PARADOKS Jurnal Ilmu Ekonomi Vol. 9 No. 1, 2026 e-ISSN : 2622-6383 The Trust Paradox: A Dual-Pathway Analysis of Digital Marketing's Impacts on Consumer Loyalty and Consumer Decision Ernando R. Supatman1*. Ronny Hansye Walean2. Deske W. Mandagi3 *Email Correspondence: ernandosupatman2024@gmail. Management Marketing. Faculty of Economic Business. Universitas Klabat. Indonesia *1,2,3 Abstract In contemporary digital markets, cultivating sustained customer loyalty and shaping purchasing decisions present significant strategic challenges for companies. This study investigates the impact of digital marketing on consumer decisions and loyalty in the financing industry in Indonesia, comparing the mediating mechanisms of brand engagement . ffective pathwa. and consumer trust . ognitive pat. Using an explanatory quantitative approach on 265 respondents selected through purposive sampling, the data were analyzed with Structural Equation Modeling (SEM). The findings reveal that digital marketing significantly influences consumer decisions and loyalty, with brand engagement acting as a partial mediator, while consumer trust does not mediate significantly, uncovering a "trust paradox" in the context of the financing industry. This study makes an original contribution by introducing a dual-pathway model that distinguishes between affective and cognitive mediation mechanisms, and by highlighting the practical implication that relational strategies grounded in emotional attachment are more effective than transactional trust-based approaches in fostering sustainable loyalty. Keywords: Digital marketing. consumer loyalty. brand attachment. consumer trust. This work is licensed under a Creative Commons Attribution 4. 0 International License. Introduction In the competitive digital era, building consumer loyalty and influencing purchasing decisions have become critical challenges for companies. Consumer loyalty today extends beyond product quality. It is shaped by accumulated experiences and sustained emotional Factors such as price, promotions, ease of information access, and direct experiences with products and services significantly influence consumer loyalty (Sundstrym & Lidholm, 2. In saturated markets, consumers tend to prefer established brands as a risk mitigation strategy when faced with abundant alternatives (Xiao et al. , 2. These dynamics compel companies to seek strategies that not only attract new consumers but also retain them over the long term. Within this context, digital marketing has emerged as a vital strategic instrument. enables firms to engage directly with consumers, deliver information swiftly and relevantly, and create added value that differentiates them from competitors (Rangaswamy et al. This approach has proven effective not only in enhancing engagement and cost efficiency but also in shaping preferences and facilitating consumer decision-making (Klein & Sharma, 2022. Mashrur & Mohamed, 2. Empirical evidence demonstrates that social media and digital engagement can significantly influence brand attitudes and loyalty across sectors, including education (Kainde & Mandagi, 2. , tourism (Mandagi et al. , and the fast-food industry (Ole et al. , 2. However, the mechanisms through which digital marketing transforms one-time decisions into enduring loyalty remain insufficiently Paradoks: Jurnal Ilmu Ekonomi 9. | 67 Two psychological mechanisms . , brand attachment and consumer trus. are believed to serve as key mediators linking digital marketing to consumer loyalty. Effective digital marketing can foster strong emotional bonds between consumers and brands, increasing the likelihood of repeat purchases (Wei et al. , 2022. Centeno & Mandagi, 2. At the same time, consumer trust serves as the foundation of every business relationship. the uncertain digital environment, trust helps reduce consumersAo perceived risks (Appiah & Agblewornu, 2. This trust develops through consistent, transparent, and cooperative interactions and has been recognized as a crucial determinant of consumer loyalty (Firoozzare et al. , 2024. Xuanze et al. , 2024. Walean et al. , 2. However, not all digital marketing efforts successfully cultivate these elements. For instance, strategies that rely solely on price promotions often fail to establish emotional bonds (Ugalde & Kyster, 2. From this background, a research gap becomes apparent. Although prior studies have confirmed the direct effects of digital marketing on consumer decision-making (Aji et , 2. and customer loyalty (Al-Dmour et al. , 2023. Mohammad, 2. , few have analyzed the simultaneous mediating roles of brand attachment and consumer trust in strengthening these relationships, particularly within the financial services industry. Previous studies in other contexts (Centeno & Mandagi, 2022. Mandagi et al. , 2024. Kainde & Mandagi, 2023. Walean et al. , 2. have highlighted the importance of emotional attachment and trust as psychological links between marketing activities and loyalty Therefore, this study aims to examine the direct and indirect effects of digital marketing on consumer loyalty through brand attachment and consumer trust. The financial services sector is experiencing a paradigm shift characterized by intense competition and increasingly digital-savvy consumers. In this setting, two critical drivers of sustainable growth are influencing initial consumer decisions and fostering long-term loyalty. Loyalty in financial services transcends mere repeat purchases. it reflects a deep psychological commitment, manifested in advocacy and resistance to competitive persuasion (Latif et al. , 2023. Srisusilawati et al. , 2. However, achieving such loyalty is particularly challenging given the high-involvement nature of financial decisions, which are influenced not only by rational evaluation but also by complex psychological, social, and emotional factors (Solomon, 2019. Song et al. , 2. The intangible nature of financial products amplifies perceptions of risk, underscoring the importance of trust and emotional Although digital marketing is widely recognized as a key enabler in this environment (Faruk et al. , 2. , substantial gaps persist in understanding the psychological mechanisms that mediate its influence on consumer outcomes. Digital marketing has evolved into a comprehensive ecosystem encompassing search engine optimization, targeted advertising, content marketing, and social media engagement (Faruk et al. , 2021. Gensler & Rangaswamy, 2. While its role in improving visibility and accessibility is well established (Enshassi et al. , 2. , recent studies emphasize that the effectiveness of digital marketing depends on the type of engagement it elicits (Ruiz et al. , 2. Informational digital marketing, such as educational content and product comparisons, appeals to cognitive processing and fosters trust, whereas relational digital marketing builds emotional attachment. These distinct approaches demonstrate that digital strategies can activate different psychological pathways that contribute to longterm loyalty (Kainde & Mandagi, 2023. Centeno & Mandagi, 2. To move beyond a simple direct-effect model, this study introduces two mediating variables . , consumer trust and brand attachmen. that represent the psychological mechanisms transforming digital marketing stimuli into long-term relationships. In high-risk contexts like financial services, trust forms the foundation of the brandAeconsumer It reflects consumersAo confidence in a providerAos reliability, integrity, and competence (Hanaysha, 2. , and is built through transparent communication, consistent service quality, and demonstrated expertise (Philander & Wimmer, 2. Trust functions as a risk-reduction mechanism. when consumers trust a brand, their perceived uncertainty diminishes (Lita & Meuthia, 2. , leading to positive decisions (Hasan & Sohail, 2. and Paradoks: Jurnal Ilmu Ekonomi 9. | 68 sustained loyalty (Yuen et al. , 2. Informational digital marketing, therefore, is particularly effective in fostering cognitive-based trust (Walean et al. , 2. Meanwhile, brand attachment refers to the emotional bond connecting consumers to a brand (Japutra et al. , 2. It extends beyond satisfaction, encompassing feelings of connection, passion, and identification. Such attachment develops through shared values, positive emotional experiences, and ongoing engagement (Barijan et al. , 2021. Centeno & Mandagi, 2. Firmly attached consumers remain loyal not out of habit but because of genuine affinity, often becoming brand advocates (Sun et al. , 2. Relational digital marketing strategies are especially effective in nurturing this affective connection (Kainde & Mandagi, 2023. Mandagi et al. , 2. The conceptual model of this study posits that digital marketing influences consumer decisions and loyalty through two distinct psychological routes: a cognitive pathway, mediated by trust, and an affective pathway, mediated by brand attachment. Existing research supports the direct impact of digital marketing on these outcomes, showing that digital platforms facilitate informed decisions by simplifying complex comparisons (Moslehpour et al. , 2. That personalized digital experiences enhance loyalty (Abdullah et al. , 2. Accordingly, this study hypothesizes: H1: Digital marketing has a positive and significant effect on consumer decisions. H2: Digital marketing has a positive and significant effect on consumer loyalty. Beyond these direct effects, the study proposes that consumer trust mediates the influence of digital marketing on consumer outcomes. Secure, informative digital interactions convey competence and integrity, fostering trust that reduces perceived risk and reinforces loyalty (Ebrahim, 2020. Pop et al. , 2022. Walean et al. , 2. H4: Consumer trust mediates the positive effect of digital marketing on consumer decisions. H6: Consumer trust mediates the positive effect of digital marketing on consumer loyalty. Simultaneously, brand attachment is expected to mediate the affective pathway. Relational digital marketing that evokes emotional resonance strengthens attachment, which in turn drives consumer decisions and long-term loyalty (Japutra et al. , 2022. Wu et , 2024. Centeno & Mandagi, 2. H3: Brand attachment mediates the positive effect of digital marketing on consumer H5: Brand attachment mediates the positive effect of digital marketing on consumer loyalty. Finally, this study incorporates hypotheses regarding the direct effects of the mediators on consumer outcomes, as supported by prior research: H7: Consumer trust positively and significantly affects consumer decisions. H8: Consumer trust positively and significantly affects consumer loyalty. H9: Brand attachment positively and significantly affects consumer decisions. H10: Brand attachment positively and significantly affects consumer loyalty. The influence between the variables can be observed in Figure 1, based on the ten hypotheses previously outlined: Paradoks: Jurnal Ilmu Ekonomi 9. | 69 Brand Attacment H10 Consumer Decision Digital Marketing Consumer Loyalty Consumer Trust Figure 1. Conceptual Mode Methodology Research Design This research uses a quantitative approach with an explanatory survey method. The purpose of the study is to test the causal relationships between variables in the proposed theoretical model. The descriptive analysis method is a research method used to systematically and accurately determine a phenomenon in a particular object. (Sugiyono. This design was chosen because it is suitable for testing the strength and direction of the relationships between exogenous variables (Digital Marketin. and endogenous variables (Consumer Decisions and Consumer Loyalt. , as well as the mediating roles of Brand Attachment and Consumer Trust. Primary data was collected through structured questionnaires and analyzed using Structural Equation Modeling (SEM) to confirm the developed theoretical model. Population and Sampling The target population in this study consists of adults in Indonesia who have utilized the services of finance companies offering pawn services for Buku Pemilik Kendaraan Bermotor (BPKB) or certificates. The inclusion criteria established are: . being at least 18 years old, . having previously used the services of a finance company with BPKB or certificate collateral, and . being willing to participate as respondents. Given the population is widespread and unlimited in size, a non-probability sampling technique, purposive sampling, was employed to recruit respondents who met specific criteria. (Ghozali, 2. The sample size was determined using the minimum sample size rule in SEM. The number of samples in this study was determined using a formula. Hair et al. In accordance with these guidelines, the present study employed the N x 5 formula, where N represents the total number of indicators . multiplied by . , yielding a minimum of 250 respondents. After screening, 265 respondents who met the criteria were further analyzed. Data Collection and Instrument Primary data were collected through online questionnaires distributed via the Google Forms platform and directly to respondents. The questionnaire consisted of two sections: Demographic data, covering gender, age, education, and residence. Intrinsic questions, which measured the five research variables. All items were measured Paradoks: Jurnal Ilmu Ekonomi 9. | 70 using a 5-point Likert scale . = Strongly Disagree to 5 = Strongly Agre. Data Analysis Technique Data analysis was conducted in two stages using SmartPLS 4. 0 software, in accordance with the Partial Least Squares Structural Equation Modeling (PLS-SEM) PLS-SEM was selected for its ability to model latent variables, making it suitable for predictive research aimed at theory development, and for its robustness to relatively small sample sizes. Stage 1: Assessment of the Measurement Model (Outer Mode. This stage evaluates the validity and reliability of the instruments. Convergent Validity: Evaluated based on a loading factor > 0. 7 and an Average Variance Extracted (AVE) > 0. 5 for each Reliability: Measured using Composite Reliability (CR) and Cronbach's Alpha with values > 0. 7, indicating good internal consistency. Discriminant Validity: Assessed using the Fornell-Larcker Criterion method, where the square root of a construct's AVE should be greater than its correlations with other constructs (Hair et al. , 2. Stage 2: Structural Model Assessment (Inner Mode. This stage tests the hypothesesAihypothesis Significance: Conducted using the bootstrapping procedure. Hypotheses are accepted if the p-value < Result and Discussion Demographic Profile of Respondents The sample consisted of 265 respondents who used a financing company's services and participated in this study. The demographic profile shows a balanced gender distribution . 8% femal. , with the majority of respondents aged 31-40 . 3%) and 41-50 . 7%). Most respondents are private employees . 3%) and civil servants . 9%), and have used the services 1-3 times . 7%). Characteristics Gender Ages times pawned Occupation Table 1. Profil Responden Category Male Female 21-30 Years 31-40 Years 41-50 Years Above 50 Years Under 21 Years 1-3 times 4-6 times 7-10 times Above 10 Times Laborers Lecturer/Teacher Housewife Private Owned Enterprise Employee Fishermen Civil Servant Farmer Motorcycle Taxi Driver Wiraswasta Frequency Percent (%) The measurement model shown in Figure 2 is evaluated to assess its validity and Convergent validity is analyzed through examination of factor loading and Average Variance Extracted (AVE) values. Paradoks: Jurnal Ilmu Ekonomi 9. | 71 Evaluation of Measurement Model Figure 2. Measurement Model Discriminant Validity Discriminant validity was tested using the Fornell-Larcker criterion by comparing the square root of the Average Variance Extracted (AVE) for each construct with the correlations between constructs. The results in Table 2 show that the square roots of the AVEs . iagonal value. for each construct are greater than the correlations with other constructs, thus meeting the discriminant validity criteria. Several indicators, namely CL4. CL5. CD2. CD5. DM3, and DM5, were removed to achieve this result. BA . CD . CL . CT . DM . BA . 0,858 0,733 0,804 0,842 0,747 Table 2. Fornell-Larcker Criterion CD . CL . CT . 0,877 0,718 0,762 0,813 0,878 0,746 0,737 0,845 0,838 DM . 0,840 Structural Model The structural model was tested to evaluate the research hypotheses. The significance of the path coefficients was analyzed using the bootstrapping procedure in SmartPLS. The results of the hypothesis testing are presented in Figure 3. Figure 3. Structural Model Paradoks: Jurnal Ilmu Ekonomi 9. | 72 The cross-loading analysis confirms the model's discriminant validity. As presented in Table 3, each indicator has the highest loading on its intended construct compared to other These results ensure that each measurement construct is unique and stands BA1 BA2 BA3 BA4 BA5 CD1 CD3 CD4 CL1 CL2 CL3 CT1 CT2 CT3 CT4 CT5 CT6 CT7 DM1 DM2 DM4 DM6 DM7 DM8 DM9 BA . 0,838 0,861 0,861 0,879 0,851 0,639 0,645 0,643 0,685 0,697 0,734 0,706 0,735 0,747 0,710 0,699 0,684 0,693 0,544 0,539 0,692 0,639 0,708 0,631 0,612 Table 3. Cross Loading CD . CL . 0,627 0,659 0,601 0,640 0,639 0,755 0,641 0,700 0,633 0,686 0,871 0,606 0,892 0,637 0,867 0,647 0,660 0,886 0,551 0,870 0,677 0,878 0,636 0,652 0,621 0,656 0,696 0,640 0,692 0,655 0,631 0,577 0,669 0,631 0,545 0,593 0,602 0,565 0,634 0,528 0,687 0,710 0,685 0,600 0,757 0,687 0,704 0,625 0,697 0,596 CT . 0,705 0,680 0,750 0,724 0,748 0,685 0,639 0,679 0,640 0,631 0,692 0,815 0,851 0,862 0,855 0,872 0,828 0,826 0,698 0,672 0,732 0,704 0,753 0,670 0,692 DM . 0,610 0,590 0,680 0,677 0,638 0,751 0,686 0,699 0,651 0,588 0,699 0,669 0,709 0,748 0,731 0,692 0,738 0,656 0,793 0,818 0,859 0,859 0,869 0,845 0,832 Reliability All constructs in the model satisfy both validity and reliability criteria. Cronbach's Alpha for all indicators is greater than 0. 70, the Average Variance Extracted is greater than 50, and Composite Reliability exceeds 0. 70, confirming the convergent validity and reliability of the measurement model. The results presented in Table 4 confirm that the measurement instrument used is consistent and reliable. Variable Table 4. Construct Reliability Composite AVE Cronbach's Alpha Reliability 0,910 0,933 0,736 0,850 0,909 0,769 0,852 0,910 0,771 0,933 0,946 0,713 0,930 0,944 0,705 Description Reliable Reliable Reliable Reliable Reliable Direct Effect The results of the hypothesis testing are presented in Table 5. The analysis shows that five out of seven direct relationships are statistically significant. Brand attachment positively and significantly affects consumer decision and consumer loyalty. Digital marketing also exerts a positive, significant influence on consumer decisions and loyalty. In contrast, consumer trust does not have a significant effect on either consumer decision or consumer Hypothesis analysis shows that digital marketing has a positive and significant effect on consumer decision-making . = 0. and on consumer loyalty . = 0. Brand attachment is also shown to influence consumer decision-making significantly . = 0. Paradoks: Jurnal Ilmu Ekonomi 9. | 73 and consumer loyalty . = 0. Conversely, consumer trust does not show a significant effect on consumer decision-making . = 0. nor on consumer loyalty . = 0. Table 5. Hypothesis Testing Results Standard T Statistics Deviation P Values (|O/STDEV|) (STDEV) 0,088 2,653 0,008** 0,105 5,351 0,000*** 0,098 1,032 0,302 0,115 0,199 0,842 0,096 5,740 0,000*** 0,087 3,440 0,001** BA i CD BA i CL CT i CD CT i CL DM i CD DM i CL Significant Yes Yes Yes Yes Notes: ***p<0,001, **p<0,005 Mediating Effect The mediation analysis results in Table 6 reveal that brand attachment significantly mediates the relationship between digital marketing and both consumer decision and consumer loyalty. Conversely, consumer trust does not show a significant mediating effect in the relationship between digital marketing and consumer decision, nor between digital marketing and consumer loyalty. Table 6. Mediating Effect Standard deviation (STDEV) T statistics (|O/STDEV|) DM i CT i CD 0,082 DM i CT i CL 0,097 0,198 DM i BA i CD 0,071 DM i BA i CL 0,087 Notes: ***p<0,001, **p<0,005 P values 0,303 0,843 0,013** 0,000*** Significant Yes Yes The results of the mediation analysis indicate that brand attachment significantly mediates the relationship between digital marketing and consumer decision-making . = . and consumer loyalty . = 0. Conversely, consumer trust does not exhibit a significant mediating effect in the relationships between digital marketing and consumer decision-making . = 0. or between digital marketing and consumer loyalty . = 0. Discussion The key findings of this study reveal two distinct psychological pathways through which the influence of digital marketing is transmitted: Strong Affective Path: The Crucial Role of Brand Attachment Research results demonstrate that brand attachment functions as a strong mediating mechanism. Digital marketing significantly fosters consumers' emotional attachment, which subsequently becomes a key driver for consumer decisions (H. and consumer loyalty (H. This is consistent with the theory that, in the context of complex financial services, emotional bonds serve as a strong psychological anchor guiding consumer behavior. (Japutra et al. , 2022. Sun et al. , 2. Relational and communitybuilding digital marketing strategies are effective in creating sustainable loyalty pathways. Malfunctioning Cognitive Pathways: The Paradox of Consumer Trust Surprisingly, consumer trust was not found to be a significant mediator (H4 & H6 rejecte. , nor did it have a direct effect on consumer decision (H. or consumer loyalty (H. These findings create a paradox in the Indonesian financial industry. Although theoretically, trust is considered a foundation. (Hanaysha, 2. In practice, consumers may view trust as a hygiene factorAia fundamental prerequisite expected from all Paradoks: Jurnal Ilmu Ekonomi 9. | 74 legitimate financing companies. Once this prerequisite is met, other factors, such as promotional attractiveness, interest rates, or the ease of competitors' processes, become more decisive in influencing final consumer decisions and consumer loyalty, even if trust in the current company has already been established. In other words, trust alone is not sufficient to "lock in" consumers within this highly competitive ecosystem. The Strong Direct Impact of Digital Marketing Digital marketing has been proven to have a substantial direct influence on both decision-making (H. and loyalty (H. This confirms previous findings. Abdullah et al. Indayani & Maulidiyah, 2. That visibility, information, and digital experiences directly shape consumer perceptions and intentions, irrespective of attachment or consumer trust. Theoretical and Managerial Implications This study contributes to theory by revealing the dual pathway model in the context of financial services, where the affective pathway . rand attachmen. is proven to be more crucial than the cognitive pathway . onsumer trus. From a managerial perspective, these findings suggest that financing companies should: . Shift from transactional to relational strategies, focus digital marketing efforts on building brand narratives, communities, and intense emotional experiences to foster brand attachment. Go beyond building trust, since trust is a given, competitive differentiation should be sought through service innovation, superior economic value, and seamless customer experiences, all of which should be communicated through digital channels. Conclusion This study concludes that digital marketing is a strong driver of consumer decisionmaking and loyalty in the financing industry. However, its main theoretical contribution lies in identifying affective pathways as more dominant compared to cognitive ones. Specifically, brand attachment serves as a substantial mediator, whereas consumer trust does not. This indicates that in a competitive financial services landscape. However, digital marketing provides an initial stimulus. it is the emotional connection and psychological bond of brand attachment that ultimately translates marketing efforts into sustained loyalty and repeat decisions. Trust, while a basic requirement, is not a sufficiently differentiating factor to secure customers. Therefore, the influence of digital marketing is most effectively channeled through strategies that build deep and emotional brand relationships, rather than merely focusing on establishing perceptual trust. This study offers distinct theoretical and practical contributions: . The Dual-Pathway Model represents the primary novelty of this study by empirically testing and validating a dual mediation model . rand attachment and consumer trus. within the context of financing companies. This goes beyond a simple direct effect model by illustrating the different psychological processes through which digital marketing operates. The Trust Paradox provides the most significant theoretical insight with the revelation of the nonsignificant role of trust as a mediator. These findings challenge conventional marketing theory, which positions consumer trust as the foundation of loyalty in financial services. They introduce a critical nuance: in a mature and competitive market, consumer trust may serve as a hygiene factor, while brand attachment emerges as a key motivational factor for This establishes new boundary conditions for the existing theory. This study strengthens the body of knowledge by applying and testing these relationships in the context of collateral-based financing services in Indonesia. This underexplored area offers valuable insights specific to this sector. Paradoks: Jurnal Ilmu Ekonomi 9. | 75 Further Practical Contributions of this study: . Strategic Reorientation for Managers: The findings provide clear and actionable guidance for marketing managers in the financing sector. Instead of allocating resources predominantly to trust-building campaigns, managers should shift to strategies designed to foster brand engagement. This includes leveraging digital platforms for storytelling, building communities, and creating personalized and emotionally resonant experiences. Resource Optimization: This helps practitioners optimize marketing budgets by highlighting that investments in relational and emotional content . uilding attachmen. yield higher returns in terms of loyalty and decision influence compared to investments that focus solely on demonstrating reliability and competence . uilding trus. This study has several limitations that open opportunities for future research. Geographic and Contextual Limitations: The data were collected from a single province in Indonesia, specifically from North Sulawesi. This restricts the generalization of findings across different cultural and regional contexts in Indonesia and beyond. Cross-Sectional Design: The use of a cross-sectional design captures perceptions at a single point in time, making it difficult to establish actual causal relationships or to observe the evolution of trust and engagement dynamics. Standard Method Bias: Although mitigated, the potential for common-method bias persists because all variables were collected from the same respondents using a single questionnaire. Given these limitations, future research should replicate the study across different industries to test the dual-pathway model in sectors such as fintech, banking, and ecommerce, and to validate the role of brand attachment relative to broader consumer Conducting longitudinal studies to track how the relationships among digital marketing, brand attachment, trust, and loyalty develop and change over time is Expanding the sampling framework to include a more diverse populationAi national or even internationalAiwould enhance the external validity of the findings. Integrating objective measurements to complement survey data with behavioral metrics, such as actual repeat purchase data and customer lifetime value, would strengthen the robustness of the results. Acknowledgment The authors would like to express their deepest gratitude to Prof. Ronny Walean for his guidance, support, and the research facilities provided throughout this study. Sincere appreciation is also extended to all colleagues and friends who assisted in distributing and collecting the research questionnaires, contributing significantly to the success of this work. The authors are also grateful for the valuable suggestions and constructive feedback that have greatly improved the quality of this manuscript. Furthermore, the authors wish to acknowledge Sir Deske W. Mandagi for his kind assistance and support during the publication process of this article. References