Predicting Consumer Purchase Intention Through E-Catalogue: A Theory Of Planned Behavior Approach

: Through Presidential Instruction No. 2 of 2022, the government supports, allocates and spends 40 per cent of the budget value to accelerate the use of domestic goods and MSMEs in government procurement. The government has provided an e-catalogue as an online shopping application that provides products from various commodities the government needs. However, this was followed by something other than the realisation of e-purchasing on the e-catalogue. The purpose was to identify the elements that impact the desire to buy via e-catalogue. Using the criterion of commitment-making officers (CMOs) in local Indonesian governments who purchased laptop items, this research used simple random selection to get 135 respondents. An examination of the data obtained from the e-questionnaire was carried out using SEM. It was shown that several factors, including self-efficacy, perceived behavioural control, attitude, and subjective norms, directly and significantly influenced the intention to purchase. The perceived web quality, however, could have been better.


INTRODUCTION
The consumption by people in Indonesia is increasing.A widespread phenomenon is online shopping.According to a recent report by the research firm We Are Social (2023), around 178.900 million Indonesians made online purchases from 2022 to early 2023, an increase of 12.800 per cent compared to the previous year.The estimated total value of Indonesians' online shopping during the previous year reached US$ 55.970 billion, equivalent to IDR 851 trillion.The majority of spending was on electronic products.
Technological developments have impacted how consumers decide when shopping to fulfil their needs.Consumers can now easily search for online product information and compare prices before purchasing (Chiu et al., 2019).
Consumer behaviour is when individuals make choices, buy and put things to use to fulfil their wants and requirements (Jusuf, 2021).The theory of planned behaviour (TPB) tries to anticipate when and where an individual wants to do something by studying past actions (Ajzen, 2020).The TPB examines the elements that impact an individual's intention to participate in an action, providing a framework for behaviour prediction.The model categorises beliefs into three groups: behavioural, normative, and control.These groups stand for the extent to which individuals feel they have control over their actions (Bosnjak et al., 2020).
Attitudes, subjective standards, perceived behavioural control, and buying intents are some of the factors that have been shown to influence purchasing behaviour in academic studies (Ajzen, 2020;Nikolas R et al., 2022;Noor et al., 2020;Persada et al., 2021;Tarawneh et al., 2020).Information system quality also impacts e-commerce user satisfaction, mediated by behaviour intention (Kumar & Lata, 2021).Improving system quality increases user trust in using the system and contributes to increased user satisfaction (Costa & Rodrigues, 2023).The level of self-assurance in their abilities is vital in determining their inclination to make an online purchase.(Y.Li et al., 2018).Customers who are sure of themselves are more likely to be enthusiastic about e-commerce and to have a favourable attitude about online purchasing.
Along with economic growth in Indonesia, digital transformation is needed to improve efficiency and productivity and encourage inclusiveness.The role of the government is to create a comfortable business environment for MSMEs.Digital marketing capabilities and literacy have successfully improved MSME business performance (Umboh et al., 2023).One aspect that the government may play in fostering economic growth is the implementation of policies and initiatives aimed at economic development (Fahrika et al., 2020).Through Presidential Instruction Number 2 of 2022, the government is committed to accelerating and prioritising products produced by MSMEs to meet the needs of the government administration by allocating and realising 40 per cent of the value of the expenditure budget.The government also provides space for local products in the digital market through the electronic catalogue (e-catalogue) application developed by the National Public Procurement Agency (NPPA).
E-purchasing through e-catalogue is one of the main alternatives to a fast, easy, and transparent procurement process for government procurement actors.The existence of ecatalogue aims to encourage the growth of the performance of domestic business actors, create a competitive business climate, encourage the development of domestic product quality, and create reasonable product prices according to the market (Suryono et al., 2021).Another convenience of electronic catalogues is that they support the government's efforts to facilitate the recording of government procurement transactions electronically (Pawe et al., 2024).In this case, the application's efficiency and convenience are essential benchmarks for all users; thus, it is more practical for transactions (Ruiz-Herrera et al., 2023).Service and website quality can also influence consumers' purchasing decisions through their purchase intentions (Saleem et al., 2022).The allure, ease of transaction, and novelty of online purchasing options contribute to consumers' desire to purchase online Procurement parties have yet to fully optimise the digital transformation of procurement, especially e-purchasing through e-catalogue.Procurement performance in 2023 shows that the value of procurement through tenders still dominates compared to epurchasing, as shown in Table 1.The national government procurement budget consists of Ministries/Institutions and Local Government.According to the Bigbox LKPP data, the national Annual Procurement Plan 2023 shows a considerable value of around IDR 1,208 trillion.However, the realisation of procurement expenditure is still quite a big gap against the Annual Procurement Plan.
Moreover, in Local Government, the value of the realisation of procurement spending is still below 50 per cent.One of the products transacted annually by the government is a laptop.The value of government laptop procurement is quite significant, and the specifications tend to be similar (Darmawan, 2022).This large gap is an issue of the purchase behaviour intentions of procurement parties through e-catalogue.Based on research on online shopping consumer behaviour, the Theory of Planned Behavior is an approach that can be used to identify purchase behaviour intentions.
Therefore, this study attempts to establish a model to explain the purchasing behaviour of government procurement parties through e-catalogue using the Theory of Planned Behavior approach.Based on the above background, the researcher asked two questions in this study: (1) What factors influence the purchasing behaviour of laptop products in e-catalogue?(2) What are the managerial implications for increasing epurchasing transactions in e-catalogue as a government shopping platform?THEORETICAL REVIEW Consumer Behavior.Consumer behaviour refers to how consumers use their resources to obtain what they want (Schiffman & Wisenblit, 2018).These resources include time, money, and energy resources.In addition, consumer behaviour is how a person, individually and in groups, perceives, selects, purchases, uses, and evaluates a product to meet their needs (Kotler & Keller, 2018).Consumer behaviour describes a product's decision-making process, starting from the acceptance, purchase, use, and identification of the goods and services used (Petcharat & Leelasantitham, 2021).

Theory of Planned
Behavior.An improved and expanded version of the earlier theory is the Theory of Planned Behavior (TPB).Human conduct may be better understood using the Theory of Reasoned Action, including consumer behaviour, focusing on intentions and intentions as predictors of behaviour (Bosnjak et al., 2020).This theory came into existence in 1967 by Icek Ajzen, a social psychologist who has extensively used their work in fields as diverse as marketing and consumer behaviour studies.Regarding the TPB, Ajzen incorporates the concept of Perceived Behavior Control (PBC), which refers to the notion that individuals have some degree of control over their activities.
They perceived Web Quality, Attitude, and Perceived Behavioral Control Purchase Behavior Intention.Additional literature in the Theory of Planned Behavior is that website reliability also contributes to developing online purchase intentions through attitude formation and trust (Ajzen, 2020).If consumers feel that the offers from the website are more trustworthy and reliable, a favourable perspective toward online shopping will be developed; this may influence their propensity to use internet shopping (Kwaku & Antwi, 2021).
Perceived web quality impacts consumer attitudes towards website use (Castro-Ochoa & Moreta-Herrera, 2023).Web quality is an essential factor influencing customer purchase intentions on social commerce websites (Cassandra et al., 2017).This is reinforced by a study that found that in mobile e-commerce, website quality impacts customer happiness and intent to buy (Chi, 2018).

H1:
Perceived web quality has a positive and significant effect on attitude.

H2:
Perceived web quality has a positive and significant effect on perceived behavioural control.
H3: Perceived web quality positively and significantly affects purchase intention.

Self-Efficacy, Attitude, Perceived Behavioral Control and Purchase Behavior
Intention.Efficacy is a belief in the ability to perform tasks well and a sense of competence, efficiency, and coping with life (Waddington, 2023).There is a correlation between how one feels about one's abilities and how one feels about using technologybased self-directed learning (Pan, 2020).The two variables of perceived control and intention to purchase are connected via the mediation of self-efficacy, which acts as a mediator.Perceived control behaviour directly and indirectly influences buy intention via self-efficacy (Y.Li et al., 2018).When customers get helpful customer service when purchasing online, their perceived control over the action grows, their self-efficacy rises, and their purchase intention becomes stronger (Al-Gasawneh & Al-Adamat, 2020; Y. Li et al., 2018;Shufiana et al., 2021).H4: Self-efficacy has a positive and significant effect on attitude.H5: Self-efficacy has a positive and significant effect on perceived behavioural control.H6: Self-efficacy has a positive and significant effect on purchase behaviour intention.

Attitude, Subjective Norms, Perceived Behavioral Control and Purchase
Behavior Intention.Several factors influence consumers' behavioural attitudes, including individual characteristics, user behaviour, user experience, user perceptions, and other external factors (Wilnei Aldir & Rafael, 2021).Attitudes towards behavioural intention show the most significant positive correlation (Wang et al., 2023).This is reinforced by consumer perceptions of the benefits of e-commerce, such as ease of searching, which increases consumers' intention to use e-commerce (Peña-García et al., 2020).
Consumers will be more likely to utilise e-commerce if they hear good things about it from people they know, whether via word of mouth, personal experiences, or objective standards (Pradeep et al., 2021).According to earlier studies, perceived social standards are a robust precursor to future conduct (Sharma et al., 2021).When people talk about online shopping, what others think of it or whether they believe it is suitable all fall under subjective norms (Dashti et al., 2019).Personal beliefs and values significantly and favourably influence consumers' propensity to purchase (Suparno, 2017).
When people discuss the ease or difficulty of engaging in a behaviour that captures their interest, they express their perceived behavioural control.This is reinforced by the results that customers' perceptions of their behavioural control have a substantial and favourable impact on their intentions to engage in direct e-commerce (Zhang & Chen, 2023).The order of importance for predicting future conduct is as follows: attitudes, perceived behavioural control, and subjective norms (Barbera & Ajzen, 2020).An individual's propensity to engage in e-commerce marketing is positively correlated with their level of behavioural control (Darsono et al., 2020).
H7: Attitude has a positive and significant effect on purchase behaviour intention.
H8: Subjective norms positively and significantly affect purchase behaviour intention.
H9: Perceived behavioural control positively and significantly affects purchase behaviour intention.
Attitude on Purchase Behavior Intention and its Mediation Role.Attitude towards online shopping can mediate the positive relationship between website quality and online purchase intention (Kwaku & Antwi, 2021).A positive attitude towards online shopping can lead to a higher intention to make purchases, while a negative attitude can result in a lower intention.Self-efficacy relates to an individual's degree of assurance in their capacity to carry out a particular behaviour, like online shopping (Y.Li et al., 2018).High self-efficacy can lead to a more positive attitude towards online shopping, which in turn can influence purchase behaviour intention.H10: Attitude mediates perceived web quality and purchase behaviour intention.H11: Attitude acts as a mediator between self-efficacy and purchase behaviour intention.

She perceived Behavioral Control on Purchase Behavior Intention and its
Mediation Role.Perceived web quality can impact consumer attitudes and purchase intention, with trust and perceived risk mediating and moderating roles in this relationship (Peña-García et al., 2020).Perceived web quality can indirectly influence purchase behaviour intention through its impact on perceived behavioural control.Self-efficacy can influence purchase behaviour intention through its impact on perceived behavioural control, which is affected by factors such as experience, modelling, expected support, and potential obstacles (Khoa, 2023;Y. Li et al., 2018).High self-efficacy can lead to higher perceived control, which can induce stronger consumer purchase intention.

METHODS
Study Design and Sample.October and November 2023 were the months in which this research was carried out.The study's methodology was a cross-sectional survey, which allowed for quantitative analysis.In this investigation, essential random sampling was used.The participants in this research were the 1,808 Regional Government Commitment-Making Officers (CMOs) who made electronic purchases of laptop items using the ecatalogue during the fiscal year of 2022.Five to ten observations are needed for each latent variable (Hair et al., 2017).Consequently, there were 18 indicators in this research, and the sample size ranged from 90 to 180 participants.One hundred thirty-five eligible participants completed an online survey using Google Docs.
Data Collecting and Measurement.This research used an online questionnaire collected via Google Forms.The questionnaire contained structured questions for CMOs who have made e-purchasing transactions for laptop products in the e-catalogue.Closedended questions were included in the questionnaire, and respondents were allowed to react using a Likert scale ranging from 1 to 5, where five denoted strong agreement and one denoted strong disagreement.The number of indicators that make up the model in this study is 18 indicators, which is attached in Table 2.Primary data for this research came from online surveys shared via social media.In addition, SEM was used in PLS software to examine the hypothesis-based model.Self-efficacy is an individual's belief about their ability to carry out a particular behaviour.Data Analysis.This study's data processing and analysis will use descriptive analysis techniques, crosstabulation, the top and bottom two boxes, and SEM-PLS.The quantitative descriptive analysis relies on mathematical and statistical analysis of numerical data to describe the study topic using variable data collected from a specific set of individuals.An analytical approach examined using the chi-square value is crosstabulation, also known as crosstabulation, which is used to find the link between influential variables and impacted factors.
To display the proportion of responses, respondent distribution analysis sorts respondents' replies into two groups: top and bottom.The Top Two Boxes are the percentage of answers that agree and strongly agree, whereas the Bottom Two Boxes are the percentage of answers that disagree and strongly disagree.When explaining the link between many variables, structural equation modelling (SEM) is a useful statistical tool for building and testing causal models.

RESULTS
The respondents involved in this study were 79.300 per cent male, and the rest were 20.700 per cent female.The age distribution of respondents ranged from 25 years to more than 40 years.In detail, respondents aged 25 to 30 reached 2.200 per cent or three respondents.Respondents aged 31 to 35 reached 4.400 per cent or as many as four respondents.Respondents aged 36 to 40 years reached 38.500 per cent or 52 respondents.Most respondents were over 40 years old (54.800 per cent) or 74 respondents.Most respondents came from agencies in the Sumatra region (39.300 per cent) or 53 people.The second largest domicile of respondents was in Java & DIY, which reached 35.600 per cent or 48 respondents.Then, 11.100 per cent of respondents came from agencies in the Kalimantan region, reaching 15 people.8.900 per cent of respondents came from agencies in the Sulawesi region and 3 per cent from Bali, reaching 12 and 4 people, respectively.Meanwhile, the most diminutive agency origin was in NTB, with 1.500 per cent of respondents or two people.Then, NTT had 0.700 per cent or one respondent.Respondents who served as CMO for 1 to 3 years reached 65 per cent or 88 people.The second largest number of CMOs for more than three years reached 33 per cent or 44 respondents.Then, 2 per cent of respondents or three people, have only served as CMO for less than one year.As many as 48 per cent or 65 respondents have the last level of education, S2.Meanwhile, Outer model evaluation is the initial stage in research testing that applies the Partial Least Square-Structural Equation Modeling (PLS-SEM) method.This evaluation aims to determine the relationship between constructs and indicator variables through validity and reliability tests, as shown in Figure 2. Indicator value measurements are valid if they have a reasonable correlation greater than 0.700 in value.Elimination or dropping occurred when an indicator's loading factor value was less than 0.700.It is feasible to evaluate convergent validity by using either the loading factor value or the average variance extracted (AVE) value.Both of these methods are appropriate.A value of the AVE larger than 0.500 was used to assess whether or not the latent variables were valid.By calculating the composite dependability score, we could gauge the indicators' degree of consistency.If the composite reliability value is more than 0.700, then the variable is considered dependable.
Using Cronbach's alpha, one can evaluate the consistency level inside a model.When the variable has Cronbach's alpha value more than 0.600, we consider that variable to be reliable or consistent.Based on Table 4, It can be inferred that each variable is deemed reliable as their Cronbach's alpha values are more than 0.600.The study model is of high quality based on the results of the outer model test.This is because the variables and indicators are reliable and accurate.Table 4 shows that the reflective measurement model evaluation results are consistent and correlated.Common method bias may arise in research when independent and dependent variables are gauged using the identical response method within a single survey.Harman's Single Factor Test is useful for detecting common method variance in CFA and can help researchers find whether the data contain common bias.To prevent common method bias within this study, it is advisable to ensure that no individual factor accounts for more than 50 per cent of the variance (Chang et al., 2020).According to the findings in Table 5, all variance values are calculated to be less than 50 per cent.Hence, there were no concerns regarding common method bias in this study.PLS-SEM Analysis.This study used SmartPLS version 4 software as a data analysis tool with 135 respondents.We used structural equation modelling (SEM) using partial least squares (PLS) to establish the connection between the two data sets.Research testing begins with an assessment of the external model.This investigation's primary objective was to identify the components that would function as indicators of the variables that were being studied.Furthermore, it seeks to determine the nature of the relationship between constructs and indicator variables by examining the reliability and validity of the data.In addition to the model, we will conduct tests to determine convergent and discriminant validity, composite reliability, and Cronbach's alpha.
The indicator value measurement is valid if the correlation value exceeds 0.700.Elimination or dropping occurred when an indicator's loading factor value was less than 0.700.The convergent validity assessment may be seen from the loading factor value and the Average Variance Extracted (AVE) value.The latent variables were deemed legitimate if the AVE value was higher than 0.500.By calculating the composite dependability score, we could gauge the indicators' degree of consistency.If the composite reliability value is more than 0.700, then the variable is considered dependable.Consistent and associated findings from examining the reflecting measurement model are shown in Table 6.As a component of the discriminant validity test, we examine how each variable correlates with the other independent variables.In the Fornell-Lacker model, in order for a variable to be highly correlated, the internal correlation of the variable must be more significant than the correlations that it has with other variables.The outcomes of the examination aimed at determining the discriminant validity are shown in Table 6.Using the cross-loading test, one may determine whether the variables have a discriminant validity.To do this, a comparison is made between the construct indicators' correlation coefficient values and those of other constructs.Indicator variables are said to have strong discriminant validity when the correlation coefficient of the indicator variable is higher than the correlation coefficient of the indicator variable with other variables.Each of the cross-loading data is shown in Table 7. Structural model evaluation is used to predict the quality relationship between latent variables or variables that cannot be measured directly and that have been built based on substance and theory.This test uses the bootstrapping procedure with SEM PLS.One of the test stages for the structural model is to see the R-squared value on the estimated path coefficient.The R-Square value is the coefficient of determination on endogenous constructs and is used to see the predictive power of the inner model.The criteria range of the R-Square value is 0.670 (strong), 0.330 (moderate) and 0.190 (weak).The R-squared value obtained in this study is shown in Table 8.The inner model was then evaluated as a component of the research testing procedure.This model evaluation aims to predict the quality of the connection between latent variables, founded in theory and substance, but cannot be evaluated directly.In the process of testing the structural model, many phases are taken.One of these procedures is determining the R-squared value in the route coefficient estimate using a bootstrapping approach with SEM PLS (Sekaran & Bougie, 2016).The coefficient of determination of endogenous components is often called the R-squared value.This number serves as a measurement of the predictive power of interior models.R-Square values that fall between the range of 0.670 (strong), 0.330 (moderate), and 0.190 (weak) are regarded to be the minimum acceptable.The results shown in Table 6 demonstrate that the R-squared values of all three variables ATT, PCB, and PBI are more than 0.570.If all of the variables are moderate, then it is possible to explain 57.400 per cent of the variation in the ATT variables, 57.100 per cent of the variance in the PCB variables, and 65.800 per cent in the PBI variables.The remaining is accounted for by other variables not accounted for in this model.The model fit indicators in Table 9, such as the RMS Theta value of 0.194 (greater than 0.102) and the NFI value of 0.693 (less than 0.900), suggest that the model does not meet the fit criteria based on these assessments.However, the SRMR value of 0.078 (less than 0.100) indicates a good fit for the model.Consequently, the model fits well with the provided data.

Goodness-of-
Hypothesis Testing Results.We tested our hypothesis using the path coefficient value derived from the Smart PLS version 4 bootstrapping results.In addition, the beginning sample value, the T-statistic, and the P-value are all returned via bootstrapping.If the path coefficient is larger than or equal to 0.050 and the t-value is greater than or equal to 1.979, then the factors substantially influence the study's outcome.In addition, the results of the hypothesis testing may be seen in Figure 2 and Table 10 via the use of the P-value, which utilises a significant level of 5 per cent.This indicates that the research hypothesis is accepted if the P-value is lower than 0.050.The data presented in the table shows a significant positive correlation between perceived web quality (H1) and attitude (t of 7.598; p of 0.000; path of 0.555).Additionally, self-efficacy has a substantial positive direct effect on attitude (t of 3.451; p of 0.001; path of 0.270).Therefore, both hypotheses are supported.This implies that the results support our hypothesis, as expected.Simultaneously, perceived web quality (H3) and self-efficacy (H4) significantly influence perceived behavioural control in epurchasing through e-catalogue.As expected, We show that self-efficacy (t of 4.853; p of 0.000; path of 0.451) and perceived web quality (t of 3.678; p of 0.000) positively impact perceived behavioural control.This finding lends credence to both hypotheses.There was no significant correlation between the perceived quality of the website and the desire to acquire an electronic catalogue (H5).This requirement is brought about by the t-value being less than 1.979, which is the case even if the route coefficient value is close to 0.040.As a consequence, this result does not provide evidence for the hypothesis.
Both hypotheses are supported by the data: attitude (H6) has a significant and positive direct effect on purchase behaviour intention (t of 3.262; p of 0.001; path of 0.283), and subjective norms (H7) have a direct positive impact on trust (t of 2.77; p of 0.006; path of 0.182).The data support both hypotheses.This implies that the results support our hypothesis, as expected.Simultaneously, perceived behavioural control (H8) and self-efficacy (H9) significantly influenced purchase behaviour intention in the ecatalogue.As expected, Both hypotheses are supported by the findings, which demonstrate that self-efficacy has a positive influence on purchase behaviour intention (t of 2.821; p of 0.005; the path of 0.281) and that perceived behavioural control has a positive influence on purchase behaviour intention (t of 2.028; p of 0.043; the path of 0.192).As an additional point of interest, it was discovered that attitude had a more substantial influence on the intention to purchase than any of the other elements.
Perceived web quality (H10) on purchase behaviour intention through attitude as a moderator is significant and has a positive direct effect (t of 2.925; p of 0.004; the path of 0.157), and self-efficacy (H11) on purchase behaviour intention through attitude as a moderator has a direct positive impact (t of 2.319; p of 0.021; the path of 0.076), supporting both hypotheses.This implies that the results support our hypothesis, as expected.The relationship between perceived web quality and purchase intention in the e-catalogue through perceived behavioural control as a moderator (H12) was insignificant.This condition is caused by the p-value, which is more than 0.050, even though the path coefficient value reaches 0.074; therefore, this result does not support the hypothesis.Simultaneously, self-efficacy (H13) positively impacts and significantly influences purchase behaviour intention in the e-catalogue through perceived behavioural control as a moderator (t of 2.038; p of 0.042; the path of 0.870), supporting both hypotheses.In addition, this research shows that attitude exerts a more significant impact on purchase behaviour intention than the other variables.

DISCUSSION
The digital transformation of government through e-catalogue has significantly changed how procurement parties carry out the procurement process.The e-catalogue application offers information in the form of lists, types, technical specifications, Domestic Component Level (TKDN), domestic products, Indonesian National Standard (SNI) products, environmentally friendly products, country of origin, prices, providers, and other information related to goods/services.Given the importance of digital transformation of the procurement process by Presidential Instruction Number 2 of 2022, the researchers tried to understand the factors that drive the purchasing behaviour intentions of Commitment Making Officers (CMOs) with the Theory of Planned Behavior approach.This research primarily aims to provide empirical evidence on the model of purchasing behaviour through an electronic catalogue with the approach of the Theory of Planned Behavior and perceived web quality and self-efficacy as a complement to these relationships.
Based on the results of the testing, direct and indirect relationships show that several factors influence the purchase behaviour intention of CMOs when buying laptop products in the e-catalogue.From this, perceived web quality and self-efficacy positively affect the attitude and perceived behaviour control of CMOs.The existence of better attitudes, subjective norms, and perceived behaviour control from CMOs will significantly increase purchase behaviour and intention to buy laptop products in the e-catalogue.This is inversely proportional to the effect of perceived web quality, which has no positive or insignificant effect on purchase behaviour intention, so the purchase behaviour intention of CMOs is not directly affected by e-catalogue web quality.
Perceptions of a website's quality affect how CMOs feel about using e-catalogue, especially the precise arrangement of the e-catalogue web display menu.An organised menu on the website can make it easier for CMOs to find product information and buy it.CMOs feel that the website's ease of use and security are more trustworthy and reliable, so a positive attitude towards online shopping will be formed.This could affect the likelihood of purchasing on a website.The importance of focusing on the quality aspects perceived by users as they browse the website will affect perceived behavioural control.The perceived quality of the e-catalogue website that users feel is better will foster the ability to overcome difficulties in conducting e-purchasing transactions.Previously, most research supported theoretical predictions regarding the relationship between perceived website quality and attitude and perceived behavioural control, which then influenced purchase behavioural intention (Baganzi et al., 2019;Barlett, 2019;Kwaku & Antwi, 2021;Rita et al., 2019).
On the other hand, the research results show that perceived website quality cannot directly influence purchasing behaviour intentions.The research results explain that epurchasing via e-catalogue is an alternative procurement method, so the quality of the ecatalogue website does not directly encourage purchasing behaviour intentions for CMOs.The perception of good website quality will encourage purchasing behaviour if there is a feeling of trust in the e-catalogue.This is confirmed by previous research that the quality and reputation of a website have a significant positive impact on purchase intention, and trust plays a mediating role in this relationship (Qalati et al., 2021).
A high level of self-efficacy related to the ability and compliance possessed by CMOs will enable them to make purchases through e-catalogue.This belief arises from previous experience, level of commitment, and confidence in dealing with technology.Individual perceptions of the use of e-catalogues, including the advantages and benefits, comfort, trust, and so on, will form a positive attitude towards the use of e-catalogues.The research results show that confidence in using the e-catalogue will tend to have a more positive attitude towards using the platform.Confidence in your ability to interact with technology will influence attitudes towards using e-catalogues.Individuals' confidence level and attitudes towards the e-catalogue platform will influence their decision to purchase through it.The research results show that confidence in using the e-catalogue effectively will help strengthen confidence in the control one has or significantly influence purchasing behaviour intentions.Thus, a positive relationship is formed between selfefficacy towards attitude and perceived behavioural control, influencing purchase behavioural intention (L.Li et al., 2022;Y. Li et al., 2018;Pan, 2020).
Regarding self-efficacy, CMOs have high confidence in existing policies and regulations.To maintain and increase CMO's self-efficacy, the government must make clear and firm policies related to government procurement of goods/services using the epurchasing method through e-catalogues.The more precise and firmer the existing regulations so that e-purchasing transactions through e-catalogues are maximised, the more regulations can be made on the imposition of sanctions on the realisation of procurement spending through e-catalogues that do not meet the target.
The results of this research show that attitude influences purchase behaviour intention positively and significantly.A positive attitude towards the use of e-catalogues can motivate CMOs to intend to shop using the platform.The formation of a positive attitude from CMOs can be influenced by feelings of pleasure in online transactions, ease of use, and perceptions of the best solution ideas.The better the perception of ease of use, the greater the intention to purchase via e-catalogue because they feel the process will be smooth and hassle-free.The results of this research provide additional insight into attitude as a factor influencing purchase behaviour intention (Peña-García et al., 2020;Wang et al., 2023).In other words, it can be said that a positive attitude towards e-catalogue will encourage CMOs to have a higher intention to use the platform in the purchasing process.The stronger the attitude of CMOs, the greater the influence on their behaviour.Therefore, this can be an evaluation material for the government to increase the convenience for providers in carrying out purchasing transactions via e-catalogue through clearly formulated regulations.
Subjective norms play an essential role in forming purchase intentions through ecatalogues by integrating social influences and behavioural norms from an individual's social environment.This research shows that subjective norms influence purchase behaviour intention positively and significantly.Subjective norms reflect the extent to which individuals feel others, such as leaders, friends, family, or other social references, expect or support them to purchase via e-catalogue.Thus, the stronger the social pressure CMOs feel to use e-catalogues, the greater their intention to purchase through the platform.Apart from that, the existence of applicable regulations will shape purchasing decisions.In other words, using an e-catalogue is a procurement regulation that must be implemented, so the intention to follow this regulation will arise by making purchases.The results of this research explain that individual perceptions regarding direction or support from leadership figures regarding the use of e-catalogues can influence CMOs' intentions to carry out epurchasing through e-catalogues (Dashti et al., 2019;Liu et al., 2023;Sharma et al., 2021).Regarding managerial implications related to subjective norms, the government, through the leadership (Budget Users/Authorised Budget Users), should be able to provide direction to procurement parties to ensure compliance with government policies relating to the procurement of e-catalogue.
Current research proves that perceived behavioural control influences purchase behaviour intention positively and significantly.Perceived behavioural control is vital in shaping purchase behaviour intentions through e-catalogues.The high level of control that CMOs have over the purchasing process via e-catalogue gives rise to a higher intention to use the platform for shopping.Confidence in managing the purchasing process effectively through e-catalogue will increase CMOs' confidence in carrying out e-purchasing transactions.The level of perceived behavioural control can be influenced by knowledge, ability to operate, and ability to overcome difficulties.This perception of capability can increase CMOs' intention to use e-catalogues for shopping (Darsono et al., 2020;X. Li et al., 2023;Zhang & Chen, 2023).Therefore, the beliefs held about the ability to control the purchasing process and perceptions about the ease of use and usefulness of the e-catalogue platform will form the intention to use the platform for shopping.In this case, the government should provide massive training for procurement parties to increase the capability to use the e-catalogue application.Training programs are critical in updating CMOs' knowledge so that they can become more confident in shopping via e-catalogue.
Government support is needed to stimulate procurement actors to transact through e-catalogue.Rewards and punishments can be applied to maximise transactions through ecatalogue.With rewards, leaders and procurement actors will be motivated to increase purchasing transactions through e-catalogue.Meanwhile, the existence of punishment will provide an evaluation for each agency that has yet to be able to maximise the value of its purchasing transactions through e-catalogue.This research shows that CMOs have behavioural intentions to transact online (e-purchasing) via e-catalogue in the future.The e-catalogue can be an effective shopping solution for procuring goods and services for government procurement parties to help save time, costs and resources while monitoring the procurement process in real time.

CONCLUSION
According to the research findings, self-efficacy and perceived online quality are two factors that substantially influence perceived behavioural control.Individuals' perceptions of the quality of the web had a favourable and substantial influence on their perceived behavioural control, and self-efficacy also had a favourable influence on the perceived level of behavioural control.
When it comes to a person's behavioural intention to purchase, the essential components that influence that intention are their attitude, perceived behavioural control, self-efficacy, and subjective norms.The degree to which customers believe they can accomplish their goals positively and substantially impacts the chance they will purchase.A positive and statistically significant correlation exists between a person's disposition and their inclination to make purchases or purchases.On top of that, there is a positive and statistically significant correlation between subjective norms and the individual's incentive to make a purchase.An individual's impression of their behavioural control was also shown to have a favourable and substantial impact on their desire to make a purchase transaction.It was not found that the perceived quality of the website had a favourable and substantial impact on the intention to purchase.
The purchasing behaviour model in this study shows that e-purchasing behaviour in e-catalogues is influenced by the perceived web quality of e-catalogues that are clear in appearance and safe to use for transactions.High self-efficacy support for existing policies or regulations also affects purchasing behaviour.The digital transformation of government procurement of goods and services must be supported by formulating the right procurement strategy to encourage local governments to increase the value of e-purchasing spending.In this process, attention can be paid to perceived web quality, self-efficacy, attitude, subjective norms, and perceived behavioural control of procurement actors, ultimately increasing the value of e-purchasing transactions in the catalogue.An alternative strategy to increase purchase behaviour intention in e-catalogue is to improve the quality of the e-catalogue website so that it is easy, safe, and comfortable to use.Another alternative is formulating clear and firm regulations for implementing epurchasing through e-catalogues.
The results of this study are expected to be helpful for the government as a policymaker.The purchase intention of the Commitment Making Officers (CMO) to make purchases through e-catalogue is essential for the government.The realisation of epurchasing through e-catalogue requires the role of CMOs as the executor.E-catalogue as a government purchasing platform provides relative advantages not obtained through other procurement methods, which is believed to impact increasing expenditure realisation.The government should regularly socialise e-catalogue regulations to build robust confidence and compliance for procurement actors in online shopping transactions through ecatalogue.In addition, the government should be able to increase the ease of procurement actors conducting purchasing transactions through e-catalogue through clear regulations.
Furthermore, clear communication, coordination, and guidance from the leadership are considered good in fostering the purchase intention of the CMOs.Therefore, through leaders (Budget Users / Authorized Budget Users), the government should be able to provide direction to actors to ensure compliance with government policies relating to procuring e-catalogue goods and services.Next, the government should be able to provide massive training for procurement actors to increase their capability to use the e-catalogue application.
Limitations and suggestions.This study has limitations because it only examines the purchasing behaviour of e-purchasing transactions through e-catalogue.In addition, the e-catalogue website is currently only used in the government sector, ministries, institutions, and local governments.The questionnaire distribution was conducted only by the local government.The questionnaire was filled out using Google Forms without any supervision from the author so that misinterpretation can occur, which makes the data biased.Furthermore, we can consider other variables influencing purchasing behaviour in e-catalogue, such as trust, perceived usefulness, and perceived risk.Adding these variables is expected to provide a broader view of purchasing behaviour.

H12:
Perceived behavioural control mediates perceived web quality and purchase behaviour intention.H13: Perceived behavioural control mediates self-efficacy and purchase behaviour intention.This study uses a combination of factors represented by perceived web quality (PWQ) and self-efficacy (SEF) about attitude (ATT) and perceived behavioural control (PBC).In addition, perceived web quality (PWQ), self-efficacy (SEF), attitude (ATT), subjective norms (SBN), and perceived behavioural control (PBC) are also represented by purchase behaviour intention (PBI), with the research model shown in Figure 1.

Figure 1 .
Figure 1.Conceptual Framework feel happy to transact e-purchasing online through e-catalogue.2. E-purchasing through e-catalogue is a good idea.3. I feel that buying through ecatalogue is a good thing people in my work environment influence my behaviour in thinking about e-purchasing through ecatalogues.2. A leadership figure's direction influences my thinking when doing e-purchasing through e-catalogue.3. The latest rules/regulations I read influence my thinking about epurchasing through e-catalogue.Perceived Behavioural Control (PBC) Perceived Behavioural Control (PBC) is a critical component of the Theory of Planned Behaviour) that predicts an individual's ability to perform a particular behaviour, which is assumed to moderate the influence of attitude and subjective PBC1 PBC2 1.I have sufficient knowledge to conduct online transactions (epurchasing) in e-catalogue.2. I can operate the e-catalogue website for online transactions (epurchasing).norms on intention and actual behavioural control.PBC3 3. I have overcome difficulties when conducting online transactions (epurchasing) through e-catalogue.Purchase Behavioural Intention (PBI) Purchase intention is the extent to which consumers are willing to buy a product.It applies to people who have yet to buy a product and those who already have experience with it.intend to transact online (epurchasing) in the e-catalogue.2. In the future, I intend to prioritise online transactions (e-purchasing) in the e-catalogue.3. I will use e-catalogue as an effective procurement shopping solution Source: Perceived web quality (Kwaku & Antwi, 2021); self-efficacy (Li et al., 2022), attitudes (Castro-Ochoa & Moreta-Herrera, 2023); subjective norms (Castro-Ochoa & Moreta-Herrera, 2023); perceived behavioural control (Castro-Ochoa & Moreta-Herrera, 2023); and purchase behaviour intention (Peña-García et al., 2020).

Table 2 .
Variables and Indicator

Table 4 .
Validity and reliability test results

Table 5 .
Common Method Biases test results

Table 6 .
Discriminant validity test results

Table 7 .
Cross-loading test results

Table 8 .
R Square testing results