lies in estimating all constructs at once, rather than separately estimating lower-order and higher-order dimensions [62].

4. Research Methodology

The data collection process and applied methods are described below. The PLS-SEM method was initially used to test the research model’s hypotheses. In the next part, the findings from the PLS-SEM analysis were further examined and organized using the TOPSIS-Sort method. This step enabled a systematic classification of systems, providing a more comprehensive understanding of the differences between industries in ERP adoption readiness.

4.1. Data Collection

A survey method was used to obtain a sufficient sample. The data were collected in 2024. The questions in the survey were adopted from the study by [63] (Appendix A). All variables were measured using a five-point Likert scale from 1 (completely false) to 5 (completely true). The questionnaire link was shared via the LinkedIn network with individuals working in companies that operate in the examined region. Respondents were previously carefully selected based on whether they are engaged in positions that involve performing work tasks in ERP and are employed in various industrial sectors. A total of 350 respondents gave complete and usable answers.

4.2. PLS-SEM Approach

Structural Equation Modeling (SEM) is an approach for modeling a set of relationships involving one or more independent variables and one or more dependent variables. SEM ensures that a pattern of relationships between several different endogenous and exogenous variables is examined [64]. PLS-SEM, as a components-based approach, offers several advantages and is suitable for research focused on refining existing theories, particularly when dealing with sophisticated models that integrate various constructs, variables, and connections. PLS-SEM is also beneficial when working with smaller sample sizes and utilizing primary or secondary data [65]. Additionally, based on the literature [66,67,68,69], it was observed that most previous research has been based on PLS-SEM.

4.3. TOPSIS-Sort Approach

Sorting problems assign a set of alternatives A = {a₁, a₂, …, a_m} to k ordered classes C₁, C₂, …, C_K, considering conflict criteria C = {c₁, …, c_n} [70]. Several MCDM methods, including the TOPSIS-Sort methodology, can be used for sorting the alternatives into different classes. Figure 2 depicts steps that should be performed to apply the TOPSIS-Sort algorithm (adapted from [70]). Furthermore, the implementation of the TOPSIS-Sort algorithm in Python 3.13, using the topsis_sort_step4_all_benefit function, is presented below. The complete code with initial values can be found in the Code Appendix or, for more details, at: https://github.com/sarsic-eng/TOPSIS-SORT.git.

# --- TOPSIS-SORT function (returns results and thresholds as Series)
def topsis_sort_step4_all_benefit(X, weights, criteria_types_norm, boundary_profiles):
    X = pd.DataFrame(X).copy()
    m, n = X.shape
    w = np.asarray(weights, dtype = float)
    w = w/w.sum()
    ctype = [str(ct).lower() for ct in criteria_types_norm]
    B = pd.DataFrame(boundary_profiles).copy()
    B.columns = X.columns
    M = pd.concat([X, B], axis = 0)
    R = pd.DataFrame(index = M.index, columns = M.columns, dtype = float)
 
    for j, ct in  enumerate(ctype):
        col = M.iloc[:, j].astype(float).values
        if ct == “benefit”:
            denom = float(col.max())
            R.iloc[:, j] = col/denom if denom != 0 else 0.0
        else:
            numer = float(col.min())
            col = np.where(col == 0.0, np.finfo(float).eps, col)
            R.iloc[:, j] = numer/col
 
    V = R * w
    v_plus = V.max().values
    v_minus = V.min().values
 
    D_plus_all = np.linalg.norm(V.values - v_plus, axis = 1)
    D_minus_all = np.linalg.norm(V.values - v_minus, axis = 1)
    Ci_all = D_minus_all/(D_plus_all + D_minus_all)
 
    Ci_alt = pd.Series(Ci_all[:m], index = X.index, name = “Ci”)
    Ci_prof = pd.Series(Ci_all[m:], index = B.index, name = “Ci”)
 
    thr_sorted = Ci_prof.sort_values(ascending = False) # Series
    thr_values = thr_sorted.values
 
    def _assign(ci):
        for idx, thr in enumerate(thr_values, start = 1):
            if ci >= thr:
                return idx
            return len(thr_values) + 1
 
    classes = Ci_alt.apply(_assign)
    results = pd.DataFrame({“Ci”: Ci_alt, “Class”: classes})
    return results, thr_sorted

5. Results and Discussion

Of the 350 respondents in this study, 62.9% were male and 37.1% were female. The majority of respondents, representing 77.3%, fall within the 30 to 49-year-old age range. In terms of educational level, over 80% of respondents hold a university degree, 40.9% have an undergraduate degree, 43.9% hold a master’s degree, and 2.3% hold a doctorate. Most respondents work in companies with more than 250 employees (64.4%), while 58.3% of respondents have been using the ERP system for over five years.

5.1. PLS-SEM Method

To establish convergent validity, the evaluation of the measurement model should include factor loading of the indicators, composite reliability (CR), and the average variance extracted (AVE). Initially, a Confirmatory Factor Analysis (CFA) was conducted to validate the structure of all latent constructs. The CFA results indicated that only System Complexity (SC) did not demonstrate statistical significance in the high-order construct of External Factor (EF). Therefore, SC was excluded from subsequent analyses within the EF group for both the Serbian and Slovenian datasets. After this modification, the proposed research model was recalculated, and the results are presented in

Table 2. Construct reliability and validity.

	Cronbach’s Alpha (Cα) *	Composite Reliability (rhoa) **	Composite Reliability (rhoc) ***	Average Variance Extracted (AVE) ****
Attitude to Use	0.807	0.830	0.911	0.837
Business Process Fit	0.750	0.821	0.845	0.592
Perceived Ease of Use	0.881	0.890	0.944	0.894
Perceived Usefulness	0.874	0.874	0.941	0.888
Work Compatibility	0.879	0.880	0.926	0.806
Social Influence	0.761	0.765	0.863	0.677
System Performances	0.858	0.862	0.899	0.640
User Manuals	0.902	0.906	0.939	0.837

Notes: * Cα ≥ 0.70; ** rho_a ≥ 0.70; *** rho_c ≥ 0.70; **** AVE ≥ 0.50.

. In the measurement model, all factor loadings meet the threshold of 0.70, indicating that all items were retained. Table 2 shows the results of the tests used to check the measurement model for construct reliability and validity. Considering that all values are above the recommended thresholds (see Notes in Table 2), and taking into account all first-order latent variable constructs, the model can be considered reliable and valid [71].

Discriminant validity was evaluated using the Fornell-Larcker criterion. The results of this test show that the square root of AVE for every variable is greater than the relationship among constructs. In addition, the value of the heterotrait-monotrait (HTMT) matrix was assessed.

Table 3. HTMT matrix.

	1	2	3	4	5	6	7	8
Attitude to Use (1)	-
Business Process Fit (2)	0.728
Perceived Ease of Use (3)	0.629	0.595
Perceived Usefulness (4)	0.799	0.795	0.653
Work Compatibility (5)	0.857	0.812	0.724	0.868
Social Influence (6)	0.623	0.725	0.459	0.581	0.654
System Performances (7)	0.644	0.738	0.637	0.625	0.644	0.486
User Manuals (8)	0.329	0.484	0.510	0.404	0.420	0.328	0.529	-

shows that all values meet the recommended threshold. Therefore, after overall evaluation, discriminant validity can be accepted for this model [72].

The SRMR model fit index (0.101) value is at the limit value. Additionally, the d_ULS value from the original sample (12.272) is less than the d_ULS 95% quantile derived from bootstrapping (19.338). Therefore, the model can be considered acceptable according to the global fit measures [73].

In the next phase, the structural model was examined to obtain the results of the path coefficient hypothesis testing.

Table 4. Results of hypothesis testing.

Hypothesis	Construct	Estimated Path Coefficient (β)	p-Value (p) *	Remark
H1	Work Compatibility → Attitude to Use	0.404	0.000 *	accepted
H2	Perceived Usefulness → Attitude to Use	0.227	0.001 **	accepted
H3	Perceived Ease of Use → Attitude to Use	0.048	0.367 n.s.	rejected
H4	External Factors → Attitude to Use	0.169	0.006 **	accepted

Note: * significant at level of 0.000; ** significant at level of 0.05; ^n.s. non-significant.

shows the structural model with the results of the hypothesis tests.

Regarding the effect of WC on AT (H₁), the statistical significance of the relationship shows that WC has a positive influence on AT. PU also has a positive, statistically significant effect on AT. The same results were also confirmed by previous research [17,32,45], suggesting that the ERP system is perceived as a useful system that contributes to business and work performance. Contrary to expectations, PEoU has no statistically significant influence on AT, thereby rejecting hypothesis H3, which contrasts with the results of previous studies [17,18,49]. When it comes to the influence of EF on AT (H4), this relationship is positive and statistically significant. This suggests that external factors, as variables that may influence AT, are a complex combination of variables [74].

Determining the weighted importance of factors took into account the partition of data into ERP users from Serbia and ERP users from Slovenia. The weighting of the significant criteria is determined by the relative frequency of the estimated path coefficients, while the weighting of the sub-criteria within each criterion is determined by the relative frequency of the factor loadings [75,76]. The general weight of the sub-criteria results from the multiplication of the weight of the criteria and the weight of the sub-criteria in the reference group of criteria. The relative weights of the criteria and sub-criteria are presented in

Table 5. Relative weights of significant criteria and sub-criteria in the model.

Criteria	Subcriteria	Serbia			Slovenia
		Weight of Significant Criteria	Weight of Subcriteria in the Criteria	Global Weight of Subcriteria	Weight of Significant Criteria	Weight of Subcriteria in the Criteria	Global Weight of Subcriteria
Work Compatibility	WC_I	0.332	0.333	0.110	0.695	0.321	0.223
	WC_II		0.343	0.114		0.346	0.240
	WC_III		0.324	0.108		0.333	0.232
Perceived Usefulness	PU_I	0.269	0.497	0.134	0.254	0.503	0.128
	PU_II		0.503	0.135		0.497	0.126
External Factors	EF–SP	0.399	0.264	0.105	0.051	0.274	0.014
	EF–UM		0.232	0.093		0.218	0.011
	EF–SI		0.223	0.089		0.235	0.012
	EF–BPF		0.281	0.112		0.273	0.014

.

Based on the results obtained regarding ERP users from Serbia, EF is the most important criterion for the acceptance of the ERP system (39.9%), and among the considered sub-criteria of EF, the weight of sub-criteria decreases in the order of BPF (0.0281), SP (0.264), UM (0.232), and SI (0.223). This means that BPF has the highest and SI the lowest importance among the EF. These findings demonstrate that organizations in Serbia recognize the importance of ensuring the ERP system aligns with current business practices and industry standards, which is crucial for its successful implementation and effective utilization. The lower importance of SI can be attributed to the fact that decisions regarding ERP system implementation depend to a greater extent on internal needs and organizational priorities than on external pressures. Furthermore, the importance of the WC criterion in the model is 33.2% and the weight of PU is 26.0%. This indicates that users in Serbia often rate the ERP system highly for its alignment with their work tasks and processes, as well as its usefulness and practical application in daily work.

For ERP users from Slovenia, WC is the most important criterion, with a weight of 69.50%. The following most important criterion in the model is PU (25.4%). Unlike Serbia, ERP users in Slovenia consider EF as the least important criterion (5.1%). The weight of sub-criteria decreases in the order of SP (0.274), BPF (0.273), SI (0.235), and UM (0.218). This indicates that ERP users from Slovenia prioritize the technical performance of the system, while assigning the least importance to UM, suggesting a higher level of technical training and user maturity.

The differences in results between Serbia and Slovenia can be attributed to the varying levels of digital maturity and institutional stability between the two countries. In Serbia, the process of digital transformation is still intensively developing, so organizations are increasingly dependent on external factors, such as regulatory requirements, market pressures, and partner recommendations. On the other hand, in Slovenia, where ERP systems are already widely integrated into business processes, the emphasis is shifted to the internal efficiency, compatibility, and usefulness of the system. Such results align with modern research, which suggests that as organizations progress through the digital transformation process, their focus shifts from external influences to internal optimization and the technical performance of the ERP system [77,78]. Furthermore, while Serbian organizations continue to adapt to market and regulatory demands, Slovenian organizations are already focusing on enhancing the functionality and productivity of their ERP systems internally.

5.2. TOPSIS-Sort Method

After validating the model in the first research phase, the second research phase was undertaken, which involved applying the MCDA technique TOPSIS-Sort to classify industries according to their predisposition to use the ERP system. In the decision model, the alternatives represent eleven types of industries, while the criteria for selecting the industries are determined based on the accepted hypotheses of the SEM model (WC, PU, and EF).

In the next phase of the analysis, the average scores of Likert-scale items across industries were used as criterion scores in the TOPSIS matrix. Although Likert scale responses are ordinal in construction, their aggregated means across industries and items can be treated as continuous interval variables. This assumption is consistent with the central limit theorem and the psychometric view that averaging a large number of items approximates the interval properties of a measurement [79]. Therefore, the use of mean values of Likert items by industry as criterion scores in the TOPSIS matrix is considered methodologically sound and aligned with previous research practices.

Table 6. Initial decision matrix.

Criteria	Work Compatibility			Perceived Usefulness		Perceived Ease of Use		External Factors
Sub-Criteria	WC_I	WC_II	WC_III	PU_I	PU_II	PEoU_I	PEoU_II	SP	SC	UM	SI	BPF
Criteria Type (max/min) Alternative	max	max	max	max	max	max	max	max	min	max	max	max
A1	4.42	4.45	4.45	4.42	4.29	4.03	3.71	4.18	2.10	3.87	4.76	4.42
A2	3.52	3.62	3.77	3.74	3.72	3.44	3.41	3.75	2.74	3.43	3.89	3.72
A3	4.08	4.21	4.13	3.96	4.08	3.96	3.71	4.04	2.28	3.67	4.60	4.29
A4	4.20	4.40	3.60	4.20	3.60	4.20	3.60	4.28	1.93	3.07	4.47	4.50
A5	3.80	4.05	4.05	4.25	4.25	4.15	4.05	4.08	1.87	3.63	4.43	4.21
A6	4.17	4.13	4.21	4.12	4.25	3.96	3.75	4.10	2.18	3.49	4.44	4.18
A7	4.25	4.25	4.75	4.00	4.00	4.00	3.50	3.95	2.50	3.33	3.92	4.06
A8	4.00	4.11	4.22	4.30	4.30	3.89	3.78	4.06	2.44	3.62	4.47	4.17
A9	3.80	4.20	4.20	4.20	4.00	3.60	3.60	3.96	2.60	3.60	4.40	4.05
A10	3.50	3.88	4.25	3.75	3.75	3.88	3.63	4.43	2.50	3.33	4.58	3.84
A11	4.14	4.36	4.23	4.64	4.45	3.91	3.86	4.06	2.12	3.11	4.52	4.44
Profile boundary B1 (c = 15%)	4.28	4.33	4.58	4.51	4.32	4.09	3.95	4.32	2.00	3.75	4.63	4.38
Profile boundary B2 (c = 35%)	4.10	4.16	4.35	4.33	4.15	3.93	3.83	4.19	2.17	3.59	4.46	4.23
Wj–S1 (ERP users from Serbia)	0.110	0.114	0.108	0.134	0.135	/	/	0.105	/	0.093	0.089	0.112
Wj–S2 (ERP users from Slovenia)	0.223	0.240	0.232	0.128	0.126	/	/	0.014	/	0.011	0.012	0.014
Wj–S3 (ERP experts from Serbia and Slovenia)	0.062	0.206	0.113	0.041	0.081	0.028	0.028	0.122	0.014	0.049	0.032	0.223

Note: A1—IT industry; A2—Automotive industry; A3—Industrial selling; A4—Banking; A5—Pharmaceutical industry; A6—Production; A7—Aviation industry; A8—Mining and metallurgy; A9—Energy industry; A10—Telecommunications industry; A11—Food and beverage industry.

presents the initial decision matrix, which serves as the starting point for applying the TOPSIS-Sort method. The values of the alternatives in relation to the criteria represent the average values of the respondents’ answers from the individual sectors, which were measured on a five-point Likert scale and were the result of an empirical study. To effectively address and mitigate the uncertainty of the initial dataset, the average values were computed based on the representative research sample.

By using this approach, the derived average values are more reliable and better represent the broader ERP user population. Furthermore, this approach significantly decreases subjective biases in assessing the importance of various criteria by incorporating a diverse range of users’ opinions [80]. The initial matrix also displays the values of the boundary profiles Bk, which were determined by considering c = 15% for boundary profile B1 and c = 35% for boundary profile B2. These boundary profiles were taken into consideration because the first change, i.e., the transition of the alternative to Class 1, happened at these boundary profile values for all scenarios. In the considered matrix, all criteria are beneficial, except SC, which is a cost criterion.

Two scenarios (S1 and S2) were defined by taking global weights of the sub-criteria determined by the PLS-SEM approach (presented in Table 5). The first scenario, S1, presents the weights determined by ERP users from Serbia. The second scenario, S2, proposes the weights established by ERP users from Slovenia. To compare the opinions of the surveyed users of the ERP system from scenarios 1 and 2, scenario S3 was additionally defined (

Table 7. TOPSIS-Sort results.

Alternatives	S1 ERP Users from Serbia			S2 ERP Users from Slovenia			S3 ERP Experts from Serbia and Slovenia
	Class 1	Class 2	Class 3	Class 1	Class 2	Class 3	Class 1	Class 2	Class 3
A1		✔			✔			✔
A2			✔			✔			✔
A3			✔			✔			✔
A4			✔			✔			✔
A5			✔			✔			✔
A6			✔			✔			✔
A7			✔		✔				✔
A8			✔			✔			✔
A9			✔			✔			✔
A10			✔			✔			✔
A11		✔			✔			✔

Note: The ✔ symbol indicates that a particular industry (A1, A2, …, A10) has been classified into the corresponding class (Class 1–3).

). This scenario included the opinions of experts from consulting organizations involved in ERP implementation that operate in Serbia and Slovenia. Experts in the field of ERP systems are highly qualified individuals with many years of work experience in various industries who possess both theoretical and practical knowledge about the implementation, integration, and evaluation of business information systems. To collect the opinions of the expert group, interviews were conducted using the traditional AHP group method [81]. Table 7 presents the results of the applied method, which categorizes industries into three distinct classes. Class 1 includes industries with a high predisposition for ERP system adoption. Class 2 comprises industries that require improvements in specific aspects of ERP implementation to enhance their overall acceptance of the system. Class 3 encompasses industry sectors with the lowest levels of ERP acceptance, indicating significant barriers or limitations to adoption.

Based on the results regarding ERP users from Serbia (S1), there are no industries in Class 1, indicating that no sector is fully ready for ERP acceptance without further improvements. The IT and food and beverage industries belong to Class 2, indicating that ERP systems are well-implemented. Still, the success of their application depends on additional work on process integration, improving employee training, and adapting software to specific needs. Class 3 encompasses the largest number of industries (automotive, industrial selling, banking, production, pharmaceutical industry, aviation, metallurgy and mining, energy industry, and telecommunications) that face some challenges in ERP implementation, probably due to process complexity, costs, and sector constraints.

According to ERP users from Slovenia (S2), no industry has a very high predisposition for ERP use; therefore, there are no industries in Class 1. The IT industry, aviation industry, and food and beverage industry are classified in Class 2. In contrast, the industrial sectors, including the automotive industry, banking, production, pharmaceutical industry, metallurgy and mining, energy industry, and telecommunication industry, are classified in Class 3, as they exhibit low readiness for ERP system acceptance.

Findings from S3, representing the opinions of experts and consultants in the field of ERP systems, align with the views of respondents from Serbia regarding industry classification. Additionally, as in the previous two scenarios, no industry is classified in Class 1.

Regarding the comparative approach for the three scenarios, there are no industries in Class 1. Serbia and experts have both the IT and food industries classified in Class 2, while Slovenia also includes the aviation industry in this class. This indicates that the aviation sector in Slovenia demonstrates a higher level of ERP readiness compared to Serbia.

Although ERP systems offer numerous benefits in integrating, automating, and optimizing business processes, specific industries may face challenges that make them less inclined to adopt them. In light of the stated, for both considered countries, most industries face difficulties with ERP acceptance. For example, the automotive industry is characterized by complex supply chains and sophisticated resource management [79]. The implementation of an ERP system enables companies in the automotive sector to centrally manage various business aspects, including production, procurement, logistics, sales, and service. This is particularly important in the context of the automotive industry, given the large number of components and parts used in production and the global nature of supply chains [82,83]. Additionally, in the telecommunications industry, specific network infrastructure requirements often require specialized software. These include specific requirements, high customization costs, security and privacy concerns, as well as the need for agility and flexibility in business [84].

The production industry is characterized by complex manufacturing processes that require detailed resource planning, constant inventory monitoring, and efficient management of production operations [85]. These processes are important for maintaining high productivity and competitiveness in the industry. Implementing an ERP system allows manufacturing companies to centrally manage all aspects of production, including raw material procurement, inventory management, production planning, quality control, logistics, and product distribution. Integrating all these functions into a single system enables better coordination between different departments and processes, resulting in more efficient management of production operations and a reduction in the time required to produce and deliver products. The same classification applies to industrial sales, as well as mining and metallurgy. In the case of industrial selling, ERP systems provide real-time insight into inventory status, which is critical for timely procurement planning and inventory management. On the other hand, in the mining industry, the automation and standardization of processes through an ERP system help to reduce costs, improve product quality, and increase labor efficiency.

Another example is the pharmaceutical industry, which is characterized by its high regulatory requirements and strict quality standards. These requirements are often a challenge for companies in this industry, which must ensure that their products meet the highest standards of safety and efficacy before they reach the market [86].

To verify the reliability of the MCDA model, a sensitivity analysis was conducted in relation to the variation in the value of the coefficients c. Figure 3 presents the results of the sensitivity analysis.

The sensitivity analysis results (Figure 3) show a clear trend, with most industries transitioning from lower to higher readiness classes as the thresholds gradually increase (from K1 to K7). In scenario S1 (ERP users from Serbia), most industries begin in Class 3 at lower thresholds (K1–K2), but with more rigid limits (K4–K7), they progress to higher classes. The most pronounced progress is visible in the IT, pharmaceutical, production, and food and beverage industries, which, at higher thresholds (K4–K7), consistently reach Class 1, indicating their relatively greater readiness for ERP implementation. In scenario S2 (ERP users from Slovenia), the classification is similar, but with a slightly more stable classification, as most industries reach Class 1 already at K4. At the same time, changes between K5 and K7 hardly occur, which suggests less sensitivity to changing borders. In scenario S3 (ERP experts), the results show the most consistent assessment. In most cases, ERP experts place industries in higher classes even at lower thresholds, which may indicate a higher perception of readiness compared to users’ self-assessments. This shows that the classification remains stable from K4 onward, meaning that the results do not change significantly after B1 and B2 reach 40% and 60%, respectively. This indicates the reliability of the identified classes through the applied TOPSIS-Sort approach.

In addition, the robustness of the model was assessed using the PROMETHEE-FlowSort technique. The PROMETHEE V-shape preference function was used, with indifference threshold (Q) set to 0 and preference threshold (P) set to 1. This ensures a greater preference for dominant alternatives. The PROMETHEE-FlowSort results are presented in

Table 8. Robust analysis for Profile boundaries B₁ (c = 15%) and B₂ (c = 35%).

Alternatives	PROMETHEE-Flowsort (ERP Users from Serbia)			PROMETHEE-Flowsort (ERP Users from Slovenia)			PROMETHEE-Flowsort (ERP Experts from Serbia and Slovenia)
	Class 1	Class 2	Class 3	Class 1	Class 2	Class 3	Class 1	Class 2	Class 3
A1	✔			✔			✔
A2			✔			✔			✔
A3			✔			✔			✔
A4			✔			✔			✔
A5			✔			✔			✔
A6			✔			✔			✔
A7			✔		✔				✔
A8			✔			✔			✔
A9			✔			✔			✔
A10			✔			✔			✔
A11		✔			✔			✔

Note: The ✔ symbol indicates that a particular industry (A1, A2, …, A10) has been classified into the corresponding class (Class 1–3).

.

The robust analysis showed a degree of agreement between the classifications obtained using TOPSIS-Sort and PROMETHEE-FlowSort methods. The IT industry consistently stood out as dominant in all three scenarios, as it was classified in Class 1 using the PROMETHEE-FlowSort method, confirming its leading position in terms of ERP integration maturity and readiness for further stages of digital improvement. The fact that the IT industry retained its highest position, according to respondents from Serbia, Slovenia, and experts, confirms the consistency of assessments and the objectively dominant position of this sector in the digital transformation process. This alignment further justifies the application of the PROMETHEE-FlowSort method, which favors the alternatives with the most pronounced performance and demonstrates the robustness of the model in identifying industries with the highest degree of ERP maturity. The IT industry is characterized by its embrace of advanced technology and a well-established digital infrastructure, which enables the efficient and rapid implementation of business processes and the adoption of modern solutions, including ERP systems. This industry sector is often at the forefront of adopting new technologies thanks to its expertise and available resources. Developed IT companies, in particular, are characterized by having highly specialized experts trained to effectively implement and manage ERP systems [87]. Furthermore, the IT industry is often driven by innovation and constantly seeks new ways to optimize business processes and achieve competitive advantages. By implementing ERP systems, companies can realize significant benefits in terms of cost reductions, productivity gains, and improved overall operational efficiency, enabling them to take a leading position in their industry.

ERP users from Serbia and experts in the food and beverage industry were classified in Class 2. Meanwhile, ERP users from Slovenia, including those in the aviation and food and beverage industries, were also classified in Class 2. Similarly to the results obtained from the TOPSIS-Sort, the PROMETHEE-FlowSort indicates that the majority of industries are classified in Class 3.

Nevertheless, by comparing the results of industry classification based on TOPSIS-Sort (Table 7) and PROMETHEE-FlowSort (Table 8), it is noticeable that TOPSIS-Sort did not classify any industry in Class 1, while PROMETHEE-FlowSort recognized all three classes. This deviation may be due to the different methodological approaches employed by these MCDA methods. The PROMETHEE-FlowSort approach utilizes pairwise comparisons and thresholds for classes, allowing for more precise differentiation of industries. In contrast, the TOPSIS-Sort approach, based on distance, does not recognize any industry as belonging to the highest class [88].

On the other hand, the Spearman rank coefficient for classifying industries using TOPSIS-Sort and PROMETHEE-FlowSort was calculated for all three scenarios [89]. For scenario S1, the Spearman rank coefficient (ρ) is 0.995 (p = 0.000), for S2, it is 0.989 (p = 0.000), and for S3, it is 0.995 (p = 0.000). This indicates a strong consistency in the ranking and categorization of the alternatives, thus confirming that both approaches can be considered reliable and mutually complementary tools to support decision-making related to ERP systems. Additionally, Cohen’s kappa coefficient (κ) has been calculated [90]. The results show substantial agreement across all scenarios, with κ values of 0.797 (p < 0.001) for S1, 0.833 (p < 0.001) for S2, and 0.797 (p < 0.001) for S3.

6. Conclusions

ERP systems are key digital tools for businesses seeking to improve business processes, optimize resources, and increase overall efficiency. Understanding the factors that shape attitudes toward the adoption of digital transformation is essential for both theoretical studies and practical implementations in industry. This study responded to the abovementioned challenge by developing and applying an innovative methodological framework, designed to provide a deeper insight into the interdependence of factors that influence the readiness of different industrial sectors to adopt ERP systems and to enable their reliable classification according to the degree of acceptance.

The scientific novelty of this paper is reflected in the development and application of a new methodological framework that integrates the PLS-SEM and TOPSIS-Sort methodologies to classify industrial sectors according to the degree of ERP system adoption. The research contribution to the field of advanced information technologies for managing business systems includes identifying and quantitatively validating key drivers that lead to increased usage of ERP systems. Therefore, this study contributes to the literature by extending the application mechanism into a more holistic, practically relevant evaluation framework.

In this context, the application of the PLS-SEM-TOPSIS-Sort approach represents a significant methodological step forward, as it combines the ability of the SEM method to identify and confirm key factors with the advantage of the MCDA approach in structuring and comparing industries according to the findings. Additionally, the high degree of concordance between the results and those obtained using the PROMETHEE-FlowSort method confirms the robustness and reliability of the applied framework, further strengthening its potential for future research and practical applications. The study makes an important theoretical contribution by enriching the current literature on the application of advanced numerical tools to the problem of assessing industry readiness to adopt ERP systems. Moreover, this framework offers a comprehensive approach to reviewing and improving research papers that integrate SEM, TOPSIS, and SORT methodologies, thereby opening up avenues for further academic research and comparative studies across various industries and regional contexts. Ensuring the applicability of the proposed framework hinges on clearly justifying the integration of methods, maintaining methodological rigor, and translating complex results into practical insights. Additionally, this research proposed a methodological framework that practitioners can utilize to develop strategies for promoting the effective adoption of ERP systems across different industries. The study offers valuable guidance for managers and policymakers, presenting actionable implications for enhancing ERP adoption and digital transformation readiness. Managers can utilize these insights to evaluate performance, allocate resources more effectively, and develop targeted strategies for ERP implementation that align with the organization’s priorities and capabilities. Furthermore, understanding the interdependence between technological and organizational factors helps decision-makers prioritize investments and adjust change management strategies, ultimately enhancing ERP adoption and ensuring sustained efficiency. At a strategic level, the results can support policy development and planning in industry, enabling stakeholders to promote digital transformation initiatives based on data-driven insights. In this way, the proposed methodological framework not only advances theoretical understanding but also offers practical guidance for improving efficiency, competitiveness, and innovation capacity across all industrial sectors.

Although the research provided comprehensive insights, it is essential to note that the dynamic nature of ERP systems, resulting from constant technological improvements and updates, limits the results to a current state view. Future research should encompass a broader range of industry sectors and consider integrating additional MCDA methods or combining the proposed methodology with machine learning tools to enhance accuracy and utilization capabilities.