Determinants of Organizational Net Benefits in Enterprise Resource Planning-Based Core Insurance Systems: An Extended DeLone–McLean IS Success Model

2.1 Information system

An information system can be defined as an integrated set of interconnected components that collectively function to collect, process, store, and distribute information. This information plays a crucial role in supporting core organizational functions, including data provision, data visualization, data analysis for decision-making, as well as communication and coordination within the organization [1]. An IS comprises several key components, namely hardware, software, data, people, and procedures. These elements interact systematically to support organizational operations and facilitate effective decision-making processes [2].

2.2 Enterprise Resource Planning

ERP refers to an integrated business system designed to unify and coordinate data across an organization within a single, structured platform that meets the needs of multiple functional areas [3]. ERP systems aim to enhance efficiency in key operational domains, such as purchasing, accounting, manufacturing, and sales, by integrating processes that were previously managed separately through legacy or standalone systems [4].

ERP systems are designed to be adaptable to organizational requirements, enabling effective data integration and process alignment across different functional units. Organizations adopt ERP systems to improve operational performance, either by supporting new business models or by optimizing business processes through better utilization of information [5].

In a broader context, information systems also play a strategic role by enabling organizations to develop products, services, and capabilities that provide a competitive advantage. This perspective underlies the concept of strategic information systems, which align with business strategies and support organizational sustainability in competitive environments [5]. Strategic information systems are described as instruments that enable the construction of knowledge, changing of knowledge, and sharing of information within the course of creation and performance of business strategies [6].

2.3 DeLone and McLean Information Systems Success Model (DeLone & McLean, 2003)

The Information System Success Model developed by Delone and McLean is among the widely used models in finding out the success or failure of embraced/adapted IS [7], as shown in Figure 2.

The following are the variables measured using the Delone & McLean IS Success Model:

2.3.1 System quality

IS quality refers to key system attributes, including reliability, availability, and usability, and represents a critical determinant of user satisfaction [9]. Within the DeLone and McLean IS Success Model, system quality is primarily associated with technical performance dimensions such as reliability, response time, security, and flexibility. However, as the counterparts of the D&M IS Success Model spread the light [10], in addition to technical efficiency, system quality is also evaluated based on user perceptions, particularly in terms of ease of use and overall user-friendliness. A well-designed system is therefore expected to enhance usability and support effective interaction.

2.3.2 Information quality

Information quality reflects the extent to which system-generated information is accurate, relevant, up-to-date, and complete. It is a fundamental factor contributing to IS success [11]. High-quality data is essential for ensuring the reliability of information and its usefulness in supporting business processes and decision-making. Conversely, poor data quality may reduce information value and undermine user confidence. Negative perceptions of underlying data quality can lead to unfavorable evaluations of overall information quality [12].

2.3.3 Service quality

Service quality in the context of IS refers to the level of support provided to end users. This includes aspects such as reliability, responsiveness, accuracy, help desk support, and user training. It reflects the extent to which IT services meet user expectations and system requirements [13]. High service quality is associated with timely support, dependable assistance, and clear communication, all of which contribute significantly to user satisfaction [14].

2.3.4 System use and intention to use

System use and intention to use represent the degree to which users engage with, or intend to engage with, a particular system. Intention to use is commonly regarded as a precursor to actual system usage. As an indicator of system effectiveness, it reflects users’ willingness to adopt and utilize the system under varying conditions. However, the relationship between intention and actual usage is not always consistent, as behavioral intentions do not necessarily translate into observable actions in all contexts [15].

2.3.5 User satisfaction

User satisfaction refers to the extent to which users feel satisfied with their experience in using an IS. It is considered a central construct in the DeLone and McLean IS Success Model and a key indicator of system success [16]. User satisfaction will be especially relevant in cases when the use of the system is necessary and possibly unplanned and where frequencies are not a sound indicator of system success [17].

2.3.6 Net benefits

Net benefits represent the overall value generated by an IS for individuals, organizations, and broader social contexts. These benefits may vary depending on the study context, system characteristics, and level of analysis. While user satisfaction and system usage indicate system success, net benefits provide a more direct measure of the outcomes achieved through system implementation [18]. Empirical evidence from a study on Thailand’s national pension system, analyzed using structural equation modeling (SEM), indicates that system quality can enhance perceived ease of use, although system instability may reduce perceived usefulness [19].

2.4 Trust and Perceived Usefulness variables

In addition to the six elements in the DeLone and McLean IS Success Model, this paper will use two other elements that enabled a meaningful impact on the IS success: perceived usefulness and trust.

2.4.1 Trust

Trust is characterized as the readiness of the individual to depend on their own belief in the competency and honesty of a third party. Within the scope of IS, trust refers to the trustworthiness of some application by a user. It is also a very important element in the acceptance of technology since users tend to isolate technology that they lack faith in. Simply stated, trust is the assurance that the capabilities, goodwill, and integrity of a service provider of an application are perceived by the users. On the other hand, distrust may predispose poor user experiences and slow down the adoption of systems [13].

2.4.2 Perceived usefulness

Perceived usefulness has been demonstrated in previous studies to have a major impact on user satisfaction with the IS [19, 20]. Perceived usefulness. Within the theory, perceived usefulness is a perception by an individual that, through using a certain system, better job performance would be involved [21].

This principle is subjective in nature since users evaluate the possibility of improving or enhancing individual or collective performance in an organization through technology [22]. The higher the perception of a user with regards to the usefulness of an IS, the greater the anticipated gains in efficiency and overall performance [23].

4.1 General description of respondents

The questionnaire will be distributed to all internal workers of the XYZ Life Insurance Company who have an account and a password to access the Core Insurance System. The population needed for this study was 270 respondents, who were the staff members of both the head office and branch offices of the XYZ Life Insurance Company. In gender composition, the respondents were stratified into two groups: male, 41.5% or 112 respondents, and female, 58.5% or 158 respondents.

The respondents were stratified according to their ages as follows: 20-30 years old, 65.2% or 176 respondents, 31-41 years old, 27.4% or 74 respondents, 42-52 years old, 6.3% or 17 respondents, and over 52 years old, 1.1% or 3 respondents.

In terms of department, the respondents were spread as follows: Policy Holder Services-10.7% or 29 respondents, Actuarial Services-11.9% or 32 respondents, Sales Channel Planning & Administration 23% or 62 respondents, Accounting 8.1% or 22 respondents, Finance 5.6% or 15 respondents, Claim 7.4% or 20 respondents, Business System 14.4% or 39 respondents, Marketing Officer Operations 9.6% or 26 respondents and Underwriting.

The percentages and numbers of the respondents according to job position will be in the eight levels of Vice President 2.6% or 7, Assistant Vice President 3.7% or 10, Senior Manager 4.4% or 12, Manager 11.9% or 32, Assistant Manager 8.5% or 23, Supervisor 8.5% or 23, Senior Officer 18.5% or 50, and Officer 41.9% or 113.

4.2 Measurement model

4.2.1 Evaluation convergent validity

Convergent validity refers to the degree to which a set of indicators consistently represents a single underlying construct by sharing a high proportion of common variance. In this context, a construct is expected to exhibit strong associations with its corresponding indicators, while maintaining weaker relationships with unrelated constructs. The evaluation of convergent validity is typically based on two primary measures, namely outer loadings and the Average variance extracted (AVE) [38].

4.2.2 Loading factors

The outer loading value is the extent of reliability of any indicator within a construct. Outer loading values of greater than 0.708 are desirable as they imply that the construct accounts for over 50 percent of the variance of the indicator [39].

Table 2. Validity test result

The results of the validity test, shown in Table 2, confirm that the coefficients of all the variables and indicators range above 0.7.

4.2.3 Average variance extracted

AVE measures the convergent validity of a scale. An AVE value of greater than 0.5 is desirable, and in this case, the indicators explain over 50 percent of the variance in the reflective indicators successfully in relation to the latent construct [40]. This suggests that the model effectively captures the relationship between indicators and their underlying construct, as shown in Table 3.

Table 3. Average variance extracted (AVE) test

AVE is unitless. AVE ≥ 0.50 indicates adequate convergent validity (the construct explains at least 50% of indicator variance).

4.2.4 Evaluation discriminant validity

Discriminant validity was assessed using the Heterotrait–Monotrait (HTMT) ratio. The criterion for acceptable discriminant validity is that HTMT values should be below 0.90 for constructs that are conceptually similar, while a more conservative threshold of 0.85 is applied when constructs are theoretically more distinct. This assessment can be further strengthened through a bootstrapping procedure, in which the upper bound of the 95% confidence interval must also remain below the specified cutoff values (0.90 or 0.85). It is important to note that HTMT values are standardized ratios and therefore do not carry units.

Table 4. Cross-loading test

Cross-loading has been applied to ensure that the loading of an indicator is maximized to the latent variable he/she is destined to measure. A construct is taken to be valid in case its cross-loading figure is above 0.7 and higher than the cross-loading numbers of items on other scales [42], this means that the indicator best describes the related construct as compared to any other construct. The results presented in Table 4 reveal that all the indicators are highly correlated with their corresponding constructs as opposed to the other constructs.

4.2.5 Reliability test

The reliability test checks the comparability and stability of a measuring instrument in giving reproducible results in similar circumstances [43]. The primary rationale behind a reliability test is to assess how far all indices of a variable provide consistent findings. The level of reliability can be checked on composite reliability and Cranach’s Alpha, with the range of recommended coefficients exceeding 0.7 to denote a high level of reliability [39], as shown in Table 5.

Table 5. Reliability test result

4.2.6 Model Fit (SRMR) and Global Model Index (GoF)

Model fit was assessed using the Standardized Root Mean Square Residual (SRMR) and complementary fit indices provided by the PLS-SEM output (d_ULS, d_G, Chi-square, and NFI). SRMR values below 0.08 are commonly interpreted as indicating an acceptable approximate model fit in PLS-SEM reporting. All indices reported in this section are unitless.

The SRMR values for both the saturated model (0.066) and the estimated model (0.079) indicate an acceptable approximate fit based on the < 0.08 guideline, suggesting that the specified structural relations do not introduce substantial residual discrepancies. The remaining indices (d_ULS, d_G, Chi-square, and NFI) are reported descriptively to provide additional context on model fit.

Table 6. Model fit indices output

In addition, a global goodness-of-fit (GoF) index was reported as a descriptive overall model quality indicator by combining the model’s average convergent validity and explanatory power. Based on the AVE values in Table 3 (average AVE = 0.662) and the endogenous R² values in Table 6 (average R² = 0.405), the GoF value obtained was 0.518. This GoF value suggests relatively strong overall model performance; however, it should be interpreted cautiously and treated as supportive rather than definitive evidence of fit.

4.3 Structural model

The assessment of the structural model was done along the framework of the conceptual research model. This criterion was decisive in determining the R² and path coefficients. The evaluation of the structural model was based on the relevance and compatibility of the research variables in the model. Figure 4 shows the variables used in this research.

Figure 4. PLS path model (Initial structural model configuration)

4.3.1 R-Square coefficient

Within the context of SEM, the coefficient of determination (R²) reflects the proportion of variance in the dependent variable that can be accounted for by the independent variables in the model. A higher R² value indicates stronger explanatory power, meaning that a larger share of the variation in the endogenous construct is captured by the proposed model.

As a general guideline, R² values of 0.67 are considered substantial, 0.33 indicate moderate explanatory capability, and 0.19 suggest weak explanatory power. These benchmarks serve as useful references for evaluating the overall robustness and predictive relevance of the structural model [44].

Table 7. R-squared value

The R² values presented in Table 7 indicate the explanatory power of the structural model for each endogenous construct. The R² value for User Satisfaction (0.431) suggests a moderate level of explanatory power, meaning that 43.1% of the variance in User Satisfaction is explained by its predictor variables. Similarly, the R² value for Net Benefits (0.408) indicates that the model explains 40.8% of its variance, which can also be considered moderate.

4.3.2 Path coefficient

The subsequent section provides an explanation of the measurements of the inner model derived from the analysis results, as shown in Figure 5.

Figure 5. PLS Structural model assessment results (Bootstrapping t-statistics and supported paths)

Table 8 present the results of the t-statistic test, indicating that 9 out of the 11 proposed hypotheses were accepted. The acceptance was based on the criterion that the t-statistic value exceeded the critical threshold of 1.96, signifying statistical significance at the 5% level. Conversely, 2 hypotheses were rejected, as their corresponding t-statistic values fell below 1.96, suggesting that the relationships proposed in those hypotheses were not supported by the data. Table 8 shows the t-statistic results, where 9 of 11 hypotheses were accepted (t > 1.96, p < 0.05), while 2 were rejected (t < 1.96).

Table 8. Path coefficient

1. H1 (Information Quality to Intention to Use):

Information quality does not significantly influence intention to use, consistent with prior e-learning studies, though conflicting with findings in e-tourism contexts.

2. H2 (Information Quality to User Satisfaction):

Information quality has a significant positive effect on user satisfaction; accurate and relevant information enhances user confidence and system experience.

3. H3 (System Quality to Intention to Use):

System quality significantly affects intention to use; reliable, accessible, and well-functioning systems encourage continued usage.

4. H4 (System Quality to User Satisfaction):

System quality does not significantly impact user satisfaction, aligning with several prior studies but differing from findings in accounting systems.

5. H5 (Service Quality to Intention to Use):

Service quality significantly influences intention to use; responsive and supportive services increase user willingness to adopt the system.

6. H6 (Service Quality to User Satisfaction):

Service quality has a significant positive effect on user satisfaction; better responsiveness and assurance improve user experience.

7. H7 (Trust to User Satisfaction):

Trust significantly enhances user satisfaction; higher perceived reliability and provider commitment lead to greater satisfaction.

8. H8 (Perceived Usefulness to User Satisfaction):

Perceived usefulness significantly affects satisfaction; systems that improve performance and task completion increase user approval.

9. H9 (User Satisfaction to Intention to Use):

User satisfaction significantly drives intention to use; satisfied users are more likely to continue using the system.

10. H10 (Intention to Use to Net Benefits):

Intention to use significantly contributes to net benefits; higher usage leads to greater organizational value.

11. H11 (User Satisfaction to Net Benefits):

User satisfaction significantly affects net benefits; satisfied users increase system utilization and overall benefits.

The inclusion of the two variables, Trust and Perceived Usefulness, along with the appraisal of the Core Insurance System implementation with reference to the implementation of the DeLone and McLean IS Success Model, has two principal implications. The participants in the study included a sample of 270 internal staff members who were active users of the Core Insurance System. Pilot data consisting of 11 hypotheses and 8 variables, which included Information Quality, System Quality, Service Quality, Use, User Satisfaction, Net Benefit, Trust, and Perceived Usefulness’s, were analyzed using a PLS-SEM.

The hypothesis results show that most hypotheses are supported. Information quality, service quality, trust, and perceived usefulness positively influence user satisfaction, as users feel more satisfied when the system provides accurate information, responsive services, and reliable performance. In addition, system quality, service quality, and satisfaction significantly increase users’ intention to continue using the system, especially when it is stable and easy to use. Finally, both user satisfaction and usage intention contribute positively to net benefits, including improved work efficiency, better business processes, and stronger achievement of organizational goals.

On the other hand, the results also identify two hypotheses that were rejected, as detailed below:

1. Information Quality does not have a significant influence on Intention to Use.
2. System Quality does not have a significant influence on User Satisfaction.

Despite these findings, this study has several limitations that open avenues for future research. Future studies may examine system personalization features (e.g., role-based dashboards or adaptive interfaces) and their impact on user satisfaction and continued usage intention. In addition, the role of data governance and automated validation mechanisms can be further explored as key antecedents of information quality and trust.

Further research may also apply multi-group analysis across departments, job levels, or user experience to identify potential differences in system perception and usage behavior. Moreover, incorporating objective operational performance metrics, such as processing time, rework rate, and ticket resolution time, would provide a more comprehensive assessment of net benefits.