The influence of big data on decision-making in the engineering procurement construction industry in Indonesia

Big data analytic capability is defined as the competency to provide business insights that use data management capabilities, infrastructure and talent to transform a business into a competitive force (Wamba et al. 2017; Yasmin et al. 2020), and the ability of a company to capture and analyse data towards a generation that is insightful in effectively managing, deploying its data, technology and talent people (Pappas et al. 2020).

Big data analytic capability has data management capabilities that can transform a business into a competitive force (Wamba et al. 2017; Yasmin et al. 2020). One way to become a competitive force begins with a decision based on insightful data (Merendino et al. 2018). This is in line with the opinion that companies can utilise data and then develop it to become successful companies (Mcafee and Brynjolfsson 2012; Westerman et al. 2014). They have the power of very large, strategic and valuable data according to their respective business interests (Marr and Ward 2019; Mishra et al. 2019).

Many organisations have adjusted their strategic decisions based on big data because the capabilities of BDA have transformed the data capturing process into a strategic one (Gupta et al. 2021). BDAC provides templates and several matrices that can bridge between DM and its BDA capabilities so that the stability of decisions based on BDA becomes strategic and more important for the quality of the resulting data analytics as a basis for effective DM (Adrian et al. 2017). It is the management of BDA capabilities that affects DM capabilities (Awan et al. 2021). BDA equipped with intelligence and knowledge is called BDAC used for DM (Lichtenthaler 2019; Shamim et al. 2019).

The advantage of BDAC is its ability to analyse big data by handling complicated and complex tasks that are difficult to perform manually (Schildt 2017). The ability to analyse big data encourages the quality of DM in insightful organisations based on data (Awan et al. 2021). Therefore, the success of an organisation depends on the ability of its managers to make the right decisions (Edwards et al. 2000) and quality based on data-based insights, business intelligence and business analytics (Awan et al. 2021). For these reasons, we propose the following hypothesis:

H1: BDAC has a positive influence on DM.

Decision-making can be defined as a process for obtaining quality information that enables decision-makers to better understand the impact of feedback processes, non-linear relationships between variables, and time delays on the performance of complex systems (Kelso 2008).

In this study, DM is central to insightful and valuable information obtained from MBV (Porter 1980, 1985; Grant 1991) in the form of competitors’ position towards the company and winning opportunities. In terms of the RBV by Barney (1986), Makadok and Barney (2001) and Peteraf (1998), it is in the form of the capabilities of existing resources, physical resources, and non-physical resources. In terms of IBV, it is a combination of human resource capabilities with data, information and knowledge (Lichtenthaler 2019; Wamba et al. 2020). It is undeniable that information is an important source of input in DM throughout strategic management (Shrivastava and Grant 1985; Westley and Mintzberg 1989). DM based on big data is categorised as valuable information creativity (Secundo et al. 2022). The quality of big data analysis is an important element for effective DM in the BDA implementation cycle, because it is often used by decision-makers in the DM process (Adrian et al. 2017).

Furthermore, research on the DM process developed towards rational approaches was conducted by Eling et al. (2014), Kahneman and Klein (2009) and Kolbe et al. (2020). In this research, we use three dimensions of DM: rationality, quality and commitment. Rational decision is a DM process that uses information sources that are relevant to the decision and depends on the process of analysing information in determining choices (Dean and Sharfman 1993b). Other researchers suggest that rational decisions generally have a positive influence on the outcome of a decision (Bourgeois and Eisenhardt 1988). The indicators of the use of rational DM are based on cognition, knowledge, and accurate and correct information (Van Riel et al. 2004; Ram and Ronggui 2018).

Previous researchers in the construction industry have emphasised on time planning and project time estimation as part of the formulation strategy (Dunovic et al. 2021). DM results in policy strategies and formulation strategies (Savio and Nikolopoulos 2013). In addition, according to Ding et al. (2018), there is need for a model in DM that is used to select strategies that suit the needs of the organisation. Among them is that DM has connected with all levels, not only hierarchically ranging from strategic, tactical and operational but also complemented by networking (Gartner 2020). For these reasons, we propose the following hypothesis:

H2: DM has a positive influence on FS.

Savio et al. (2013) stated that IS has an important role in the DM process. According to Song et al. (2018), DM is inherently linked to the strategy to be used. Similarly, Bryson and Roering (1988) convey that decision-makers will select strategies based on the decisions made. The initial method for making decisions is crucial because it concerns those who will implement the strategy (Hickson et al. 2003). Furthermore, decisions can be used to develop and integrate strategies that are essential for the organisation to succeed (Jones et al. 1992; Rusmin et al. 2014). IS can also be defined as a dynamic, iterative and complex process that encompasses various managerial activities in turning strategic plans into reality to achieve organisational goals (Tawse and Tabesh 2021).

The DM process can be utilised to develop and integrate strategies that are critical to organisational success (Savio and Nikolopoulos 2013). A comprehensive DM process fosters the development and integrates decisions towards a more strategic direction (Hitt et al. 2017). According to Tawse et al. (2018), DM affects organisational performance through IS. Additionally, the decision process is also used by the organisation to develop and integrate strategies for success (Liberation and Trope 1998). Given the various considerations, there is a great deal of interest in examining the relationship between DM and an organisation to be successful in implementing strategies, which can motivate the parties involved in IS (Mintzberg 1978; Hey and Knoll 2011; Song et al. 2018). A comprehensive DM process is one that directs decisions towards a more strategic direction (Fredrickson 1985). Similarly, Oyewobi et al. (2016) state that DM affects organisational performance through the IS.

However, success of the IS is a critical factor for organisations in running their business. Implementing an effective strategy is a pivotal stage in an organisation’s success and can serve as a potential source of competitive advantage (Hitt et al. 2017). For these reasons, we propose the following hypothesis:

H3: DM has a positive influence on IS.

Rumelt et al. (1991) state that the initial understanding of strategy is about differentiating the company from its competitors, while the general management task is to adjust and update the company’s resources over time, competition and changes that occur. This was further developed by Alkhafaji (2011). FS can be defined as a process that involves making strategic decisions regarding the organisation’s mission, philosophy, goals, policies and methods to achieve organisational goals.

Furthermore, according to Mintzberg (2019), FS is the process of strategy development and includes internal and external analysis of the organisation and industry and defines its goals. Agyapong et al. (2020) stated that the ability of IS is more important than the quality of the strategy itself. Salavou (2015) stated that there are general approaches that have the potential to succeed in outperforming competitors in their industry, including cost leadership. McGee et al. (2016) stated that FS needs to involve various choices as a consideration in implementing the operations strategy, because according to Tawse et al. (2018), there is a difference between what is needed when implementing a strategy and the experience and emotions of other involved parties, so that the company’s ability to formulate strategies is more developed than the ability to implement (Mintzberg 1978).

There is awareness from all parties involved in the FS and IS when making a transition from the formulation strategy to the IS to try to jointly overcome the existing problems (Engert and Baumgartner 2016). For this reason, specific factors are needed when making the transition (Lau et al. 2020) with in-depth analysis (Engert and Baumgartner 2016). Conant et al. (1990) indicate that formulation strategy or planning strategy has an influence on IS. Stakeholders believe that better formulation strategies are needed to ensure the success of implementation strategies (Kalyan et al. 2025). However, Coricelli et al. (2018) state that the transition from FS to IS runs successfully when there is involvement of all interested parties in its implementation by accommodating the aspirations of all circles or levels within the organisation. For these reasons, we propose the following hypothesis:

H4: FS has a positive influence on IS.

Rumelt et al. (1991) state that the initial understanding of strategy is about what distinguishes between companies, differences with their competitors. According to Grant (2001) and Johnson et al. (2008), FS is a process that involves making strategic decisions that are influenced by external and internal factors. If there is a relationship between strategy and FP, it is essentially satisfying stakeholders who can be considered to have growth, profitability, market value, customer satisfaction, employee satisfaction, social performance and environmental performance (McGee 2006; Santos and Brito 2012).

Therefore, in order to satisfy stakeholders, FS must be based on in-depth analysis by considering external factors. From MBV, competitive advantage from an external factor power perspective is seen, with sources of power coming from market monopolies, barriers to entry and high bargaining power (Tallman 1991; Collis 1994). From previous research as well, it is stated that there is a positive relationship between the development of a comprehensive strategy and FP (Papadakis 1998; Khatri and Ng 2000). Proponents of strategic planning or FS describe the potential to positively impact organisational outcomes, especially FP (Halachmi 1986; Poister et al. 2022).

Other researchers state that the impact of FS on performance produces mixed results (Poister et al. 2022). Furthermore, the impact of various aspects of strategic planning was examined, where there were differences in setting targets, issuing internal analysis results, conducting external analysis and developing actions (Poister et al. 2013, 2022). The results of research by Ateş et al. (2020) and Balogun and Johnson (2005) state that the failure to implement the planned strategy is due to lack of commitment in implementing the strategy. For this reason, it is necessary to create a mechanism that involves components in the organisation so that they have a sense of responsibility in the success of the company’s performance (Coricelli et al. 2018).

The diversification strategy, which is part of the strategy formulation, has less significant impact on performance. Likewise, the results of research from Fredrickson and Mitchell (1984) show that there is a negative relationship between FP and a comprehensive strategic formulation when the environment is unstable. There is a significant difference in performance between high diversification and low diversification (Montgomery and Collis 1995). But according to Yukl et al. (2005), there are mixed findings on the relationship between diversification strategy and FP. For these reasons, we propose the following hypothesis:

H5: FS has a positive influence on FP.

Implementation of strategy can be defined as everything related to various managerial activities in order to implement planning strategies such as the creation of the type of organisational structure, the type and source of information systems, suitable leadership and the type of control mechanism (Mintzberg and Westley 2010; Alkhafaji 2011). IS can also be defined as a dynamic, iterative, complex process consisting of various managerial activities in turning strategic plans into reality to achieve organisational goals (Tawse and Tabesh 2021).

The IS process is aimed to improve the achievement of the firm’s business mission and vision through quality products and customer service (Rani 2019). Effective strategy implementation requires compatibility between the elements that exist within the internal organisation (Venkatraman and Camillus 1984). Implementing an effective strategy is a critical stage of an organisation’s success and a potential source of competitive advantage for the organisation (Hitt et al. 2017; Tawse and Tabesh 2021).

Hoskisson and Hitt (2007) and Johnson et al. (2008) stated that the IS needed by an organisation is to improve FP. Furthermore, Pryor et al. (2010) have stated that it is a major challenge in the IS toward coordinating and integrating the activities of participating individuals and functions. According to Johnson et al. (2008) and Sirmon and Hitt (2009), IS is required by an organisation to improve FP. The success of project analysis is widely used in determining the project success factors, including quality improvement, project efficiency and effectiveness (Rehan et al. 2024), where the success of projects in the EPC industry can be synonymous with the success of the firm.

Firm performance is an important topic in strategic management and international management (Chan et al. 2008). FP can be defined as the results and basis that can satisfy stakeholders, which are considered to have growth, profitability, customer satisfaction and performance (McGee et al. 2008; Santos and Brito 2012). The results of the study by Fatima et al. (2024) emphasise that to measure the FP of construction projects, a dimensional approach is needed, involving cost, time and quality.

From the RBV perspective, FP is derived from two organisational capabilities: operational capability and dynamic capability (Eisenhardt and Martin 2000; Helfat and Peteraf 2003). There is a failure in strategy implementation when the strategy has been formulated by top management but not implemented by managers or the strategy is implemented but the results are not satisfactory (Cândido et al. 2015). The manager as a leader in the construction project plays an important role to achieve performance in a project and finally, the accumulation of a project’s performance is FP (Imran et al. 2025). For these reasons, we propose the following hypothesis:

H6: IS has a positive influence on FP.

From the explanation about the relationship between latent variables, we have proposed six hypotheses, which are the model of our research, as shown in Figure 1.

Fig. 1:

Source(s): Author’s own creation.

The unit of analysis in this research survey is the EPC company, which include a private company, state owned enterprise and joint venture or foreign investment that have products or services produced by providing enormous impacts and benefits to the community, nation and state, such as electrical energy, fertilisers, motor vehicle fuel and cement.

In addition, in 2021, the construction industry made the fourth largest contribution to Indonesia’s national economy at 10.48% of GDP after the industrial sector (18.3%), trade (12.71%) and agriculture (11.39%) https://www.bps.go.id.

The development of a survey through a unit of analysis that considers the contribution to a country’s economy will have a significant impact on the country’s income and the prosperity of its people. EPC companies as units of analysis play a very important role in this research, represented by company leaders or their equivalents, because they are decision-makers who can cause a company to succeed or fail.

Before the research survey was conducted, all questions in this study were checked for sentence and word understanding by two people who were not respondents, but students who had graduated from S3 strategic management Faculty Economic and Business (FEB) UI. The results of their checking changed the sentences and words for better than previously found.

The research questions of this study for each construct are developed based on previous research, including research questions on the BDAC construct by Akter et al. (2016) and Shamim et al. (2020), the DM construct by Dooley and Fryxell (1999) and Kolbe et al. (2020), the FS construct by Agyapong et al. (2016) and Parnell and Carolina (2010), the IS construct by Cepeda-Carrion et al. (2012) and Pollanen et al. (2017) and the FP construct by Ji-fan Ren et al. (2017) and Wang et al. (2016).

The purpose of survey in this research is to provide scientifically gathered information to work as a basis for our conclusions.

The sample of this research was collected from EPC service companies operating in Indonesia, which are members of the Association of Indonesian National Design and Build Companies (AINDBC) as a population. From 116 companies incorporated in AINDBC, 76 companies responded to the research survey. These comprise companies owned by the Indonesian government, multinational companies and private companies.

Company leaders such as BOD and managers who were respondents in this study representing their respective companies were asked to complete the research questionnaire 1 h before the seminar on the use of big data in DM at EPC service companies. The aim of answering questions from company leaders provide answers that are completely factual, without embellishment or bias, and more objective in nature.

Respondent data collection in this research survey is the two-stage approach method of survey (Naresh and Satyabhusan, 2016): pre-test stage and main test stage. At the pre-test stage, the research survey was distributed via email dated, 22 November 2022 and the period of data collection at the pre-test stage is approximately 4 months. The purpose of conducting a pre-test is to find out how much the respondent understands the questions asked, so that at the main test stage the questions submitted to the respondent are better and more understanding (Presser et al. 2016). Pre-test is also used to detect the level of misunderstanding, ambiguity, bias or doubt or other difficulties that respondents may face (Perneger et al. 2015).

The data collection method begins with creating a questionnaire containing questions related to the dimensions of each construct derived from literature review, previous research results and combined with the objectives of the study. All questions are packaged in the form of surveys to respondents who are members of AINDBC, which are adjusted to the profile of the respondents to be addressed, including company leaders or their representatives with groupings based on positions within the company.

This research questionnaire uses a Likert scale from 1 to 6 (Byrne 2016) as follows: 1 = strongly disagree, 2 = disagree, 3 = somewhat disagree, 4 = agree, 5 = strongly agree and 6 = very strongly agree. The mean value is the average value of a measurement, and the standard deviation is the average deviation from the mean value of the measurement results. The consideration of reducing the tendency of respondents is to choose neutral ones so that they are smarter in determining their choices.

The purpose of determining the standard deviation is to understand the variation or spread of the data relative to its mean value. If the standard deviation is larger, the data are more varied or further from the average value, and conversely, if the standard deviation is small, the data are closer to the average value. The size of the standard deviation value is influenced, among other things, by the range of measurement scale used and the respondents’ perception or understanding of the questions.

The results of data collection at the pre-test stage were obtained from 30 respondents who represented 30 companies or 30 analysis units from 116 companies incorporated in AINDBC or around 25.86%. From the results of data processing on 30 companies, there are several low factor loadings <0.7. From the results of pre-test data processing using SPSS version 25 (IBM), there are 75 indicators consisting of 57 indicators having positive statements and 18 having negative statements. Of the 18 negative statements, 17 have a Kaiser-Meyer-Olkin (KMO) value <0.05 and Cron-bach’s alpha <0.5, where the minimum limit of KMO >0.05 and Cronbach’s alpha >0.5. (Hair et al. 2010).

Furthermore, the second research survey was distributed after making improvements to the editorial of several research statements, especially on their factor loading values which tend to be <0.5 (Chin 1998; Hair 2010). The distribution of the research survey in the second stage is different from the first stage because it is through google. doc media before the webinar event on the EPC industry on 12 July 2023.

Data processing in this research uses SEM – PLS version. 4.0.9.5 (IBM). This is done by focusing on explaining variance in the dependent variable when examining the model (Pappas et al. 2020). The purpose of using Partial Least Squares – Structural Equation Modeling (PLS-SEM) is to make predictions that maximise the explanation of variance in the dependent variable (Anderson et al. 2019) and develop theory in exploratory research (Sarstedt et al. 2014).

According to Henseler and Chin (2010) and Hair et al. (2019), in measuring a model using Smart PLS, one should go through the two-stage approach. The first approach is through model measurement testing, such as Reliability test is used to measure the internal consistency of items in each category by determining the Cronbach’s alpha or composite reliability (CR) value (Lee 2004). This measures how well the items in each category measure the same concept. A Cronbach’s alpha or CR value >0.6 indicates that the dataset is acceptable (Kalyan et al. 2025). Cronbach’s alpha describes the reliability of the sum (or average) of x measurements, where x measurements can represent x raters, x occasions, x alternative forms or questionnaire items (Bonett and Wright 2015).

Then, validity test refers to the ability of research survey questions to measure how accurately something is being measured (Lee 2004). It consists of convergent validity, which is indicated by a factor loading value of >0.5, although ideally it should be >0.7 (Chin and Marcoulides 1998; Hair 2010; Sarstedt et al. 2014) and discriminant validity, which is indicated by an average variance extracted (AVE) value of >0.5 (Sarstedt et al. 2014). AVE is a measure used to assess convergent and discriminant validity, which is defined as the variance in the indicator or observed variable that can be explained by the construct of the latent variable (Naresh and Satyabhusan, 2016).

The second approach is structural model testing, such as R², Q², F², variance inflation factor (VIF) and path coefficient, where R² is a measure of how strongly the variance within the endogenous variable can be explained by the exogenous variable (Hair 2019). According to Chin (1998) and Henseler and Chin (2010), the values of R² are as follows: R² = 0.67 (substantial), R² = 0.33 (moderate), and R² = 0.19 (weak). Q² is a measure that describes the accuracy of a prediction due to the relationship between exogenous or endogenous variables in predicting other endogenous variables. The value of Q² ranges from –1 to +1. If the value is close to +1, it indicates a strong relationship between exogenous and endogenous variables (Hair et al. 2011), where the value of Q² does not reflect the quality of the prediction (Hair et al., 2014).

In addition, F² is a measure that describes how strong the relationship is between exogenous variables and endogenous variables. It measures the magnitude of the effect independently of sample size and provides an indication of the relevance of an effect measured based on calculations; the values of F² are 0.02 (small), 0.15 (medium) and 0.35 (large) (Cohen and Levin 1989). VIF is a measure used to determine the level of collinearity of each construct. According to Hair (2019) and Sarstedt et al. (2023), the VIF value is considered critical if it is ≥5, there is potential for collinearity if 3 ≤ VIF ≤ 5 and ideally the VIF value should be <3. Path coefficients in SEM models are interpreted like standardised regression coefficients.

Furthermore, the coefficient linking the path between the dependent variable and the independent variable measures the expected increase in the dependent variable if the independent variable increases by one standard deviation, while the other independent variables in the regression equation are held constant (Henseler 2017). The coefficients represent the relationship between independent constructs and dependent constructs, with path coefficient values ranging from –1 to +1. Coefficients closer to +1 indicate a strong positive relationship, while coefficients closer to –1 indicate a strong negative relationship (Sarstedt et al. 2014).

In this research, the variables measured among others were BDAC, DM, FS, IS and FP, as shown in Table 3.

All variables measured in this research are latent or cannot be measured directly or through the perceptions of respondents. In this study, there are six variables that can be defined as operational variables, such as: AIC is the ability of a company to manage organisational resources and implement computer systems capable of performing processes like humans (Mikalef et al., 2021; Agrawal et al., 2019). BDAC is the ability of a company to utilise data so that it can provide added value or insight through the use of technology (Pappas et al., 2018; Akter et al., 2016). FS is the planning process for initial decision-making includes, among other things, the philosophy, mission, and policies for achieving organisational goals that are influenced by internal and external factors. IS is a process for realising what has been planned, including organisational structure and information systems, as well as mechanisms for controlling their implementation (Tan et al., 2019; Aydiner et al., 2019). FP is the output of a process that meets the expectations of stakeholders (Straub et al., 2020; Ji-fan Ren et al., 2017).

From the results of data processing on these 76 respondents, there are 49 indicators which have a factor loading >0.5 (Chin and Marcoulides 1998; Hair 2010; Sarstedt et al. 2014), from 54 of the total indicators, as shown Appendix 1. It refers to validity test, wherein 49 indicators are the full requirement. Therefore, 49 indicators were processed further and 5 indicators were dropped (Chin 1998; Hair 2010; Sarstedt et al. 2014).

The value of Cronbach’s alpha and CR in accordance with >0.6 and AVE >0.5 (Hair et al. 2019) are shown in Table 4.

Based on the structural model test as explained in section 3.3, the following are the results of structural model test using the SEM – PLS version. 4.0.9.5:

• The R² value for the variables: DM is 0.217 (low), FS = 0.654 (high), IS = 0.727 (high) and FP = 0.289 (low).

• The F² values for the variables: BDAC → DM = 0.071 (small), DM → FS = 1.886 (large), DM → IS = 0.159 (medium), FS → IS = 0.371 (large), FS → FP = 0.021 (small) and IS → FP = 0.221 (medium).

• The Q² values for the variables: DM = 0.150 (good), FS = 0.203 (good), IS = 0.333 (good) and FP = 0.217 (good).

• VIF values for the variables; BDAC → DM = 4.496 (possible collinearity), DM → FS = 1.000 (no collinearity), DM → IS = 2.886 (no collinearity), FS → FP = 3.163 (possible collinearity), FS → IS = 2.886 (no collinearity) and IS → FP = 3.163 (possible collinearity).

Based on the results of the structural test, the model shows that this research model is acceptable and can be continued to a coefficient path analysis test, as in the following paragraph.

This research study includes six hypotheses, of which five are supported by the path coefficient values presented in Table 5, while one hypothesis is not supported. The hypothesis that was not proven involves the FS and FP variables, which have a path coefficient value of –0.220. Meanwhile, other variables are accepted.

Referring to Table 5, all hypotheses are accepted except for H5, which does not prove the effect of strategy formulation on company performance. This indicates that the existing dimensions of the formulation strategy are not able to explain directly to FP. Based on the results of the model measurement test on the formulation strategy construct, the marketing strategy dimension has the highest factor loading of 0.909 with a t-value of 42.172 and a p-value of 0.000, an AVE value of 0.562 and a Cronbach’ alpha value of 0.897.

However, the results of structural tests on the formulation strategy construct on the company performance is constructed with a path coefficient of –0.220, t-value 1.079 and p-value 0.140. It can be translated that the formulation strategy construct does not have a positive and significant effect on company performance or H5 is rejected. The effect of formulation strategy on company performance is inseparable from the marketing strategy dimension in the FS construct where the marketing strategy dimension has the highest factor loading (0.909).

The results of this research are in line with the results of other studies which indicate that the diversified strategy, which is part of the formulation strategy, has less significant impact on performance (Rumelt et al. 1991). Likewise, the results of research from Fredrickson and Mitchell (1984) state that there is a negative relationship between company performance and a comprehensive strategic DM process, which is part of the strategy formulation when the environment is unstable.

The rejection of H5 also proves that there are differences in marketing strategies based on the research of Cavusgil et al. (1994), where there is a strategic relationship between marketing strategy and FP supported by a set of potentials that distinguish it from more comprehensive competitors. Then, Shaw (2012) also stated that the potential in question is an internal factor in the form of product characteristics and market characteristics. Market characteristics between mass production and EPC service markets are much different where EPC market conditions are highly dependent on external conditions, including petroleum prices and a stable environment.

Therefore, it becomes clearer that there are differences in product characteristics and market characteristics, which are strategic components of the relationship between marketing strategy and FP. In the EPC industry, the products produced by companies are determined by customers in terms of specifications, quality, time and price. Thus, it can be interpreted in this study that the products produced are not mass-produced but rather specific products, which are very different from mass-produced products (consumer goods). Other research states that the impact of strategic planning on performance yields mixed results (Poister et al. 2013, 2022).

Furthermore, Fatima et al. (2024) examined the impact of various aspects of strategic planning, where there were differences in setting targets, releasing internal analysis results, conducting external analysis and developing actions or behaviours.

The results of this study have implications for the BDAC to support the calculation and prediction process. The DM needs BDAC to provide a lot of data for calculations and predictions. For BDAC to provide data and information, company managers must prepare, among others, BDA management capability, BDA technology capability and BDA talent capability.

The results of this study also have implications for the quality of DM based on BDAC. In this study, DM has a positive and significant influence on strategy formulation. In the DM process, the commitment dimension is the most significant aspect of this study, reflecting a sense of responsibility for decisions agreed upon by company leaders and involving commitments to parties related to project completion and success. The rationality dimension plays a crucial role in producing decisions that are faster, of higher quality and more accurate. BDAC contributes rationality by providing quality and up-to-date data and information necessary for projects and other purposes. Quality decisions can be used for the next process and are easier to track based on the origin of the data and information flow.

In the IS that has a lot to do with stakeholders, managers must understand the characteristics of each stakeholder so that the project implementation can run according to the specified time and agreed budget. Managers also need to pay attention to the ability of human resources to carry out formulation strategies and implementation strategies based on their qualifications and competencies. The ability of human resources that matches the needs of the company with those of each employee is able to carry out implementation strategies in completing EPC projects to meet the expectations of customers.

The results of this study also have implications for the company’s planning stage, which requires dimensions that support the FS. In the formulation strategy, the marketing strategy dimension is a critical dimension for companies in marketing the company’s products and services. This conclusion is related to the uniqueness of the products and services of the EPC industry, which is different from other industries. The EPC industry’s marketing strategy targets stakeholders involved in project development, such as the project owner and the vendor to enhance effectiveness and precision.