INTRODUCTION:

Vietnam’s chemical industry plays an important role in the national economy but remains less competitive compared to regional and global markets. With approximately 1,818 enterprises and 2.7 million workers (2020), the industry faces challenges such as low international market share (0.5%), limited technological capacity, and low public awareness of chemical risks, which can lead to environmental and health concerns. In particular, the agricultural chemical sector struggles to compete with regional producers, underscoring the need for improved product quality, safety, and technological innovation.

Purchase intention, defined as a customer’s willingness and perceived ability to purchase a product or service, can be influenced by various factors such as perceived value, celebrity endorsement, packaging, and customer knowledge¹. In the chemical sector - where products can pose potential risks - customers tend to prioritize reputable suppliers offering safe and high-quality products. However, research on purchase intention in Vietnam has mostly focused on industries like fashion and technology, with limited studies targeting chemical companies.

This study aims to fill this gap by identifying and analyzing factors influencing customers’ purchase intention at Hoang Vu Chemical Trading Co., Ltd. A research model was developed based on international studies and tested through a survey of customers in Da Nang. The results will provide practical managerial implications to enhance purchase intention and improve business performance in the chemical trading sector.

LITERATURE REVIEW:

Purchase intention

Purchase intention (PI), understood as a consumer’s likelihood or decision to buy a specific product or brand², is widely conceptualized as a central construct linking consumers’ attitudinal evaluations with actual purchasing behavior. It functions not only as a predictor of future sales but also as a diagnostic indicator of marketing program effectiveness³. In the digital context, Online Purchase Intention (OPI) further extends this theoretical role by emphasizing how technology-mediated environments shape intention formation⁴. Consistent with this theoretical framing, Bais⁵ reinforces that PI is influenced by an interplay of demographic, psychological, and contextual factors rather than by isolated stimuli.

Within this broader framework, demographic variables are theorized as structural antecedents that segment consumer responses. Evidence from Praveer and Shrivastava⁶, as well as Devasenathipathi and Mark⁷, demonstrates that demographic profiles shape product evaluations, satisfaction, and repeat purchase likelihood, highlighting how demographic segmentation aligns with consumer decision-making models. Psychological determinants, particularly consumer perceptions and value orientations, form a second major stream of theoretical influence. Studies by Doniavi and Pourfatemi⁸ and Pani and Pradhan⁹ show that perceived value, authenticity, cultural meaning, and customer orientation drive intention formation, reinforcing expectancy-value and perception-based theories. Marketing communication also functions as a stimulus shaping consumer attitudes and intentions. Research by Kumar and Sharma¹⁰ and Kumar and Asok Kumar¹¹ positions advertising, pester power, and digital promotional strategies as external cues that affect consumers’ cognitive responses, consistent with stimulus–organism–response (SOR) models.

Parallel to these factors, digital interaction variables—particularly electronic word-of-mouth (eWOM) and online trust—represent a distinct theoretical domain in contemporary PI research. Singh and Singh¹² highlight how website reliability, product variety, and digital trust shape online purchase decisions, illustrating the theoretical shift toward risk–trust frameworks in e-commerce. Contextual and behavioral outcomes, such as satisfaction and product performance, have also been linked to PI through confirmation and value–satisfaction theories. Chowdhury¹³ shows that quality, price fairness, and after-sales service enhance satisfaction, thereby strengthening future purchase intentions. Research on switching behavior further expands this understanding by demonstrating how comparative evaluations drive intention formation. Jha, Deshmukh, and Joseph¹⁴ reveal that attitudes, perceived service quality, and trust shape switching intention in mobile number portability, illustrating dynamic intention processes in competitive markets.

Methodologically, PI is commonly measured using Likert scales encompassing transactional, preferential, referential, and explorative dimensions¹⁵. However, PI remains an imperfect predictor of actual behavior¹⁶ due to intention–behavior gaps influenced by situational or personal changes³. Theoretical advancements in OPI highlight the central roles of psychological constructs such as trust and perceived risk, where trust enhances intention and perceived risk inhibits it³. Product quality, advertising, brand image, brand awareness, brand equity^2,17, and perceived value¹⁸ operate as key cognitive antecedents. Brand personality¹⁹—defined as human traits associated with brands—also shapes PI, with evidence from Mexico showing that specific BP dimensions significantly predict consumers’ willingness to purchase¹⁹. Together, these streams of literature underscore that PI is best understood through an integrated theoretical lens combining demographic segmentation, psychological evaluation, communication stimuli, digital trust dynamics, and satisfaction-related mechanisms.

PURCHASE BEHAVIOR:

Low et al. define customer behavior as the process by which customers make decisions regarding the selection and disposal of products or services²⁰. Consumer behavior encompasses a sequence of decisions about what, why, when, how, where, how much, and how often individuals or groups purchase. It involves the activities of searching for, purchasing, using, and evaluating products and services that are expected to meet personal needs²¹.

Achrol and Kotler view consumer behavior as the specific actions individuals take when deciding to purchase, use, or dispose of products or services⁵. Similarly, Lourenco conceptualizes it as the dynamic interaction of cognitive, behavioral, and environmental factors through which individuals adapt their lives²².

In marketing, the study of consumer behavior enables firms to identify consumer needs, preferences, and habits—understanding what products they purchase, the reasons for those purchases, brand choices, purchase methods, timing, locations, and quantities. Behavior includes human actions in purchasing and using products, alongside the psychological and social processes before, during, and after purchase. He identifies four major groups of factors influencing consumer behavior: cultural, social, personal, and psychological. Consumer behavior refers to the actions and decision-making processes individuals undertake when selecting, using, or discarding products or services.

RESEARCH MODEL AND HYPOTHESES:

Based on the theoretical foundations of the Theory of Reasoned Action (TRA), the Theory of Planned Behavior (TPB), and the Technology Acceptance Model (TAM), combined with elements inherited from the related research articles analyzed above, the author proposes a research model to examine factors influencing customers’ purchase intention at Hoang Vu Chemical Trading Co., Ltd. The model includes four variables: customer knowledge, perceived value, celebrity endorsement, and product packaging.

H1: Customer knowledge has a positive impact on perceived value

H2: Customer knowledge has a positive impact on purchase intention

H3: Celebrities have a positive impact on perceived value

H4: Celebrities have a positive impact on purchase intention

H5: Perceived value has a positive impact on purchase intention

H6: Products packaging has a positive impact on purchase intention

Figure 1: The study’s Proposed Theoretical Framework

RESEARCH METHODOLOGY:

A total of 350 valid responses will be collected in Da Nang via an online survey (Google Forms) distributed through social media platforms such as Facebook and Zalo. The dataset will be analyzed using SPSS 26.0, employing descriptive statistics, reliability assessment via Cronbach’s Alpha, exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and structural equation modeling (SEM). Additionally, mean differences will be examined using the Independent Samples T-test and One-Way ANOVA. The findings will inform conclusions and objective managerial implications aimed at developing effective strategies to enhance customers’ purchase intentions. The quantitative analysis methods that will be used for the data analysis part include:

Testing the reliability of the scale by Cronbach's Alpha: This test reflects the degree of correlation between the observed variables in the same factor. The standard to test the reliability of the scale is that the measurement variables have the total correlation coefficient of Corrected Item - Total Correlation ≥ 0.3, then the variable meets the requirements.

Exploratory Factor Analysis (EFA): This method helps to evaluate two important types of values of the scale: convergent value and discriminant value. The condition for exploratory factor analysis is to satisfy the following requirements: Factor loading > 0.5; KMO (Kaiser-Meyer-Olkin) in the range 0.5 ≤ KMO ≤ 1, Bartlett test has statistical significance (Sig. < 0.05).

Confirmatory factor analysis (CFA): is one of the techniques that allow testing how well the measured variables represent the factors. The CFA method is used to confirm the univariate, multivariate, convergent, and discriminant validity of the factor scale.

Linear Structural Model (SEM): This model is used to analyze complex relationships in causal models. SEM is used to estimate the measurement models and the structural model of the multivariable theory problem. The criteria to evaluate the overall goodness of fit when analyzing actual data are the Chi-square test with p > 0.05, TLI and CFI values from 0.8 or more, Chi-square/df < 3, RMSEA < 0.08.

RESEARCH FINDINGS AND RESULTS:

Testing the Reliability of Scale by Cronbach's Alpha

After assessing the Cronbach’s Alpha coefficients for all scales in the research model, the results (Table 1) indicate that five scale groups demonstrated good reliability, encompassing 22 observed variables. All scales and observed variables met the reliability requirements, and no factors or items were removed during the analysis.

Table 1: Cronbach's Alpha test results

Status	The scale	Number of observed variables	Number of response variables	Cronbach's Alpha coefficient
1	Perceived value	4	4	0,935
2	Celebrity	5	5	0,903
3	Products Packaging	5	5	0,958
4	Customer knowledge	4	4	0,956
5	Purchase intention	4	4	0,907

Source: Processing results from survey data, 2024

Exploratory factor analysis (EFA)

The results of EFA analysis are shown in Table 2 with the KMO coefficient = 0.923; the Bartlett test value is significant (Sig < 0.05), all observed variables have factor loading coefficients greater than 0.5 so no variables are excluded.

Table 2: EFA results of the scale

Variable	Factor
Variable	1	2	3	4	5
BB5	0,897
BB3	0,888
BB4	0,883
BB2	0,869
BB1	0,866
NT3		0,846
NT2		0,835
NT1		0,817
NT4		0,816
NT5		0,774
KT4			0,917
KT2			0,892
KT3			0,875
KT1			0,869
CN4				0,933
CN3				0,931
CN2				0,917
CN1				0,881
YD3					0,921
YD2					0,890
YD1					0,869
YD4					0,867

BB = Products Packing, NT = Celebrity, KT = Customer Knowledge, CN = Perceived Value, YD = Purchase Intention

Source: Processing results from survey data, 2024

Confirmatory Factor Analysis (CFA):

After exploratory factor analysis, the author adjusted the scale and performed confirmatory factor analysis (CFA) to check the scale again. Confirmatory factor analysis was used to retest the measurement model that we found from the EFA analysis. After analyzing the CFA, we obtained indicators to assess the fit of the measurement model with the actual data: Chi-square index=372,077, df=199, Chi-square/df=1,870, P=0,000, GFI index=0,926 (≥ 0,9 is good), CFI index= 0,981 (≥ 0,95 is very good), PCLOSE index=0,860 (≥ 0,01 is acceptable), TLI index= 0,978 (≥ 0,9 is good) and RMSEA=0,045 reach relatively low values (≤ 0,06 is good). Overall, these indices show that this measurement model has a relatively high level of measurement fit.

BB = Products Packing, NT = Celebrity, KT = Customer Knowledge, CN = Perceived Value

Figure 2: Result of confirmatory factor analysis (CFA)

Testing research hypothesis using SEM

Evaluate the fit of the model by linear structural model SEM:

After the measurement model has been validated by confirmatory factor analysis (CFA), we determined the linear structural model (SEM). The model fit indices indicate an adequate to excellent fit to the survey data: Chi-square index=412,975, df=200, Chi-square/df=2,065, P=0,000, GFI index=0,919 (≥ 0,9 is good) CFI index=0,977 (≥ 0,95 is very good), PCLOSE index=0,479 (≥ 0,05 is good), TLI index=0,973 (≥ 0,9 is good) and RMSEA=0,050 (≤ 0,06 is good). These indexes all satisfy the requirements when assessing general suitability.

BB = Products Packing, NT = Celebrity, KT = Customer Knowledge, CN = Perceived Value, YD = Purchase Intention

Figure 3: Linear structure model

Testing the linear structural model SEM

Hypothesis	Relationship	Regression coefficients are not standardized	Standardized regression coefficient	SE	CR	P	Hypothetical results
H1+	CN<---KT	0,302	0,404	0,034	8,867	***	Accept
H2+	CN<---NT	0,379	0,418	0,044	8,707	***	Accept
H3+	YD<---CN	0,297	0,315	0,051	5,833	***	Accept
H4+	YD<---KT	0,118	0,168	0,035	3,419	***	Accept
H5+	YD<---NT	0,199	0,233	0,043	4,642	***	Accept
H6+	YD<---BB	0,219	0,264	0,037	5,835	***	Accept

BB = Products Packing, NT = Celebrity, KT = Customer Knowledge, CN = Perceived Value, YD = Purchase Intention

Source: Processing results from survey data, 2024

The parameter estimation results indicate that all relationships in the research model are statistically significant (p < 0.05), thereby supporting all proposed hypotheses. Based on the standardized regression coefficients, the findings reveal:

· Holding other variables constant, a one-unit increase in Customer Knowledge leads to a 0.404-unit increase in Perceived Value.

· A one-unit increase in Celebrity Endorsement results in a 0.418-unit increase in Perceived Value.

· A one-unit increase in Perceived Value leads to a 0.315-unit increase in Purchase Intention.

· A one-unit increase in Customer Knowledge results in a 0.168-unit increase in Purchase Intention.

· A one-unit increase in Celebrity Endorsement leads to a 0.233-unit increase in Purchase Intention.

· A one-unit increase in Product Packaging results in a 0.264-unit increase in Purchase Intention.

These results confirm the positive influence of all examined factors on purchase intention within the proposed model.

Testing theoretical model estimation using Bootstrap:

For this study, the official survey sample comprised 425 respondents. To perform the Bootstrap procedure, the author generated 1,000 resamples to satisfy the requirement of the initial sample representing the population. The execution results are presented as follows.

Table 3: Estimation results using Bootstrap

Relationship			SE	SE-SE	Mean	Bias	SE-Bias
CN	<---	KT	0,050	0,001	0,400	-0,004	0,002
CN	<---	NT	0,054	0,001	0,420	0,001	0,002
YD	<---	CN	0,068	0,002	0,318	0,003	0,002
YD	<---	KT	0,048	0,001	0,165	-0,002	0,002
YD	<---	NT	0,058	0,001	0,231	-0,002	0,002
YD	<---	BB	0,050	0,001	0,265	0,002	0,002

BB = Products Packing, NT = Celebrity, KT = Customer Knowledge, CN = Perceived Value, YD = Purchase Intention

Source: Processing results from survey data, 2024

The Bootstrap analysis results show that all absolute CR values (Bias/SE-Bias) are ≤ 2, indicating that the estimates from the original sample are, on average, close to the population estimates. Both the bias and its standard error are small and stable. Therefore, it can be concluded that the adjusted estimates in the SEM model are statistically robust and reliable.

DISCUSSION:

This study identified four key factors influencing customers’ purchase intentions at Hoang Vu Chemical Trading Co., Ltd.: (1) customer knowledge, (2) perceived value, (3) celebrity endorsement, and (4) product packaging. A mixed-method approach was employed, with qualitative analysis used to refine the scales and quantitative analysis conducted on 425 valid survey responses collected via Google Forms. Data were analyzed using SPSS 26.0 and AMOS 20 to validate the research model and address the study objectives.

The research findings indicate that perceived value (β = 0.315) exerts the strongest influence on purchase intention. Consequently, the company should continuously monitor market trends, enhance products based on customer feedback, implement competitive pricing strategies, and improve service quality and after-sales support to foster customer loyalty. Product packaging (β = 0.264) ranks second, serving both protective and promotional functions. Attractive, brand-consistent, and sustainable packaging with clear information, regularly updated according to market feedback, can strengthen brand appeal.

Celebrity endorsement (β = 0.233) also enhances brand recognition and purchase intention. Selecting representatives aligned with brand values, integrating them into multi-channel campaigns, leveraging social media influence, and monitoring performance will optimize this effect. Lastly, customer knowledge (β = 0.168), though the least influential, remains important. Providing educational content, offering online resources, training sales staff, adapting messages based on feedback, and fostering customer communities can strengthen trust and purchasing decisions.

CONCLUSIONS:

The limitations of this study largely stem from its research scope. The survey was conducted exclusively with customers in Da Nang City, which restricts the generalizability of the findings. Furthermore, the use of a convenience (non-probability) sampling method with 425 respondents limits representativeness. Constraints in time, human resources, and funding also resulted in a modest sample size, which may have allowed subjective responses to influence the results. In addition, given the complexity of factors affecting purchase intention, the research model addresses only a subset of these variables, offering a general rather than comprehensive view.

Future research should address these limitations by increasing sample size to enhance the robustness of findings, and by expanding the survey to diverse geographic regions for broader applicability. Subsequent studies should also incorporate insights from a wider range of domestic and international literature to refine research models and measurement scales. Moreover, adding more relevant influencing factors would allow for a more detailed, multidimensional understanding of purchase intention, thereby enabling the formulation of more targeted and objective recommendations.

REFERENCES:

1. Babin B. Consumer self-regulation in a retail environment,. J Retail. 1995;71(1):47–70.

2. Mirabi DV, Akbariyeh H, Tahmasebifard H. A Study of Factors Affecting on Customers Purchase Intention. 2015;2(1).

3. Morwitz VG, Steckel JH, Gupta A. When Do Purchase Intentions Predict Sales?

4. Akar E, Nasir VA. A review of literature on consumers’ online purchase intentions. J Cust Behav. 2015 Nov 23;14(3):215–33.

5. Bais SS, Murthy A, Roy T, Mehendale S. Impact of Electronic word-of-mouth on Brand image and Purchase Intention: A Methodological and Thematic Review. Asian J Manag. 2018;9(1):774.

6. Praveer SR, Shrivastava AK. Demographic Inclination towards Purchase Factors: A Study on Fashion Apparels. 2010;

7. Devasenathipathi T, Mark J. Influence of consumer demographic characteristics on factors attracting the purchase of two-wheelers – a post-purchase survey of two-wheeler holders. Asian J Manag. 2012;3(4):204-9..

8. Doniavi A, Pourfatemi A. Identifying and prioritizing the factors affecting customer orientation in Urmia production cooperatives and providing proper solutions by using hierarchical techniques. Asian J Manag. 2016;7(2):110.

9. Pani D, Pradhan SK. Integrating effect of consumer perception on buying intention: a conceptual model of buying behaviour for tribal handicrafts products. Asian J Manag. 2017;8(4):1149-58.

10. Kumar MR, Sharma K. Pester Power and Advertisements Influence on Purchase of Food Products in a Convenience Store. Asian J Manag. 2017;8(2):204.

11. Kumar RR, Kumar SA. Analyse the factors of promoting arts organizations and its products through electronic mode (online mode): Certain issues and challenges. Asian J Manag. 2017;8(3):451.

12. Singh A, Singh P. Factors Determining online Purchase behaviour of The Consumers in Punjab (A study of Urban and Rural Consumers). Asian J Manag. 2018;9(1):133.

13. Chowdhury PP. Factors affecting Buyer Behavior: A Study on the Consumers of Two-Wheelers (Motorcycles) of Dhaka City. Asian J Manag. 2019;10(1):33.

14. Jha AK, Deshmukh GK, Joseph S. Customers Intention to Switch towards Mobile Number Portability in Chhattisgarh-A Study. Asian J Manag. 2016;7(3):213.

15. Achrol RS, Kotler P. Marketing in the Network Economy. J Mark. 1999;63:146.

16. Morwitz V. Consumers’ Purchase Intentions and their Behavior. Found Trends® Mark. 2012;7(3):181–230.

17. Shah SSH, Aziz J, Fatima M, Sherazi SK. The Impact of Brands on Consumer Purchase Intentions. 2012;

18. Younus S, Rasheed F, Zia A. Identifying the Factors Affecting Customer Purchase Intention. 2015;

19. Toldos-Romero MDLP, Orozco-Gómez MaM. Brand personality and purchase intention. Juan Carlos Sosa Varela D, editor. Eur Bus Rev. 2015 Aug 10;27(5):462–76.

20. Low GS, Lamb CW. The measurement and dimensionality of brand associations. J Prod Brand Manag. 2000 Nov;9(6):350–70.

21. Grant R, Clarke RJ, Kyriazis E. A review of factors affecting online consumer search behaviour from an information value perspective. J Mark Manag. 2007 June;23(5–6):519–33.

22. Lourenco F, Casey BJ. Adjusting behavior to changing environmental demands with development. Neurosci Biobehav Rev. 2013 Nov;37(9):2233–42.

Received on 20.10.2025 Revised on 18.11.2025

Accepted on 13.12.2025 Published on 14.02.2026

Available online from February 18, 2026

Research J. Science and Tech. 2026; 18(1):25-32.

DOI: 10.52711/2349-2988.2026.00004

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Creative Commons License.