Service Marketing Mix and MOOC Enrollment in Thailand: Exploring Brand Image as a Mediator

Abstract

This research develops and verifies a structural model of enrollment intention for Thai MOOC, the national learning platform, based on empirical data from a survey of 475 learners. Partial least squares structural equation modeling (PLS-SEM) was used to analyze the data, and the results indicated a strong model with high predictive capability. Among the seven dimensions of the service marketing mix, product, promotion, process and place had a significantly positive association with brand image perception. In contrast, perceived value, people, and physical evidence had no significant relation. Brand image perception was established as a mediator, representing the channel through which the significant marketing mix factors associated with the intention to enroll in Thai MOOC. These findings suggest that to induce enrollment, government-backed MOOCs should focus on content quality and platform accessibility ahead of conventional service aspects, while utilizing promotion to establish a strong brand image.

1. Introduction

The United Nations Sustainable Development Goal 4 highlights the importance of delivering inclusive and equitable quality education, in addition to promoting lifelong learning opportunities for everyone [1]. Consequently, education is a key pillar of national socioeconomic development and a competitive advantage in the global knowledge economy. Contemporary workforce dynamics require countries to facilitate comprehensive human capital development. This involves strategic upskilling programs to enhance professional competencies, systematic reskilling programs for career advancement, and proactive skill acquisition initiatives for future market opportunities. The expansion of knowledge acquisition frameworks has been necessary to address accelerated socioeconomic change and shifting labor market needs. This necessity culminated in the advent of massive open online courses (MOOCs), which were officially launched in 2008. MOOCs function as democratized learning environments that provide high-quality instruction to worldwide audiences through internet-facilitated access [2,3,4,5]. Prominent MOOC platforms, such as Coursera, Udacity, edX, FutureLearn, and Khan Academy, have positioned themselves as leading educational technology providers within the digital learning landscape. The core goals underlying MOOC innovation include two overarching dimensions: (1) the development of universally accessible learning opportunities that overcome geographical and time-limiting constraints while ensuring pedagogical superiority; (2) the systematic elimination of educational disparities through the offering of cost-free or significantly low-priced online learning materials [6,7,8,9].

MOOCs manifest unique pedagogical and structural benefits compared to traditional e-learning modes, owing to some basic characteristics. First, MOOCs display universal access and demographic inclusiveness through their ability to host large learner populations simultaneously regardless of gender, age, or educational backgrounds, thus breaking customary hindrances to educational access [10,11]. Second, MOOCs support personalized learning pathways through their intrinsic flexibility as technology-supported educational platforms accessible through various digital devices such as smartphones, tablets, and computers, with course participation independent of institutional affiliation or organizational membership prerequisites [10,12]. Third, MOOCs offer extensive curriculum variety that caters to heterogeneous learner needs and academic interests while overcoming the constraints of customary disciplinary divisions and standardized educational systems. Fourth, MOOCs integrate collaborative learning processes that support peer-to-peer communication and learner-educator interaction through embedded communication platforms, including discussion forums and collaborative workspaces, to promote community-based knowledge development [12,13]. Fifth, MOOCs implement credentialing systems through which successful completion of courses results in awarded certificates that function as validated credentials for job applications and academic progression within higher education institutions [10,12].

Throughout the Asian continent, national MOOC initiatives have been developed strategically to support lifelong learning goals, as represented by Thai MOOC, which launched in March 2017 [14]. As the primary initiative of the Thai Cyber University Project (TCU), Thai MOOC has become Thailand’s foremost digital learning platform. Through collaborations with over 120 domestic educational institutions, it offers more than 750 free online courses to a user base exceeding two million learners [14,15]. Institutional success of the platform is based on three pillars: extensive curricular breadth across a wide range of academic and professional fields, the availability of validated electronic certification mechanisms, and conformity to nationally standardized quality assurance metrics [16,17,18]. To secure and expand this institutional success, the platform has engaged in systematic upgrade, including advanced security protocols through linkage with national identification databases, a centralized credit banking system for academic portability, and enhanced digital credentialing [14,15].

Despite its key institutional successes and sustained technological innovation, Thai MOOC confronts a strategic challenge of paramount importance in improving the mechanisms of learner acquisition and retention to facilitate future institutional expansion. While the platform’s existing marketing efforts show operational activity, they lack a credible theoretical foundation. This is evidenced by insufficient systematic research examining marketing tactics that effectively develop brand image perception and increase enrollment intention in government-funded educational environments. This theoretical void creates operational challenges.

The governance context of government-backed MOOC platforms introduces additional theoretical complexity warranting systematic investigation. Ref. [19] demonstrates that governance structures and sociocultural factors fundamentally shape service system outcomes, particularly in public-sector initiatives where institutional mandates transcend commercial logic. The authors of ref. [20] establish that institutional readiness and sectoral development capacity critically determine platform adoption patterns. These governance dimensions distinguish Thai MOOC from commercial platforms, as its success depends not only on marketing effectiveness but on alignment with national educational priorities, inter-institutional coordination mechanisms, and cultural norms regarding digital learning adoption. This institutional embeddedness necessitates marketing frameworks specifically calibrated to mission-driven environments rather than profit-maximizing contexts. This research offers significant theoretical and practical contributions with direct benefits for multiple stakeholder groups; educational policy makers platform administrators, researchers, higher education institutions, learners and potential students.

Without understanding key marketing determinants, resource allocation becomes inefficient and strategic opportunities for learner engagement remain unrealized. Such theoretical ambiguity inherently constrains the platform from developing integrated marketing initiatives requisite for fulfilling its national educational commitments. Thus, the formulation and empirical validation of an inclusive theoretical model of enrollment intention goes beyond intellectual curiosity; it is a prime strategic necessity to equip Thai MOOC and similar government subsidized MOOC platforms with the empirical basis for evidence-informed strategic decision-making and institutional viability.

Considering Thai MOOC’s strategic needs to maintain growth, a key issue is posed by the absence of an evidence-based marketing model. This is exacerbated by an evident research gap; existing research on Thai MOOC has concentrated mainly on pedagogical concerns such as course design and learning strategies, omitting the vital link between the marketing mix, brand image, and enrollment intention [21,22]. A systematic study is thus needed to close this gap and identify which marketing factors have the greatest impact in this specific environment. A tested model would furnish the empirical basis necessary for crafting effective, integrated marketing strategies. This research addresses this gap by pursuing two objectives: first, to adapt and empirically test established service marketing and brand perception frameworks within the specific context of government-backed MOOC platforms in Southeast Asia; second, to identify which elements of the conventional 7Ps framework retain relevance in mission-driven, technology-mediated educational environments and which require reconceptualization. Fulfilling these aims will make an important contribution by providing a practical, evidence-based tool to boost marketing effectiveness for Thai MOOC and similar platforms in mission-driven environments, supporting ultimately the aim of fostering lifelong learning.

This research makes three distinct theoretical contributions. First, it extends service marketing theory by empirically identifying boundary conditions under which conventional 7Ps frameworks require modification—specifically demonstrating that digital scalability, technology-mediated delivery, and mission-driven mandates fundamentally alter the salience of perceived value, people, and physical evidence constructs. Second, it advances brand theory in educational contexts by validating mediation models, establishing brand image as the cognitive mechanism through which marketing activities associate behavioral intentions in high-involvement educational decisions. Third, it contributes to MOOC literature by providing the first empirically validated framework specifically designed for government-backed platforms operating under distinct institutional logics from commercial counterparts, addressing a critical gap in understanding how public-sector digital learning platforms should approach strategic marketing. This study therefore examines how marketing effectiveness serves as an enabler of sustainable educational transformation, where increased enrollment translates directly into expanded educational equity and resource-efficient knowledge dissemination.

2. Literature Review and Hypothesis Development

2.1. Consumer Decision-Making

Consumer decision making theory provides a core theoretical foundation for explaining the processes by which external marketing stimuli are translated into behavioral intentions through cognitive processing mechanisms. Ref. [23] characterize consumer decision making as a dynamic process with three basic elements: input factors consisting of external marketing stimuli, processing mechanisms reflecting psychological assessment including brand image creation, and output expressions in the form of behavioral intentions. The theoretical model shows specific relevance in educational settings, where enrollment decisions reflect high involvement decision making processes and complex cognitive assessments requiring extensive information processing and systematic evaluation of intangible service features. Theoretical application of service marketing concepts to educational contexts mandates paradigm adaptation from traditional goods marketing theory, necessitated by experiential nature inherent in educational services, the focus on long term outcome attainment, and the primacy of institutional credibility in student choice processes [9]. The extended service marketing mix framework, comprising seven components (7Ps), offers a holistic analytical framework necessary for investigating educational service provision [17]. Unlike tangible products, education services such as MOOCs are inherently experiential and intangible in nature, making the traditional 4Ps model (Product, Perceived Value, Place, Promotion) theoretically inadequate for holistic analysis Ref. [24]. The additional facets of People, Process, and Physical Evidence are found to be effective in encapsulating the entire service experience environment [25]. The People aspect subsumes instructor proficiency and support staff quality, Process includes the end-to-end learner journey from registration to certification, and Physical Evidence integrates the virtual learning environment architecture and platform interface design [24,25]. The 7Ps framework implementation thus enables thorough scrutiny of all service touchpoints impacting learner perception and, in turn, enrollment decision making in service driven contexts [17,26,27]. 7Ps marketing mix framework has found ubiquitous adoption across educational institutions worldwide for optimizing student recruitment and retention outcomes [10,28], with empirical studies verifying that individual components of the framework have differential effects on institutional branding performance [13].

While digital learning environments involve multiple theoretical frameworks including Technology Acceptance Model (TAM), perceived usefulness constructs, and learning motivation theories—this study positions brand image as the primary mediating mechanism for three theoretical reasons. First, MOOC enrollment represents a high-involvement decision requiring extensive cognitive processing and institutional evaluation [29]. Unlike routine technology adoption decisions where TAM constructs suffice, educational service selection demands holistic assessment of institutional credibility, outcome quality, and long-term value—dimensions that brand image integrates into unified perceptions [23]. Second, consumer decision-making theory establishes brand image as a higher-order cognitive abstraction that consolidates multiple lower-order evaluations, including technology acceptance, perceived usefulness, and platform experience, into gestalt institutional judgments [23]. Brand image functions as the integrative schema through which learners synthesize disparate marketing stimuli into actionable behavioral intentions. Third, empirical research in educational contexts demonstrates that brand image mediates the relationship between functional service attributes (including technology features) and enrollment intentions, with its three-dimensional structure (functional, emotional, reputational) capturing both utilitarian and affective evaluation processes [5,30]. While technology acceptance and learning motivation represent important parallel processes, brand image serves as the overarching cognitive framework that encompasses these dimensions within government-funded educational platforms where institutional credibility supersedes individual feature evaluations.

2.2. Service Marketing Mix Framework

The services marketing mix framework requires systematic examination of how each variable contributes to brand image perception in higher education. Empirical evidence shows different mechanisms and influence magnitudes across service industries [31]. In digital learning environments, Product quality serves as the primary brand differentiator because, unlike commercial services where quality judgments remain subjective, educational outcomes manifest through measurable competency acquisition and career advancement [5,32]. This tangibility makes course quality the foundational mechanism through which functional brand image develops in MOOC contexts.

Perceived value in mission-driven, no-cost educational platforms like Thai MOOC operates on different principles compared to commercial educational services. Perceived value of educational services is linked to brand reputation through impressions of affordability and perceptions of accessibility, and not through signals of quality, particularly as students make decisions on long-term return on investment based on the prospects for career development [33]. The Place element in the context of e-learning has evolved from geographical accessibility advantages to encompass digital presence and platform functionality, where the aesthetics of interfaces, mobile compatibility, and multi-device accessibility enhance the educational brand image by impacting the convenience of learning and engagement potential [29].

Education marketing requires communications strategies rooted in credibility that depart radically from the practices of product promotion. Empirical research validates that promotional practices focused on learning outcomes, motivated by success stories, validated by experts, and based on evidence provide higher returns in brand image than traditional advertising appeals, since educational consumers require credible validation of educational value instead of emotional appeals [29,32]. The People dimension encompasses the competence of instructors, quality of administrative services, and norms of peer relationships, with recent studies [30,34] providing evidence that faculty expertise, quality of instruction, and student support services exhibit strong association with the perception of educational brands, by virtue of their linkages with satisfaction from the learning experience and academic success.

The Process aspect involves registration processes, mechanisms for monitoring learning progress, assessment methods, and certification procedures. Mahajan and Golahit [35] illustrated that optimized academic processes have a positive effect on brand reputation through decreased administrative complexity and improved learning process quality. Physical Evidence in conventional educational settings includes campus infrastructure and facilities; however, the advancement in technology has directed focus towards virtual evidence factors, such as platform design, digital content, and virtual learning environment quality [34].

2.3. Brand Image

Brand image perception acts as the primary mediating process connecting service marketing mix elements to enrollment intention through cognitive appraisal processes, in which students consolidate their marketing mix exposure into consolidated institutional judgments that guide behavioral choice. This theoretical model is based on consumer decision making theory and brand perception studies, where Kotler and Keller [29] confirm that brand perceptions act as mediating cognitive abstractions that convert marketing stimuli to behavioral reactions. In the context of education, this mediation process echoes the high involvement character of enrollment choice, demanding extensive institutional consideration through brand image development before committing to educational investment [10].

Empirical support for education brand image mediation illustrates how functional, emotional, and reputational brand aspects separately mediate the marketing activities and enrollment intention relationship through cognitive integration processes and behavioral intention development mechanisms [30]. Alcaide-Pulido et al. undertook in-depth empirical analysis of three unique higher education brand image aspects, namely functional, emotional, and reputational dimensions, thus providing a strong theoretical context for the assessment of educational brand image perception [29].

2.4. Research Gaps

This study addresses three contextual gaps in the current MOOC literature. While marketing frameworks have been extensively validated in commercial higher education contexts, their applicability to government-backed, mission-driven platforms in developing economies remains underexplored. First, there is a geographical gap in which the majority of current research is drawn from Western MOOC environments [3], whereas Southeast Asian government backed MOOC platforms are situated within unique cultural, economic, and educational frameworks that demand regionally specific research approaches. Second, there is an institutional gap in that current studies overwhelmingly investigate commercial platforms like Coursera and edX or university-based MOOC programs [36], whereas mission driven government backed platforms like Thai MOOC exhibit inherently different institutional logics, funding models, and strategic imperatives that demand differing marketing strategies. Third, there is a theoretical integration gap in that MOOC research has traditionally focused on learning analytics and technology acceptance models [37], systematically overlooking in-depth application of marketing theoretical frameworks necessary for strategic platform growth and learner acquisition strategies. Therefore, the hypotheses of this study are as follows:

H1a–g. 

Each service marketing mix (7Ps) positively associates with the brand image perception of Thai MOOC.

H2. 

Brand image perception positively associates with intention to enroll in Thai MOOC.

H3a–g. 

Each service marketing mix (7Ps) positively associates with the intention to enroll in Thai MOOC, mediated by brand image perception.

3. Methodology

3.1. Research Design

This research uses a quantitative, cross-sectional research design to create and empirically test a model of MOOC enrollment intention. The choice of partial least squares structural equation modeling (PLS-SEM) as the method of analysis is an intentional methodological decision, based on its appropriateness for the objectives and model characteristics of this study. The choice is guaranteed for several important reasons. Firstly, one of the main aims of this research is to explain principal predictors of enrollment intention and examine novel theoretical relationships, which maps onto PLS-SEM’s established strength as a prediction-focused and theory-building approach. Secondly, the hypothesized model is complex in nature, integrating the elaborate 7Ps service marketing mix framework, and PLS-SEM excels at analyzing such models. Thirdly, since this study transfers established marketing theory to the new environment of a government-funded MOOC platform, the exploratory focus of PLS-SEM is more suitable than the more constrained, theory-confirming nature of covariance-based structural equation modeling (CB-SEM) [38].

Figure 1 depicts the suggested theoretical framework of the structural model of enrollment intention in the Thai MOOC environment. The conceptual model includes three main constructs: service marketing mix (7Ps), brand image perception, and intention to enroll in Thai MOOC. The following hypotheses are included in the theoretical model.

3.2. Data Collection

The target population for the present research consists of both prospective students showing interest in Thai MOOC and current users of the platform, creating a virtually unlimited population parameter. Adopting an analytical strategy of PLS-SEM, sample size determination was carried out using G*Power software (version 3.1.9.7) [39], which incorporated seven predictors derived from the service marketing mix model. The statistical parameters were set as follows: effect size = 0.15, Type I error probability = 0.05, and statistical power = 0.95, which yielded a minimum required sample size of 153 respondents. The present study successfully obtained a sample size of 475 participants, well above the G*Power recommendation and thus providing adequate statistical power for the subsequent analyses.

Ethical compliance was maintained through extensive informed consent processes operationalized by the consolidated online consent feature in the survey software. The survey opened provided complete disclosure about research aims, participant expectations, and data handling procedures. Survey continuation required overt consent confirmation through active clicking of the ‘Continue’ button. This procedure ensured voluntary participation. This electronic consent model ensured that survey completion was an overt agreement to research participation while upholding participant autonomy through ongoing withdrawal avenues via survey abandonment.

Data were collected through an online survey from 1 June to 31 July 2024. The survey targeted potential learners and actual users of Thai MOOC who were actively engaged in social media platforms associated with Thai MOOC, including Facebook, Instagram, YouTube, and TikTok. This study employed non-probability convenience sampling of Thai MOOC social media followers. While this approach restricts generalizability, it aligns with the study’s objective: constructing and testing an association model within the target market of current and potential users. Results generalize only to digitally connected, socially engaged Thai MOOC followers who are substantially more technologically literate, educationally motivated, and positively predisposed toward the platform than the broader population of potential learners. A four-stage randomization protocol was conducted via sequential phases utilizing standardized randomization procedures. First, Facebook was randomly chosen from Thai MOOC’s four main social media platforms (Facebook, Instagram, YouTube, TikTok) via Random.org (https://www.random.org, accessed on 25 September 2025) number generation with sequential numerical designation (1 to 4) for each platform. Next, Thursday was randomly chosen from the seven-day weekly cycle using the same numerical designation and randomization procedures. The third phase consisted of random selection of 10:00 a.m. from hourly intervals within the 9:00 a.m. to 4:00 p.m. operational window. Lastly, stratified random sampling was conducted within these parameters, dividing participants by engagement intensity levels (high, medium, low) to provide representative coverage within Thai MOOC’s follower demographics while controlling for platform-specific and temporal sampling biases.

Though online survey methodology supported operational efficiency and wide demographic reach, several methodological limitations are worthy of consideration. Sole reliance upon social media-based recruitment poses potential self-selection bias, as participation was limited to individuals already evidencing Thai MOOC digital platform engagement, possibly overrepresenting technologically adept and intrinsically motivated learners. This sampling approach may systematically underrepresent potential learners with limited digital literacy, fewer socioeconomic resources, or rural geographical locations with limited internet connectivity, thus limiting generalizability to Thailand’s wider population. To partially counter these limitations, the study employed stratified sampling based on engagement levels (high, medium, low) to provide for varying degrees of platform use. Temporal randomization protocols further minimized potential temporal bias in user accessibility patterns. However, we acknowledge that self-selection bias likely influences our findings in three specific ways: (1) overestimation of brand image perceptions, as respondents already favorable toward Thai MOOC are overrepresented; (2) potential inflation of path coefficients, particularly for brand image-to-intention relationships; (3) underrepresentation of barriers to adoption among digitally disconnected populations. The high proportion of government officers (48.2%) who face mandatory training requirements further introduces systematic bias toward extrinsically motivated learners. Results must therefore be interpreted within the qualified context of digitally connected, already-engaged populations rather than representing the general Thai population’s attitudes toward MOOCs.

3.3. Measurement of Constructs

The measurement instrument utilized a five-point Likert scale (1 = Strongly Disagree, 5 = Strongly Agree) to operationalize the study’s primary constructs. The service marketing mix (7Ps) was assessed with a 26-item scale developed for the Thai MOOC context. This scale encompassed seven dimensions: five items for Product, such as “Courses on Thai MOOC meet my needs”; three for Perceived Value, including “The knowledge gained is worth the time and effort spent”; three for Place, like “I can study courses on Thai MOOC anytime, anywhere”; four for Promotion, such as “There are positive reviews from other Thai MOOC users”; four for People, including “The instructors are knowledgeable and have expertise”; four for Process, like “The enrollment process is simple and not complicated”; and three for Physical Evidence, such as “The Thai MOOC website has an interesting design and is easy to use”. The brand image perception construct was operationalized with 14 items adapted from Alcaide-Pulido et al. (Alcaide-Pulido et al., 2022 [40]). These measured three core dimensions: Functional (5 items), assessing perceptions like Thai MOOC provides a distinct and innovative learning experience compared to traditional educational institutions”; Emotional (5 items), gauging feelings such as “I feel fond of and attached to Thai MOOC”; and Reputational (4 items), evaluating beliefs like “Thai MOOC has a strong and stable foundation”. Finally, enrollment intention was measured with a four-item scale, including statements such as “I intend to enroll in a course on Thai MOOC”. Prior to data collection, the instrument’s content validity was established using the index of item-objective congruence (IOC), with all items exceeding the 0.67 criterion. A pilot test involving 30 respondents subsequently confirmed the instrument’s excellent internal consistency, yielding Cronbach’s alpha coefficients between 0.915 and 0.944, a range considered excellent [41].

3.4. Estimation Method

The theoretical model proposed was analyzed using PLS-SEM methodology instead of CB-SEM techniques. PLS-SEM was chosen for exploratory research settings because it has greater analytical rigor and demands more conservative result interpretation than CB-SEM [38]. All analyses were performed using SmartPLS 4 software [42] with established analytical procedures. The analytic process involved three consecutive stages. First, confirmatory factor analysis was conducted to remove measurement items showing factor loadings less than the 0.7 cut-off criterion [38]. Then, overall evaluation of internal consistency, reliability, and validity parameters was carried out for the theoretical model hypothesized [38]. Lastly, structural model estimation was carried out, followed by systematic exploration and empirical testing of the hypothesized framework. Reliability testing included dual evaluation criteria using Cronbach’s alpha coefficients and composite reliability measures, tested alongside convergent validity measures [38]. Average variance extracted (AVE) was tested using the 0.5 threshold value criterion for acceptable convergent validity [38]. Discriminant validity testing included comparative testing between latent variable correlation coefficients and square root of AVE values such that AVE square root values were greater than respective correlation coefficients, thus demonstrating sufficient discriminant validity [38]. Hypothesis testing routines utilized bootstrap resampling technique with 5000 iterations at 95% confidence intervals under the PLS algorithmic framework to provide robust statistical inference and accuracy in parameter estimation [38].

4. Results

The use of single source survey data necessitates an explicit exploration of the possibility of common method bias (CMB), which may potentially inflate relationships among theoretical constructs. To control for this methodological flaw, several procedural remedies were implemented, such as randomization of question order, ensuring anonymity of participants, and using temporal separation in the measurement protocols of constructs. A statistical test was conducted using Harman’s single factor test, and it was observed that the first unrotated factor explained 40.3% of the total variance, falling below the conventional threshold criterion of 50%, thus suggesting that CMB does not represent a substantial methodological threat to construct validity [43].

Prior to structural model assessment, we conducted three robustness checks to ensure empirical credibility. First, multicollinearity diagnostics revealed all variance inflation factor (VIF) values ranged from 2.38 to 4.24, below the conservative threshold of 5.0, indicating no problematic multicollinearity among predictor variables [38]. Second, outlier analysis using Mahalanobis distance (D²) at p < 0.001 identified 8 potential multivariate outliers (1.7% of sample); sensitivity analysis with and without these cases showed no substantive changes in path coefficients (maximum difference = 0.023), confirming result stability.

4.1. Profile of the Respondents

Most of the respondents were female, aged 31–36 years old, working as government officers, and possessing a bachelor’s degree, as shown in

Table 1. Demographic Characteristics of Samples.

Respondent Profile		Frequency (%)	Respondent Profile	Frequency (%)
Gender	Male	210 (44.2)	Occupation
	Female	265 (55.8)	Government officers	229 (48.2)
Age	19–24	96 (20.2)	Employees	88 (18.5)
	25–30	68 (14.3)	Business owners	66 (13.9)
	31–36	105 (22.1)	Students	66 (13.9)
	37–42	76 (16.0)	Not mentioned	26 (5.5)
	43–48	64 (13.5)	Education Level
	49–54	38 (8.0)	Below bachelor’s degree	78 (16.4)
	55–60	17 (3.6)	Bachelor’s degree	240 (50.5)
	Above 60	11 (2.3)	Above bachelor’s degree	157 (33.1)

Source: Own Survey.

. The high percentage of government officers (48.2%) is a hallmark of the Thai MOOC population. Employees in the Thai public sector are frequently obligated to finish yearly training hours to attain key performance indicator (KPI) goals for professional development.

4.2. Measurement Model

Prior to the SEM analysis, a normality test ascertained that all the indicators had normality since their values of skewness and kurtosis fell within acceptable limits of ±1 and ±1.5, respectively [44]. This reveals that the data are suitable for parametric statistical analysis.

Table 2. Comprehensive Measurement Model Evaluation.

Construct	Indicators	Mean	S.D.	Loading	Alpha	ρA	ρC	AVE
Service Marketing Mix (7Ps)
Product (PRD)	PRD1–PRD5	4.375–4.482	0.581–0.614	0.893–0.907	0.935	0.936	0.951	0.794
Perceived Value (PRC)	PRC1–PRC3	4.48–4.516	0.567–0.596	0.896–0.932	0.901	0.903	0.938	0.835
Place (PLC)	PLC1–PLC3	4.484–4.507	0.571–0.606	0.895–0.927	0.902	0.902	0.939	0.836
Promotion (PRO)	PRO1–PRO4	4.396–4.463	0.602–0.626	0.897–0.919	0.930	0.930	0.950	0.826
People (PEO)	PEO1–PEO4	4.478–4.507	0.567–0.599	0.905–0.935	0.939	0.940	0.957	0.846
Process (PRS)	PRS1–PRS4	4.408–4.444	0.604–0.637	0.909–0.925	0.938	0.939	0.955	0.842
Physical Evidence (PHY)	PHY1–PHY3	4.421–4.463	0.601–0.623	0.922–0.947	0.928	0.928	0.954	0.874
Brand Image Perception
Functional (FCN)	FCN1–FCN4	4.440–4.500	0.553–0.597	0.877–0.901	0.915	0.915	0.940	0.797
Emotional (EMO)	EMO1–EMO5	4.410–4.450	0.598–0.628	0.892–0.917	0.944	0.944	0.957	0.816
Reputational (REP)	REP1–REP4	4.430–4.490	0.584–0.603	0.896–0.911	0.925	0.925	0.947	0.817
Intention to Enroll
Intention to Enroll (INT)	INT1–INT4	4.440–4.480	0.585–0.624	0.890–0.918	0.921	0.922	0.944	0.809

Note. Alpha, Cronbach’s alpha; AVE, average variance extracted. Source: Own Survey.

presents the comprehensive measurement model evaluation results. All constructs demonstrated excellent psychometric properties. Cronbach’s alpha coefficients ranged from 0.901 to 0.944, composite reliability (ρC) from 0.938 to 0.957, and AVE from 0.794 to 0.874. These values substantially exceed recommended thresholds of 0.7, 0.7, and 0.5, respectively [45]. Factor loadings for all indicators ranged from 0.877 to 0.947, surpassing the 0.7 criterion for adequate item reliability. The results confirm robust internal consistency, composite reliability, and convergent validity across all constructs in the service marketing mix (7Ps), brand image perception dimensions, and intention to enroll, validating the appropriateness of all measurement items for their respective latent variables.

Data in Table 2 show high means (4.37–4.52) and low standard deviations (0.55–0.64). We recognize this low variance as a methodological limitation, as it can decrease statistical power in SEM analysis. This pattern is likely due to a self-selection bias, since the sample was recruited from current followers of Thai MOOC who would be inclined to perceive the platform positively. In spite of this limitation, the excellent reliability and validity figures (AVEs > 0.79, α > 0.90) and significant path coefficients indicate adequate variance was available for a proper analysis of this user group [38]. Thus, the results should be taken to be representative of active and digitally connected users instead of the general population.

Discriminant validity was confirmed using the HTMT criterion [45], with all ratios ranging from 0.615 to 0.765, well below the conservative 0.85 threshold, presented in

Table 3. Discriminant Validity: Heterotrait–Monotrait Ratio of Correlations (HTMT).

	PRD	PRC	PLC	PRO	PEO	PRS	PHY	BIP	INT
Product (PRD)	-
Perceived Value (PRC)	0.640	-
Place (PLC)	0.657	0.716	-
Promotion (PRO)	0.650	0.641	0.663	-
People (PEO)	0.666	0.727	0.728	0.704	-
Process (PRS)	0.683	0.661	0.695	0.695	0.716	-
Physical evidence (PHY)	0.673	0.703	0.698	0.697	0.740	0.765	-
Brand image perception (BIP)	0.671	0.695	0.705	0.661	0.705	0.669	0.678	-
Intention to enroll in Thai MOOC (INT)	0.615	0.692	0.694	0.614	0.692	0.641	0.655	0.678	-

Source: Own Survey.

. The highest HTMT value of 0.765 between Process and Physical Evidence indicates satisfactory discriminant validity across all constructs, supporting the measurement model’s validity.

4.3. Evaluation of the Structural Model

The main criteria for evaluating the structural model were multicollinearity, coefficient of determination (R²), effect size (f²), predictive relevance (Q²), and model fit [38]. The coefficient of determination (R²) of brand image perception and intention to enroll in Thai MOOC were 0.889 and 0.714, respectively. Both constructs had relatively high R² value exceeding 0.7 (70%); therefore, the explanatory power of this study is considered strong, as shown in

Table 4. Explained Variance (R²).

Construct	R2	Adjusted R2
Brand image perception (BIP)	0.889	0.888
Intention to enroll in Thai MOOC (INT)	0.714	0.714

Source: Own Survey.

.

Effect size interpretation using Cohen’s criteria [46], (0.02 = small, 0.15 = medium, 0.35 = large) reveals important practical insights beyond statistical significance. As shown in

Table 5. F2 Effect Sizes.

Construct	Brand Image Perception (BIP)
Product (PRD)	0.310
Perceived Value (PCS)	0.014
Place (PLC)	0.030
Promotion (PRO)	0.092
People (PEO)	0.000
Process (PRS)	0.028
Physical evidence (PHY)	0.007

Source: Own Survey.

, the Product to Brand Image Perception relationship demonstrates a large effect size (f² = 0.310), indicating substantial practical importance and justifying major resource allocation to course quality enhancement. Promotion (f² = 0.092), Place (f² = 0.030) and Process (f² = 0.028) exhibit small effect sizes, indicating limited practical significance. Notably, People demonstrates zero practical association (f² = 0.000), while Perceived Value and Physical Evidence show negligible effects (f² = 0.014 and 0.007), suggesting minimal practical importance in the MOOC context. These effect size interpretations indicate that product enhancement should be the primary strategic priority, followed by promotional activities, with other elements supporting tactical roles.

The Q² value of this study is greater than zero for the endogenous latent variable, indicating that the PLS path model possesses strong predictive relevance for the latent variable, as shown in

Table 6. Predictive Relevance (Q²-value).

	Cross-Validated Redundancy		Cross-Validated Communality
	Q2	Prediction Capability	Q2	Prediction Capability
Brand image perception (BIP)	0.882	0.882	0.886	0.886
Intention to enroll in Thai MOOC (INT)	0.737	0.737	0.777	0.777

Source: Own Survey.

. The predictive relevance based on the cross-validated redundancy for the latent variables BIP and INT was classified as high (Q² > 0.35), and the predictive power based on the cross-validated communality was also considered high (Q² > 0.35), as shown in Table 6. This implies that the model has significant predictive power.

Structural model analysis (

Table 7. Structural Model Results.

Direct Effect Testing	β	t-test	p Values	f-sq	Results
PRD → BIP (H1a)	0.491 ***	8.184	0.000	0.310	Supported
PRC → BIP (H1b)	0.093	1.910	0.056	0.014	Not Supported
PLC → BIP (H1c)	0.139 **	2.768	0.006	0.030	Supported
PRO → BIP (H1d)	0.212 ***	4.194	0.000	0.092	Supported
PEO → BIP (H1e)	−0.003	0.053	0.958	0.000	Not Supported
PRS → BIP (H1f)	0.141 *	2.369	0.018	0.028	Supported
PHY → BIP (H1g)	−0.072	1.277	0.202	0.007	Not Supported
BIP → INT (H2)	0.845 ***	44.519	0.000	2.499	Supported
Indirect Effect Testing	β	t-test	pValues	f-sq	Results
PRD → BIP → INT (H3a)	0.179 ***	4.231	0.000	-	Supported
PRC → BIP → INT (H3b)	0.079	1.901	0.057	-	Not Supported
PLC → BIP → INT (H3c)	0.118 **	2.752	0.006	-	Supported
PRO → BIP → INT (H3d)	0.179 ***	4.231	0.000	-	Supported
PEO → BIP → INT (H3e)	−0.003	0.053	0.958	-	Not Supported
PRS → BIP → INT (H3f)	0.119 *	2.360	0.018	-	Supported
PHY → BIP → INT (H3g)	−0.061	1.276	0.202	-	Not Supported

Note: *** p < 0.001, ** p < 0.01, * p < 0.05. Source: Own Survey.

and

Table 8. Direct, Indirect, and Total Effect.

	Brand Image Perception (BIP) R2 = 0.889				Intention to Enroll in Thai MOOC (INT) R2 = 0.714
	DE	t-Test	IE	TE	DE	t-Test	IE	t-Test	TE
Product (PRD)	0.491 ***	8.184	-	0.491 ***	-	-	0.179 ***	4.231	0.415 ***
Perceived Value (PRC)	0.093	1.910	-	0.093	-	-	0.079	1.901	0.079
Place (PLC)	0.139 **	2.768	-	0.139 **	-	-	0.118 **	2.752	0.118 **
Promotion (PRO)	0.212 ***	4.194	-	0.212 ***	-	-	0.179 ***	4.231	0.179 ***
People (PEO)	−0.003	0.053	-	−0.003	-	-	−0.003	0.053	−0.003
Process (PRS)	0.093 *	2.369	-	0.093 *	-	-	0.119 *	2.360	0.119 *
Physical evidence (PHY)	−0.072	1.277	-	−0.072	-	-	−0.061	1.276	−0.061
Brand image perception (BIP)	-		-	-	0.845 ***	44.519	-		0.845 ***

Note: *** p < 0.001, ** p < 0.01, * p < 0.05; DE = Direct Effect, IE = Indirect Effect, TE = Total Effect Source: Own Survey.

) reveals that four of seven marketing mix dimensions significantly predict brand image: Product (β = 0.491, p < 0.001), Promotion (β = 0.212, p < 0.001), Place (β = 0.139, p < 0.01), and Process (β = 0.141, p < 0.05), supporting H1a, H1c, H1d, and H1f. Perceived Value, People, and Physical Evidence showed no significant associations, rejecting H1b, H1e, and H1g. Brand image strongly predicts enrollment intention (β = 0.845, p < 0.001, supporting H2), establishing mediation as Product, Place, Promotion, and Process associate with intention through brand image (indirect effects: β = 0.415, 0.118, 0.179, and 0.119, respectively, all p < 0.05), confirming H3a, H3c, H3d, and H3f.

Figure 2 illustrates the tested structural model showing the relationships among the service marketing mix elements, brand image perception, and intention to enroll in Thai MOOCs.

5. Discussion

5.1. Direct Effects: Marketing Mix→Brand Image Perception

Product was the strongest predictor of brand image perception (β = 0.491, p < 0.001), consistent with education marketing literature that indicates that course quality is the predominant force behind institutional perceptions [9,17,24]. The key to success of Thai MOOC is offering diverse, expert-lecturer-taught courses from renowned Thai institutions to meet different learner aspirations for segments of lifelong learning. This finding affirms that in mission-driven educational platforms, content quality becomes the most important brand differentiator over traditional service attributes. Promotion showed the second highest correlation (β = 0.212, p < 0.001), aligning with literature citing credibility-based educational promotion success [29,32]. Thai MOOC’s multi-channel communication approach via social media, student testimonials, and institution collaborations generates positive brand attitudes by highlighting educational results rather than commercial inducements [10]. Process and Place also demonstrated moderate yet significant associations (β = 0.141 and 0.139, respectively), aligning with research on Lean learning processes and online accessibility [28,47]. Easy registration, straightforward navigation, and cross-device functionality of Thai MOOC facilitate brand perception through increased user experience, especially important for heterogenous demographic cohorts with disparate digital literacy levels [6,24].

5.2. Non-Significant Results: Theoretical Implications

The inability to detect significant impacts for Perceived Value, People, and Physical Evidence demonstrates fundamental theoretical shortcomings of conventional service marketing in government-funded digital learning websites. The non-significant effect of our Perceived Value construct (β = 0.093, p = 0.056) must be interpreted cautiously due to operationalization limitations. Our items measured perceived value and opportunity cost rather than monetary price [7,33]. Additionally, our sample’s restriction to already-engaged users likely attenuated variance in value perceptions. The theoretical relevance of pricing mechanisms in free government-sponsored platforms remains an open question requiring different methodological approaches, such as experimental designs manipulating certification costs or comparing free versus paid credential options.

The non-significant association of People (β = −0.003, p = 0.958) and Physical Evidence (β = −0.072, p = 0.202) should not be interpreted as construct irrelevance but rather as limitations in operationalization and sampling. First, our Physical Evidence may overlap conceptually with Process items. Second, our sample of already-engaged users likely demonstrates restricted variance in perceptions of instructor quality and platform design, having already overcome barriers these factors might pose for new users. Third, the technology-mediated nature of MOOCs may require reconceptualization of these constructs rather than their dismissal. Furthermore, this may be a sign of generation change whereby digital native learners prefer content access to instructor relationships, or cultural dimensions of Thai education emphasizing knowledge acquisition over interpersonal relations [34]. Physical Evidence irrelevance is the notion that service virtualization has made conventional physical quality cues obsolete in virtual learning environments, where virtual experience design and platform usability take the place of physical facilities as important quality signals [6,48].

5.3. Mediation Effects: Brand Image as Strategic Mechanism

A key contribution of this research is the establishment of brand image perception as the mediator between key marketing mix factors and enrollment intention (β = 0.845, p < 0.001). This transforms the status of brand image from a mere influencing variable to the very cognitive nexus where marketing activities are transmuted into behavioral consequences. Essentially, brand image functions as a cognitive schema that learners employ to reduce the high-involvement decision of enrolling in a course [30]. Instead of evaluating each marketing cue separately, potential learners synthesize these cues into one unified, gestalt perception of the brand, which in turn becomes the overriding predictor of behavioral intention. This research breaks down this brand image into its functional, emotional, and reputational components, and the results demonstrate a definite synergy among the elements of the marketing mix in developing these aspects [23]. The Product dimension, with the highest total effect (β = 0.415), forms the foundation of the functional brand image; relevant, high-quality courses create the platform’s essential value and credibility. This functional base is then supported by an efficient Process and convenient Place, which fulfill the promise of quality with a smooth user experience. It is only on this strong functional foundation that Promotion is able to develop the emotional and reputational aspects. The success stories and testimonials promoted build credibility and user affinity precisely because the product is seen to be strong. In the absence of a quality product, promotion would ring hollow and be unable to construct a positive brand image. Strategically, this whole mediation model suggests that discrete marketing initiatives are inadequate. For Thai MOOC, success demands a move towards integrated brand management, wherein each marketing action is considered in terms of its effect on the overall brand image. This is especially important in a non-commercial, mission-driven environment. In this setting, a robust, reliable brand image, particularly its reputational and functional dimensions must take the place of customary market signals. It becomes the main currency of value, promising learners that their time and effort investment will pay off [25,28,49].

5.4. Paradigm Shift: Digital-Native Educational Marketing

Cumulatively, these findings demonstrate a paradigm shift in theory away from conventional service marketing towards digital-native theories with a focus on virtual evidence, tech-mediated delivery, and mission-driven value propositions. The findings indicate that MOOC environments require new marketing models emphasizing digital content quality and user experience over traditional service factors. This contributes to theory by defining boundary conditions under which classical service marketing relationships do not apply in digitally native learning settings [10]. The most saturated affective dimension of brand image aligns with literature prioritizing affective bonds in educational brand development [32], and indicates that the success of Thai MOOC relies on establishing emotional bonds through engaging learning experiences and not necessarily the delivery of functional services.

Beyond conventional marketing outcomes, this study’s findings have significant implications for educational sustainability. The dominance of Product quality as the primary brand predictor suggests that sustainable MOOC success depends fundamentally on delivering substantive educational value rather than superficial promotional appeals, aligning with principles of sustainable consumption where quality supersedes quantity [32], The accessibility findings directly address digital inclusion imperatives within sustainable education frameworks across devices and connectivity contexts as a result of technological barriers among rural and economically disadvantaged populations, thereby advancing equity dimensions of SDG 4.3 [1].

6. Research Implications

6.1. Theoretical Implications

The findings contradict conventional service marketing paradigms. They demonstrate that digital learning environments require modified theoretical models rather than straightforward applications of traditional 7Ps frameworks. The dominance of product association, coupled with insignificant correlations for perceived value, people, and physical space, delineates digital scalability and technology-enabled delivery as essential boundary conditions that transform core service marketing relationships. This extends the theory of service marketing by defining the conditions under which older models break down and digital education services as a theoretical class that needs special frameworks with priority to content quality over interpersonal service aspects. The intangible effects of pricing extend service marketing theory by separating mission-driven and profit-driven service environments as a key theoretical differentiation.

The study offers important contributions to brand theory in that it empirically validates mediation models in education settings, showing that brand image is the mediator by which marketing action exerts relationship with behavioral intentions for high-involvement choices. The substantial mediating effects of the brand contest alternative direct effect models and confirm that brand perception is the essential cognitive mechanism governing educational decision-making behavior. Such multi-faceted mediation of brand identity through functional, emotional, and reputational dimensions empowers education branding theory by introducing how the marketing mix elements relate to separate dimensions of the brand, with product quality primarily driving functional perceptions and promotional activities impacting emotional and reputational dimensions.

The research also contributes to educational marketing theory through the empirical validation of content-orientated service models with more focus on educational outcome provision compared to traditional service process attributes. The empirical marketing mix effects hierarchy demonstrates that educational services require marketing models that reflect their unique attributes as experience goods with long-term outcome orientations, contrary to traditional educational marketing approaches where emphasis is placed on institutional infrastructure and interpersonal relationships.

6.2. Managerial Implications

Thai MOOC administrators should implement a strategic plan based on the high effect size and practical significance of product enhancement as the primary predictor for enrollment intention, as indicated by our empirical findings.

Product enhancement represents the first strategic priority. This involves three key activities. First, conduct quarterly learner demand analysis in target groups. Second, establish partnership agreements with specialized organizations through standardized quality criteria. Third, organize design-thinking workshops with course development teams and create detailed learner personas to guide course customization. This product-centered approach reflects our finding that course quality and range serve as the key brand differentiators in government-funded learning platforms.

Promotion optimization constitutes the second strategic priority. Administrators should develop an integrated social media plan across Facebook, Instagram, TikTok, and YouTube with platform-specific content calendars featuring learner testimonials and success stories. The promotional approach must focus on education outcomes and career development benefits rather than traditional advertising appeals. Our findings demonstrate that credibility-based promotion significantly influences brand image perception. Campaign activation must involve collaborations with education influencers and subject matter experts to endorse courses, supported by real-time engagement monitoring dashboards to assess impact within various demographic segments.

Process simplification forms the third implementation priority. User experience must be optimized to address the moderate effect achieved in our research. Administrators must reduce registration processes to a maximum of three clicks with social media integration. They should implement twenty-four-hour chatbot support with human escalation for complex questions. The process improvement strategy should include learner profile-driven course recommendation systems and quarterly user testing with immediate feedback incorporation. This enables continuous improvement in platform usability and learner satisfaction.

Platform accessibility (Place) enhancement comprises the fourth strategic imperative. This focuses on digital accessibility and multi-device optimization. Administrators must provide complete mobile responsiveness with native iOS and Android applications while adding Thai-English multilingual functionality for greater demographic coverage. The accessibility project must include achieving full compliance certification and monitoring geographic usage trends to optimize rural participation. Platform accessibility plays a significant role in favorable brand image perception through greater convenience and learning engagement capacity.

This strategic plan addresses the significant marketing mix elements identified in our study. Perceived value, people, and physical evidence demonstrated low practical importance in the Thai MOOC environment. Effect size hierarchy must inform implementation to improve resource use effectiveness and enrollment returns. Product development receives the highest priority, followed by promotion optimization, process streamlining, and accessibility upgrading.

7. Limitations and Future Research

7.1. Limitations of the Study

The results of this research should be considered within the framework of some important limitations. The foremost limitation arises from the sampling technique, which drew participants from Thai MOOC’s social media platforms. This method incurred a self-selection bias toward participants who were already interested in and positively inclined toward the platform, so the sample is not representative of the general Thai population, and the results are not generalizable. This bias is likely to influence the inferences of the study by possibly overestimating the size of the observed relationships and may account for the non-significance of dimensions such as ‘Perceived Value’ and ‘People,’ which might be more prominent in the case of a less-motivated, general population. Additional limitations are the use of self-reported information, which subjects the research to the possible influence of CMB despite procedural remedies, and the cross-sectional nature of the design, which provides associations at a single time point and does not explain absolute association.

7.2. Future Research

Future research should pursue several complementary directions to continue the exploration of MOOC enrollment behavior. Firstly, comparative analyses of university-affiliated and standalone MOOCs would provide model generalizability to different institutional contexts. Secondly, the theoretical model can be expanded by investigating other mediating variables such as perceived risk, social influence, and technology acceptance, or moderating variables such as digital literacy levels, cultural orientation, and socioeconomic status. Third, qualitative inquiry methods, including in-depth interviews and focus groups, would provide deeper understanding of learner decision-making, affective bonds with MOOC sites, and the nuanced forces behind enrollment intention that may not be captured in quantitative models. Fourth, longitudinal investigations tracing learners from initial awareness through course completion could inform the dynamic nature of brand perception and how it evolves across the learning experience. Fifth, cross-cultural validation studies across various Southeast Asian nations would increase the model’s generalizability within the region and explain culture-specific variables. Lastly, the use of mixed methods combining quantitative modelling with ethnographic research may uncover contextual factors related to the introduction of MOOCs under different socio-economic and geographical conditions.

8. Conclusions

This study provides empirical validation of service marketing and brand perception frameworks within government-sponsored MOOC contexts, addressing gaps in understanding how conventional marketing models apply to mission-driven, publicly funded digital learning platforms in Southeast Asian contexts. The results of this research indicate that product, promotion, process, and place dimensions significantly explain Thai MOOC enrollment intention through the mediation of brand image perception as the cognitive mechanism underlying government-sponsored online learning platforms. The research contributes to service marketing theory by outlining boundary conditions under which traditional 7Ps models must be modified in the digital education context. The research validates a three-dimensional brand image model comprising functional, emotional, and reputational attributes. The model includes empirically validated measurement instruments that are open to replication. This conceptual framework has significant practical implications for Thailand’s strategic MOOC development and for both domestic and international MOOC platforms. The results propose product quality as the most essential strategic lever, followed by promotion efforts, process refinement, and access improvement, thereby creating evidence-based propositions for making marketing strategy development and resource allocation decisions for the developing online learning ecosystem.