Antecedents of Electronic Word of Mouth (eWOM) Adoption in the Purchase of Cosmetics in Ecuador: Does Gender Moderate Relationships?

Abstract

Social networks have emerged as a powerful tool for communication and marketing, enabling two-way user participation and fostering a conducive environment for electronic word of mouth (eWOM). This study examines how social influence, eWOM engagement, and perceived eWOM credibility affect its adoption, while also exploring the moderating role of gender in these relationships. Using a sample of 371 cosmetics consumers in Ecuador, a causal model was estimated using the PLS method. The findings confirm that social influence significantly impacts eWOM engagement and credibility, contributing to its adoption. Additionally, gender moderates the relationship between informational influence and eWOM credibility. This study also highlights the need to replicate the model in different contexts and with more diverse samples to enhance the generalizability of the findings. By enriching Erkan and Evans’ IACM theoretical framework, this research extends the understanding of eWOM dynamics in the cosmetics sector. Moreover, it offers a comprehensive perspective on the eWOM adoption process by incorporating underexplored variables and providing empirical evidence from an understudied region, such as Latin America.

1. Introduction

In recent years, the exponential growth of social network usage has strengthened its role as a powerful marketing and communication tool. In 2024, there were 5.78 billion active users, representing approximately 70.5% of the world’s population, with a daily average screen time of 6 h and 38 min [1]. This growth, along with the platforms’ innovative features that enable efficient dissemination of products and services [2], has created new business opportunities, resulting in a significant increase in social media marketing investment, which reached nearly 720 billion dollars that year [1].

The COVID-19 pandemic further accelerated this trend in countries such as Ecuador, where the number of active social media users grew by 16.7% since 2021, reaching 83.7% of the national population by the end of 2024 [1]. This expansion has opened new opportunities for local companies across sectors, including cosmetics, which has experienced 10% annual growth in recent years, generating over 1.1 billion dollars annually [3].

In the research context, the growing relevance of social networks has stimulated extensive studies on eWOM behavior from both the sender’s and receiver’s perspectives. From the sender’s viewpoint, recent studies have examined how individuals generate and disseminate eWOM, focusing on factors such as motivation and content virality [4], the role of brand communities in shaping eWOM messages [5], and how social and cultural dynamics influence message diffusion [6]. On the receiver’s side, research has explored how users process and evaluate eWOM messages, with studies analyzing cognitive and emotional responses to online reviews [7,8], the moderating effects of trust and source credibility in eWOM adoption [9] and the way constructs such as eWOM credibility, perceived usefulness, and psychological distance collectively shape consumer perception and acceptance in digital service environments [10].

Despite the growing body of research on eWOM, there remains a critical gap in understanding how consumers adopt and act upon this information, particularly from the receiver’s perspective. Although participatory and two-way communication platforms offer fertile ground for exploring the eWOM adoption process and its influence on purchasing decisions, research focusing specifically on the receiver’s perspective remains limited. This study addresses this gap by examining key constructs such as social influence, perceived eWOM credibility, and consumer engagement, which play a crucial role in the adoption process.

While these constructs have been widely acknowledged in eWOM research, their specific role in the pre-purchase consultation phase remains insufficiently explored [11,12,13,14]. Most studies on the social influence and perceived credibility of eWOM focus on digital platforms in general, without delving into the specific context of social networks [15,16]. Studies on eWOM engagement are also limited because the concept is relatively new [13,14]. Moreover, these constructs have yet to be examined collectively within a unified framework, highlighting a critical gap in understanding their interplay and influence on consumer decision-making. By integrating these elements, this research provides a more holistic understanding of the eWOM adoption process in social network environments, while also emphasizing the importance of theoretical integration to explain consumer behavior in increasingly complex digital ecosystems [17,18].

Likewise, the literature recognizes the importance of the applicability of research in different contexts, considering that eWOM behavior can vary depending on cultural and social factors [19,20,21,22]. In fact, recent studies highlight significant differences in review behaviors based on the generation gap and gender [23,24]. However, most of the existing research has been conducted in Western and Asian markets, leaving Latin America as a largely unexplored region in this field [25,26,27,28]. There is still a lack of empirical studies in the Latin American context and, specifically, research that closely examines disparities in eWOM consultation behaviors among various demographic segments. Although it has been addressed in several studies, gender remains an area with inconclusive results that vary considerably between different cultural contexts. Some studies have found differences between men and women in information processing and usage behavior on digital platforms [24,29,30], while others do not highlight any significant differences between genders [26]. This discrepancy underscores the need for further research in diverse cultural settings, particularly in Latin America, where digital consumer behavior may differ due to socio-economic and market-specific factors.

As an emerging digital economy, Ecuador presents a unique opportunity to analyze these dynamics [1,3]. Considering the interest in studying the influence that reviews have on purchasing decisions and the literature gap regarding the role of gender, the aim of this study is to investigate the eWOM adoption process within the Ecuadorian market, to determine which variables and relationships explain the effect that eWOM queries have on cosmetic purchasing behaviors. To achieve this, a dual objective is pursued. On the one hand, to analyze the relationships between social influence, eWOM engagement, perceived credibility of eWOM, and adoption of eWOM, differentiating the two dimensions of social influence recognized in the literature: normative and informational [16]; and on the other hand, to study the moderating effect that gender has on these relationships to find out if there are differences in the eWOM adoption process between men and women.

This work contributes to existing literature in several ways. Firstly, it offers a context-specific perspective on the eWOM adoption process in Ecuador, a Latin American market that remains underrepresented in academic discussions. By focusing on the cosmetics sector, an industry where gender differences in consumer behavior are particularly pronounced, this study provides valuable insights into the role of gender in eWOM adoption. Secondly, this study provides empirical support for recently investigated variables in eWOM behaviors, such as perceived credibility and engagement, in an environment that is conducive to understanding their impact, such as social networks. This responds to current calls for more research linking engagement and trust-based variables in localized contexts [31,32].

Finally, unlike most research that predominantly examines eWOM from the sender’s perspective or within broad digital platform contexts, this study focuses specifically on the receiver’s perspective in social networks, which offer unique functionalities for user interaction. Furthermore, it seeks to enrich the literature by providing evidence on gender differences in eWOM adoption behavior in a region where such studies remain scarce [26,27,32]. The findings of this research are expected to be applicable in similar cultural contexts and industries.

2. Theoretical Background

2.1. Social Influence on eWOM Behavior

Social influence plays a fundamental role in individual behavior, where cognitive, affective, and behavioral attitudes are shaped by social demands [33]. Consequently, numerous marketing studies have used this variable as a key concept to explain consumer behaviors. In 1950, Leibenstein introduced the concept of the “bandwagon effect” to describe how individuals’ decisions, including purchasing choices, are influenced by others [34]. Although several theories have attempted to explain this phenomenon, Deutsch and Gerard’s [35] dual-process theory is one of the most frequently used to examine social influence from an eWOM perspective. This theory distinguishes between two types of social influence: normative, related to the desire for approval, and informational, referring to the acceptance of information as evidence of reality.

Normative influence, also known as approval-based compliance [36], relates to the consumer’s desire to belong. Through observation of the behavior of others, the individual tends to create judgements about a product or service that shapes their consumption behavior [37]. In digital environments, particularly on social media, this influence manifests through visible social approval mechanisms such as likes, shares, and comments [13,38]. These mechanisms function as social cues that signal group norms and reinforce conformity, especially among users seeking social validation [11,36]. In the cosmetics sector, where appearance and self-presentation are often linked to identity and peer perception, normative influence is amplified through visual content, influencer endorsement, and platform-specific beauty standards [39,40,41].

Informational influence, by contrast, is based on the willingness to accept others’ experiences and opinions as reflections of reality. Consumers often view peer insights as valuable sources of information for making decisions [42]. In digital settings, informational influence is intensified by the persistent availability of reviews and the credibility cues embedded in social platforms, such as reviewer profile visibility, follower count, or engagement history [33,42,43]. In the cosmetics category, where product trial is subjective and results are often individualized, informational eWOM (e.g., user reviews, before-and-after videos, or skin-type-specific recommendations) becomes a crucial mechanism for building trust and guiding decisions [41,44,45].

Both dimensions of social influence impact eWOM behavior since consumers rely on the opinions and recommendations of others with whom they feel connected to form judgements about products [46]. In fact, much of the research surrounding online reviews is based on Sussman and Siegal’s [47] information adoption model (IAM), which is presented as a combination of the technology acceptance model (TAM) and the elaboration likelihood model (ELM) to explain how individuals adopt and act upon persuasive information in mediated environments. Informational influence, by contrast, is based on the willingness to accept others’ experiences and opinions as reflections of reality. Consumers often view peer insights as valuable sources of information for making decisions [44,46].

2.2. eWOM Engagement

The rise of digital channels as dominant communication tools has intensified scholarly interest in online interactions [48,49]. Despite the recent boom, the concept of engagement has its foundations in two areas of marketing: relationship marketing, which proposes the link with the customer as the central axis of the commercial relationship to ensure sustained long-term demand [50]; and service-dominant logic, which is based on creating relationships with consumers through proposals that generate value [51].

Hollebeek’s [52] ideas have been widely used in recent research to address the construct of engagement [13,14,38]. These hypotheses highlight the multidimensionality of the concept by defining it as a process that implies a bidirectional and dynamic relationship where both the entity or object of the engagement and the engaged individual contribute to the interaction process. Furthermore, Hollebeek identifies three key dimensions of engagement: affective, cognitive, and behavioral. The affective dimension of engagement is based on the emotional attachment and connection that the individual feels towards the brand or product; the cognitive dimension implies that the consumer has made a rational and conscious evaluation of the object of the engagement; while the behavioral dimension refers to tangible behavior that demonstrates the individual’s engagement in the form of actions [52].

In digital contexts, platform functionalities such as likes, comments, reposts, and hashtags have significantly expanded the possibilities for expressing engagement. These features support the emergence of new patterns of participation that blur the boundaries between content consumption and content creation. This evolution has led scholars to refine the concept of eWOM engagement, understood as the willingness to seek, share, or interact with peer-generated product information in digital environments [13].

Within social networks, user engagement behaviors can be classified into two distinct types: passive and active [53]. Passive engagement includes actions such as viewing, scrolling, or reading product reviews without interacting, whereas active engagement involves more participatory behaviors such as commenting, tagging others, reposting brand-related content, or generating original posts related to personal product experiences [53,54].

In the cosmetics sector, this distinction is particularly relevant, as consumers frequently use visual and interactive elements to express their experiences with products. For example, tagging a brand in a “makeup of the day” post, sharing unboxing videos, or creating skincare routines on platforms like Instagram or TikTok are clear manifestations of active engagement [38,55]. Meanwhile, browsing product tutorials or reading comments without reacting represents passive engagement. Both forms contribute to the dissemination and reception of eWOM, but they differ in their level of visibility and potential influence.

Following Hollebeek’s dimensional approach, several studies have shown that specific social media behavior, such as posting content, taking screenshots, tagging friends, or reacting to brand communications, can be mapped onto the three engagement dimensions [43]. However, to truly understand the impact of these actions on decision-making, it is essential to distinguish both their intensity and nature, particularly for experience-based products like cosmetics.

2.3. Perceived Credibility of eWOM

Perceived credibility is the consumer’s assessment of a message’s truthfulness and trustworthiness [15]. This perception is based on two key dimensions: source credibility, which concerns the perceived trustworthiness and expertise of the message sender, and content credibility, which relates to the clarity, coherence, and accuracy of the message itself [56].

In the context of eWOM, perceived credibility plays a pivotal role, where purchasing decisions are highly influenced by subjective user experiences. Consumers often rely on peer-generated reviews and testimonials to assess product effectiveness and safety [17,23]. Moreover, perceived credibility serves as a direct antecedent of eWOM adoption, a concept that is further elaborated upon in the following section.

2.4. eWOM Adoption

eWOM adoption behavior is defined as the process of internalizing the message that can trigger an action on the part of the consumer [57]. This construct has been used by various studies on eWOM behavior, encompassing the complex process that involves the individual’s assimilation, interpretation, evaluation, and reaction to the information received [7,9,11,46].

To explain how individuals adopt eWOM, this study draws on the information acceptance model (IACM) proposed by Erkan and Evans in 2016 [18]. The IACM is particularly suitable for social media contexts, as it expands upon the traditional IAM by incorporating not only message quality and source credibility but also the receiver’s attitude toward the information and the social dynamics of digital platforms [18]. This broader framework is especially relevant for eWOM behavior in fast-paced, emotionally charged, and highly visual environments like social media [7,58].

Unlike the IAM, which focuses mainly on an objective evaluation of message attributes, the IACM accounts for the influence of users’ attitudes, motivations, and contextual interactions. This makes it a more appropriate theoretical lens for the cosmetics sector, where identity expression, visual appeal, and peer validation are integral to how consumers engage with and adopt product-related information [9,59].

Social media platforms actively facilitate eWOM adoption by offering features that encourage both information-seeking and interaction, such as search functions, comment sections, hashtags, and recommendation algorithms. Consumers who intentionally seek out product-related content are more likely to encounter relevant eWOM and assess it as useful to their purchase decisions [17].

Therefore, a holistic approach that considers both the structural availability of information and the user’s attitudinal disposition is essential for understanding how eWOM is adopted [60]. The IACM framework supports this perspective, offering a robust theoretical basis for analyzing how consumers in social networks process peer-shared content and integrate it into their decision-making. This understanding is also critical for the development of targeted and effective online marketing strategies in visually intensive, experience-based sectors like cosmetics [31,40].

3. Research Model and Hypotheses

3.1. Effect of Normative Social Influence on eWOM Engagement

Social networks allow individuals to identify with certain groups that share similar beliefs and norms [11]. This connection is further enhanced by platform algorithms, which promote interaction among like-minded users and foster participation in virtual communities [61].

Along these lines, normative social influence maintains that the individual’s behavior is influenced by the social norms and expectations of their reference groups, and that their objective is to fit in with the intrinsically required standards [35]. In the context of eWOM on social networks, individuals show a greater willingness to participate in discussions and share information when they perceive that these actions are socially acceptable and valued by their online community [62]. Studies suggest that normative influence in digital settings drives behavior through the desire for acceptance and a cost-benefit evaluation of conformity [38,63].

Participation in social network reviews is aligned with various dimensions of engagement [64] as platform features expand engagement beyond mere information consultation to include actions such as reacting, saving, or commenting [38]. Both the nature and intensity of these interactions affect users’ levels of eWOM engagement [15].

In this regard, Gvili and Levy (2018) emphasize that digital social capital, shaped by peer influence and shared community values, plays a key role in eWOM participation [15]. Similarly, Mushtaq and Ahmad (2016) note that normative social pressure strongly encourages users to share and engage with brand-related content in pursuit of online validation [65]. Furthermore, Cheung et al. (2014) demonstrate that social conformity significantly influences the perceived relevance and trustworthiness of eWOM content, which in turn fosters engagement behaviors [66].

According to these contributions, group norms become an essential component of social capital by influencing user behavior and fostering a sense of belonging among members [67]. Considering that eWOM engagement is closely related to many of the consumer attitudes that develop based on these norms, it is assumed that this type of social influence will contribute to consumers’ levels of eWOM engagement. Therefore, the following research hypothesis is formulated:

Hypothesis 1 (H1). 

Normative social influence positively influences eWOM engagement.

3.2. Effect of Informational Social Influence on eWOM Credibility

Informational social influence refers to the way in which individuals take the experiences and opinions of others as evidence of reality and, based on this information, construct their own value judgements [35]. On social networks, this behavior is amplified by the amount of information that is shared. Consumers actively seek information to optimize their decision-making processes, which implies minimizing efforts and reducing costs associated with the search for alternative purchases [62].

Perceived eWOM credibility reflects how trustworthy, authentic, and realistic consumers find others’ recommendations [15]. This perception of credibility can be influenced by various factors such as the reputation of the source, the consistency of the message with the consumer’s previous experience, or the number of endorsements the recommendation receives [68]. When consumers consider the recommendations found in eWOM queries to be credible, they are more likely to trust them and use them as a guide to construct their ideas and shape their behavior [33].

Additional evidence highlights that informational cues such as argument quality and message clarity substantially influence perceptions of credibility in online reviews [66]. Moreover, source expertise and message consistency have been identified as primary antecedents of eWOM trust, reinforcing the role of informational influence in shaping consumer perceptions [67]. Consumers’ reliance on peer-shared information also increases when they perceive a high level of informational richness and objectivity in the review content [17].

Based on these works and considering that informational influence can lead to a review being considered more credible if others support it or consider it reliable, we consider that this type of social influence will increase the consumer’s perception of eWOM credibility, and, therefore, put forward the following hypothesis:

Hypothesis 2 (H2).

Informational social influence positively influences perceived eWOM credibility.

3.3. Effect of Engagement on eWOM Adoption

Social networks also promote two-way interaction between users through their various functionalities [14]. In addition to allowing consumers to generate reviews, these platforms facilitate engagement around this content. Therefore, social networks transcend the action of searching for information, since they provide users with the ability to interact with messages in a more dynamic way, allowing them to comment, react, share, or save the information received [38].

This interaction is aligned with different engagement dimensions [15]. For example, the behavior of seeking or sharing information is related to both the cognitive and emotional dimensions of engagement [14] while actions such as saving or commenting are more aligned with the behavioral dimension [38]. Furthermore, active interaction with content, such as reacting or tagging, reinforces these dimensions by facilitating a deeper and sustained connection with the information and the digital community [13].

On the other hand, in the process of adopting a message posted on social networks, the consumer’s attitude towards the information plays a crucial role in the acceptance or rejection of the message [18].

The depth of engagement, whether emotional or behavioral, is directly linked to information processing intensity, which affects subsequent adoption behavior [52]. Furthermore, the willingness to engage with peer content has been shown to predict eWOM influence on purchase intent, establishing engagement as a reliable precursor to adoption in social media contexts [43].

Given that eWOM engagement reflects the degree of involvement and attitude of the consumer towards reviews, we consider that this engagement will contribute to the level of eWOM adoption, raising the following hypothesis:

Hypothesis 3 (H3).

eWOM engagement positively influences eWOM adoption.

3.4. Effect of Perceived Credibility on eWOM Adoption

Perceived credibility plays a key role in the reception of eWOM, as consumers typically embrace recommendations that they consider credible, truthful, and objective [68]. This perception extends to the credibility of the message, both the source and the channel. These two aspects are key in the persuasion and acceptance of eWOM, i.e., in how eWOM information affects the purchase decision [67,69].

In this sense, the information acceptance model (IACM) by Erkan and Evans highlights the importance of perceived credibility as a crucial element in the information acceptance process. This model highlights that credibility not only affects the way in which consumers process information on a deeper level via the central route but also influences how they interact with superficial signals via the peripheral route when evaluating the authenticity and reliability of the message [70]. In this peripheral route, the perception of experience and the reliability of the source are considered as external aspects of the message that are crucial in the formation of opinions and the adoption of information [23].

Considering that perceived credibility stands out as a relevant attribute in the information acceptance process [68], it is assumed that in the context of eWOM behavior, this credibility will significantly influence eWOM adoption, leading to the following hypothesis:

Hypothesis 4 (H4).

Perceived credibility of eWOM positively influences eWOM adoption.

3.5. Moderating Effect of Gender in the eWOM Adoption Process

Recent studies highlight significant gender differences in online information processing and eWOM query behaviors [19,24,71]. Along these lines, factors such as the perception of credibility, the influence of the social component, and participation patterns may vary according to gender.

For example, it has been found that women are more likely to trust opinions and personal experiences shared online, while men may place more importance on objective information [24]. Likewise, eWOM consultations are observed to have a more pronounced impact on men’s willingness to repeat a service than in the case of women [19]. Additionally, social norms and gender expectations can influence individuals’ online behavior. Specifically, men may be more susceptible to social norms in the eWOM acceptance process, while for women, informational influence has a more significant weighting when evaluating reviews [71].

However, it is important to consider that the role of gender may vary depending on sociocultural patterns, the specific contexts of the sector of analysis, and the generation gap [24,26]. There is no agreement in the literature on this issue. While [26] did not find gender differences in the eWOM adoption process, other research has observed important differences. For example, research in Asia indicates that men are more likely to experiment with products or services recommended through eWOM, even if it involves some risk [72]. In contrast, studies in Africa suggest that women show greater receptivity towards eWOM adoption and are more susceptible to its influence on their online purchasing decisions [73]. Research in Europe highlights that the impact of eWOM recommendations made by women increases as the level of gender equality in society increases [22].

Along these lines, a recent study conducted in several U.S. cities identified clear gender-based differences in digital information processing [24]. This is consistent with the findings of Kwahk and Kim [33], who observed that gender significantly influences how individuals evaluate and adopt online information. Specifically, both studies highlight that men and women differ not only in their cognitive strategies when interpreting eWOM but also in the weight they assign to social and informational cues during the decision-making process.

Furthermore, differences in risk perception between genders are also corroborated by prior work. For instance, women consistently express greater concern about transaction security in digital environments [30], a pattern that complements our findings by illustrating how trust dynamics shape gendered engagement with online reviews. Additionally, recent research indicates that the moderating role of gender in eWOM behaviors may be influenced by individual-level variables, such as personality traits, suggesting that gender-based effects are not homogeneous but rather contingent upon broader psychological and contextual dimensions [74]. These cultural and psychological nuances have also been documented in regional studies on digital consumption behavior [75], reinforcing the need to consider context-specific factors when analyzing gender-based differences in eWOM adoption.

Considering this empirical evidence, it becomes theoretically and practically relevant to explore how gender moderates the eWOM adoption process in specific sectors and cultural contexts [19,22,62]. In particular, the cosmetics industry in Ecuador offers a unique setting where gendered consumption patterns, identity expression, and social media usage converge. In such environments, the mechanisms through which social influence, perceived credibility, and engagement shape adoption behavior may operate differently between men and women [31,55]. Moreover, platforms such as Instagram and TikTok, which are highly visual and interactive, may intensify affective or social dimensions of engagement, which could resonate differently across gender segments.

These theoretical considerations are directly linked to the conceptual model proposed in this study, which integrates normative and informational influence as antecedents of eWOM engagement and perceived credibility, leading to adoption. We posit that gender may moderate these paths by influencing the degree to which men and women engage with eWOM behaviors.

Accordingly, the following hypotheses are proposed regarding the moderating role of gender in the model:

Hypothesis 5 (H5).

Gender moderates the following:

Hypothesis 5a (H5a).

The effect of normative social influence on eWOM engagement;

Hypothesis 5b (H5b).

The effect of informational social influence on perceived eWOM credibility;

Hypothesis 5c (H5c).

The effect of eWOM engagement on eWOM adoption;

Hypothesis 5d (H5d).

The effect of perceived eWOM credibility on eWOM adoption.

Figure 1 presents the research model utilized in this study, offering a visual representation of the main constructs and hypotheses. This model serves as a structured framework for analyzing the relationships under investigation.

4. Methodology and Results

4.1. Data Collection and Sample

To evaluate the proposed model, a quantitative study was conducted using a self-administered online survey based on a structured questionnaire. The target population is comprised of men and women residing in Ecuador, aged 16 to 65, who are regular consumers of cosmetic products and active users of social networks. The study is exempt from formal ethics approval under Article 43 of ACUERDO No. 00005-2022, issued by the Ecuadorian Ministry of Public Health.

Data was collected through non-probability snowball sampling via social media platforms (WhatsApp, Facebook, and Instagram). This approach was selected due to the difficulty of obtaining a probabilistic sample of social media users with an interest in cosmetics. While snowball sampling enabled access to participants who actively engage with eWOM, it also introduces limitations regarding generalizability and potential sampling biases, such as the overrepresentation of certain social circles and the underrepresentation of less digitally active users. These limitations are acknowledged, and caution is advised when interpreting the findings more broadly.

4.2. Instrument and Measurement Scales

To measure the constructs, previously validated scales from earlier studies were adapted to align with the specific context of this research. A 7-point Likert scale was employed, as it offers greater sensitivity and variance in responses compared to shorter scales, thereby enhancing measurement reliability [76]. This choice also ensures consistency with existing eWOM research, facilitating cross-study comparisons. Response options ranged from 1 = strongly disagree to 7 = strongly agree [9,16,26,38].

Normative and informational social influence was measured using the scale developed by Bearden [77], which remains widely utilized in contemporary literature due to its demonstrated reliability and validity [32,42]. The construct of eWOM engagement was assessed using the scale proposed by Hollebeek [64], while perceived eWOM credibility was measured with the instrument developed by Cheung [70]. Finally, eWOM adoption was evaluated using the scales proposed by Fang [63] and Yan [58].

Given the reliance on self-reported data, common method bias (CMB) was assessed to mitigate potential distortions in the findings. Harman’s single-factor test was conducted, confirming that no single factor accounted for most of the variance, indicating that CMB was not a significant concern [78]. Additionally, procedural remedies, including respondent anonymity and variation in question wording, were implemented to reduce response bias.

A total of 371 valid questionnaires were collected for analysis. In adapting the constructs to the Ecuadorian context, minor linguistic and cultural adjustments were made to enhance clarity and relevance, particularly for items related to consumer behavior and online engagement. The sociodemographic profile of the sample, along with its purchasing behavior and digital engagement characteristics, is presented in

Table 1. Sample profile.

Gender		Education		Occupation
Female	65%	Without studies	0.4%	Employee	20.8%
Male	35%	Elementary	3%	Self-employed	13.5%
Age		Secondary/college	27.5%	Housewife	7.2%
16–20	26.7%	University	60.2%	Student	56.4%
21–25	36.4%	Postgraduate studies	8.9%	Unemployed	2.1%
26–30	16.5%	SM accounts (1)		SM most used (2)
31–35	5.1%	Facebook	95%	Facebook	74%
36–40	3.8%	Instagram	74%	Instagram	56%
41–45	3.1%	LinkedIn	14%	LinkedIn	1.3%
More than 41	8.4%	Twitter	27%	Twitter	2.7%
Frequency consultation		TikTok	42%	TikTok	5.9%
Very often	23.3%	YouTube	51%	Daily SM use
Quite often	24.2%	Frequency purchase		<1 h	22.5%
Sometimes	38.5%	<frequently than others	42%	2–4 h	52.1%
Sporadically	11%	Same as others	43%	4–6 h	15.7%
Never	3%	>frequently than others	15%	>6 h	9.7%

(1) and (2) are multi-answer questions.

. Most respondents belong to the millennial and centennial generations, ranging in age from 16 to 40 [79]. These cohorts demonstrate high digital engagement, frequently consulting product reviews across multiple social media platforms, with Facebook, Instagram, and TikTok being the most popular.

In the measurement model, the retention of low-loading items was supported by both theoretical and empirical considerations. Items with factor loadings below the standard 0.5 threshold were reviewed and kept when they contributed to the construct’s conceptual validity [80]. Sensitivity analyses were performed to ensure that their retention did not compromise overall model reliability or validity.

4.3. Reliability and Validity of Measurement Scales

To test the hypotheses, a structural equation model was applied using the partial least squares method with the SmartPLS4 software. Most items exhibit significant factor loadings, although two items related to normative influence and one to informational influence did not reach the minimum required threshold. Despite this, we decided to keep them to guarantee content validity, considering their statistical significance in the analysis. The reliability of the scales was evaluated through the Cronbach’s alpha coefficient and composite reliability (CR). It is notable that the α-Cronbach values in all constructs exceed the recommended threshold of 0.7 [81], ranging between 0.83 and 0.93, which suggests a solid internal consistency of the items used in each construct. CR values are between 0.83 and 0.93, exceeding the recommended cut-off point of 0.7 [82]. Furthermore, the average variance extracted (AVE) values are within the appropriate range, between 0.65 and 0.7, which supports the convergent validity of the measurement instrument (

Table 2. Measurement model.

Construct	Indicators (First Order)	Load	T	α	CR	AVE
Normative influence (NI)	NI1	0.711	10.248	0.884	0.884	0.562
	NI2	0.741	12.561
	NI3	0.690	13.562
	NI4	0.678	10.253
	NI5	0.878	18.106
	NI6	0.782	13.184
Informational influence (II)	II1	0.784	13.211	0.837	0.836	0.562
	II2	0.669	14.469
	II3	0.747	20.147
	II4	0.791	17.696
eWOM engagement (EE)	EE1	0.849	30.976	0.933	0.933	0.666
	EE2	0.767	23.767
	EE3	0.864	34.322
	EE4	0.793	25.030
	EE5	0.853	30.633
	EE6	0.821	28.895
	EE7	0.697	18.075
Perceived credibilityof eWOM (PC)	PC1	0.769	25.023	0.911	0.910	0.718
	PC2	0.794	27.064
	PC3	0.757	24.219
	PC4	0.860	27.214
	PC5	0.840	26.757
	PC6	0.878	23.610
	PC7	0.805	22.038
eWOM adoption (EA)	EA1	0.822	40.184	0.929	0.929	0.653
	EA2	0.858	45.081
	EA3	0.829	43.435
	EA4	0.879	41.516

α = Cronbach’s alpha; CR = composite reliability; AVE = average variance extracted.

).

To evaluate discriminant validity, the Fornell–Larcker criterion [83] and the heterotrait-monotrait (HTMT) index ratio were used. The square root of each construct’s AVE exceeded its correlations with other constructs, and HTMT index values remain consistently below the suggested threshold of 0.9, confirming the discriminant validity of the measurement scales used in the study (

Table 3. Fornell–Larcker criterion and HTMT. First-order indicators.

Constructs	1	2	3	4	5
1. II	0.749	0.599	0.631	0.490	0.636
2. NI	0.600	0.750	0.578	0.461	0.556
3. PC	0.635	0.576	0.816	0.759	0.847
4. EE	0.492	0.558	0.757	0.808	0.837
5. EA	0.636	0.460	0.848	0.836	0.847

).

4.4. Structural Model Analysis and Multigroup Analysis

The R² values of 0.403, 0.809, and 0.312 were significant and suggest that the model has a good-to-moderate fit [84]. Specifically, the model explains 40.3% of the variance in perceived eWOM credibility, which represents a moderate explanatory power according to Falk and Miller’s [85] criteria and reflects the model’s effectiveness in capturing the influence of informational and social cues on credibility evaluation.

In the case of eWOM adoption, the R² value of 0.809 indicates an exceptionally high level of explained variance, which is uncommon in consumer behavior studies. This result suggests that the combination of perceived credibility and engagement, two constructs theoretically anchored in the IACM, is a particularly strong predictor of eWOM adoption in the cosmetics sector, where decisions are often social, visual, and experience-driven. Conversely, the 31.2% variance explained for eWOM engagement represents a modest but acceptable level, consistent with prior findings that position engagement as a complex, multidimensional construct influenced by both individual disposition and contextual platform factors [52,86].

To assess the risk of multicollinearity within the structural model, variance inflation factor (VIF) values were analyzed for all predictor constructs. All values ranged between 1.000 and 1.996, well below the commonly accepted threshold of 3.3 [87], indicating no multicollinearity concerns. In addition, bootstrapping with 5000 subsamples using the percentile method was applied to estimate the precision of the path coefficients. The 95% confidence intervals for all significant paths did not include zero, confirming their robustness and statistical reliability.

The relationships between the variables present significant connections (

Table 4. Structural model results.

HYP	Relationships	Path Coefficient	T	p	f2	Result
H1	NI → EE	0.558	13.073	0.000	0.675	Supported
H2	II → PC	0.319	6.739	0.000	0.675	Supported
H3	EE → EA	0.456	7.351	0.000	0.564	Supported
H4	PC → EA	0.502	7.679	0.000	0.464	Supported

). On the one hand, normative influence presents a positive and significant impact on eWOM engagement, suggesting that opinions and social expectations influence eWOM engagement (p < 0.001, path coefficient of 0.558). This influence is complemented by the significant relationship between informational influence and the perception of eWOM credibility, highlighting how the search for information influences trust in this source of information (p < 0.001, path coefficient of 0.319).

Furthermore, it is observed that eWOM engagement is positively related to the adoption of this behavior, indicating that those more involved in eWOM communication are more likely to adopt the recommendations (p < 0.001, path coefficient of 0.456). This trend is reinforced by the significant relationship between perceived eWOM credibility and eWOM adoption, showing that trust in the shared information influences the decision to adopt the recommendations (p < 0.001, path coefficient of 0.502).

Before carrying out the multigroup analysis, configural invariance, compositional invariance, and, finally, equal composite mean values and variances through Measurement Invariance of Composites (MICOM) were applied (

Table 5. Findings of permutation-based invariance measurement testing.

Construct	Step 1	Step 2		Partial Measurement Invariance?	Step 3 (a)			Step 3 (b)			Full Measurement Invariance?
	Configural Invariance	Compositional Invariance			Equal Variances?			Equal Means?
		Original Correlation	Confidence Interval		Diff.	Confidence Interval	Equal?	Difference	Confidence Interval	Equal?
NI	Yes	0.996	0.996. 1.000	Yes	−0.002	−0.204. 0.187	Yes	−0.003	−0.248. 0.231	Yes	Yes
II	Yes	0.998	0.992. 1.000	Yes	0.004	−0.211. 0.221	Yes	0.005	−0.253. 0.256	Yes	Yes
PC	Yes	1.000	1.000. 1.000	Yes	−0.001	−0.202. 0.205	Yes	0.008	−0.288. 0.309	Yes	Yes
EE	Yes	1.000	0.999. 1.000	Yes	−0.004	−0.205. 0.205	Yes	0.009	−0.264. 0.268	Yes	Yes
EA	Yes	1.000	1.000. 1.000	Yes	−0.003	−0.215. 0.199	Yes	0.003	−0.260. 0.283	Yes	Yes

). Overall, the findings indicate that the constructs are adequately measured and comparable across groups, with the majority achieving full measurement invariance. This result makes it possible to carry out significant comparisons of means and variances between groups for the examined constructs.

Table 6. Multigroup analysis estimation (gender).

HYP	Relationships	Path Coefficient FEMALES	Path Coefficient MALES	Difference Paths	Henseler MGA	p-Value	Results
H5a	NI → EE	0.548	0.618	−0.07	−0.071	0.811	Rejected
H5b	II → PC	0.723	0.510	0.213	0.213	0.011	Supported
H5c	EE → EA	0.423	0.517	−0.094	−0.094	0.758	Rejected
H5d	PV → EA	0.545	0.427	0.118	0.118	0.199	Rejected

presents the results of the multigroup analysis (MGA) based on gender. Regarding the effects of social influence, on the one hand, it can be observed that there is a significant relationship between normative influence and eWOM engagement in both men and women, with path coefficients of 0.548 and 0.618, respectively. However, the difference between genders does not achieve statistical significance. On the other hand, the relationship between informational influence and perceived eWOM credibility is significant for women (0.723), but rejected for men (0.510), with a considerable difference between genders (0.213).

Regarding the effects of engagement and credibility on the adoption of eWOM, the path coefficients are 0.423 and 0.545 for women and 0.517 and 0.427 for men. These coefficients indicate the strength of the relationship between eWOM engagement and eWOM adoption, and between perceived eWOM credibility and eWOM adoption. However, the p-values (0.758 and 0.199, respectively) indicate that these relationships are not statistically significant and are therefore rejected.

5. Discussion and Implications

5.1. Implications for Theory

Social influence, eWOM engagement, and perceived credibility of social network reviews emerge as key factors shaping how eWOM queries influence cosmetic purchasing decisions in Ecuador. Previous studies in tourism, fashion, and tech have already highlighted the relevance of these variables in digital settings [26,31]. The current context fosters the formation of virtual communities that share values and opinions and allows for immediate consumer exposure to peer reviews. This environment encourages various forms of participation closely linked to eWOM engagement [13,14,38,88].

The results of this study confirm all the hypotheses related to the proposed eWOM adoption model, offering empirical support within the underexplored context of eWOM behavior from the receiver’s perspective in Latin America. In line with prior literature, social influence was found to significantly impact both eWOM engagement and the perception of review credibility, supporting the idea that subjective norms, particularly those tied to cultural identity, influence users’ engagement with reviews [15,65]. The act of consulting others’ experiences aligns with the IACM framework [18], especially in environments where platforms allow users to observe others’ reactions to content [23,66].

Additionally, Boldureanu et al. (2025) have shown that eWOM credibility, perceived usefulness, and psychological proximity significantly influence consumer perception and acceptance, reinforcing the explanatory power of these variables in digital purchasing decisions [10].

Moreover, the results confirm that individuals who engage more deeply with eWOM content and perceive reviews as credible are more likely to adopt recommendations, a pattern consistent with tourism and hospitality studies [14,23]. This underscores the importance of user engagement and message trustworthiness as central drivers in the adoption process.

In terms of gender, this study uncovers partial differences. Informational influence significantly affects the perception of eWOM credibility in women, consistent with prior research showing that women tend to rely more on informational cues when evaluating message trustworthiness [71]. However, other gender-based differences were not statistically significant, challenging prior results. For example, earlier work suggests men are more affected by normative influence [19] while women tend to engage more actively [20]. The limited gender differences in this study may reflect Ecuador’s shifting norms, potentially reducing behavioral gaps [24].

Theoretically, this study not only supports but extends the IACM in three key ways. First, it validates IACM in a Latin American context, where little empirical testing has been performed, thus contributing to geographic and cultural generalizability. Second, it introduces eWOM engagement as a precursor to adoption, a construct not originally included in the IACM framework but theoretically aligned with its underlying premises about user interaction and perceived value. Third, it applies the model from the receiver’s perspective, which is often overlooked in favor of the sender or content-centric perspectives. This extension is particularly relevant in social network environments, where users are not passive recipients but active co-creators of meaning and value in review ecosystems

5.2. Managerial Implications

This study offers several actionable insights for marketing professionals in the cosmetics sector operating in Latin America, particularly in Ecuador. The results highlight the strategic relevance of electronic word-of-mouth (eWOM) as a central component of digital communication strategies. Rather than treating eWOM as an organic by-product of customer behavior, companies should proactively integrate it into their content, influencers, and media planning cycles to strengthen brand engagement and purchase intention.

5.2.1. Activate Social Influence Through Micro-Influencers

The significant impact of social influence on both eWOM engagement and credibility supports the strategic use of paid eWOM tactics via micro-influencers. These individuals cultivate authentic relationships within specific audience segments, fostering trust and enhancing message diffusion [89].

When targeting female consumers and younger cohorts, especially those identified in this study as more susceptible to informational influence, brands should prioritize visual and short-form content platforms such as TikTok or Instagram reels. These environments are more conducive to affective engagement and virality, amplifying the impact of influencer endorsements.

5.2.2. Design Platform-Specific Content Strategies

Given that eWOM engagement significantly predicts adoption, brands should tailor content strategies based on platform dynamics and consumer behavior patterns. Promoted reviews and sponsored user-generated content) can be effective tools to enhance visibility and trust, particularly when supported by algorithmic amplification and audience segmentation tools.

In Ecuador, where digital penetration is high among millennials and centennials but uneven across sociodemographic groups, customizing content by audience segment and platform type is essential for optimizing reach and impact.

5.2.3. Implement Robust eWOM Management and Listening Protocols

The strong effect of perceived credibility on eWOM adoption underscores the need for structured eWOM governance. Brands should develop and enforce review management protocols to monitor, amplify, or mitigate the influence of user-generated reviews [9].

This includes investment in social listening tools and AI-driven sentiment analysis [90], enabling real-time monitoring of consumer feedback, swift resolution of negative experiences, and identification of emerging trends that shape purchase behavior.

For SMEs and local beauty brands, deploying agile customer response workflows via platforms such as WhatsApp Business or Instagram Direct Messaging can enhance trust, demonstrate responsiveness, and build long-term customer relationships in high-contact industries like cosmetics.

5.3. Future Lines of Research

While this study contributes to the literature on eWOM behavior from the receiver’s perspective, it has several limitations that open promising directions for future research. First, the sample is demographically skewed toward millennial and centennial generations, segments known for their high digital literacy and frequent social media use [91]. Future studies should explicitly compare generational cohorts to assess whether the observed relationships hold across age groups, particularly among older consumers who may process eWOM differently.

Second, this research did not account for valence, i.e., whether the reviews analyzed were positive, negative, or neutral. Given that prior work has shown consumers are more motivated to post negative reviews [92,93] and that review valence influences trust and decision-making [94], future studies could adopt an experimental approach to manipulate review polarity and observe its impact on eWOM credibility and adoption.

Third, although eWOM engagement is part of the proposed model, it was treated as a unidimensional construct. Future research could expand this by integrating the three subdimensions—affective, cognitive, and behavioral engagement [52]—to assess their individual contributions to adoption. This would help disentangle whether emotional connection or rational evaluation plays a stronger role in different contexts (e.g., luxury vs. functional cosmetics).

Fourth, platform-specific variables such as the frequency of use or type of social media platform (e.g., visual vs. text-based, ephemeral vs. persistent content) may act as moderators in the eWOM process [95]. For instance, consumers who use Instagram daily may process reviews differently than those who engage primarily on Facebook [5]. Segmenting the analysis by platform could reveal whether engagement or credibility functions differently across channels, offering a more granular understanding of the digital purchase journey.

Finally, considering Ecuador’s specific sociocultural characteristics, future research should explore how broader factors, such as economic background, education, or urban vs. rural residence, interact with gender and digital behavior in shaping eWOM adoption patterns. This could be approached through mixed-method designs or cross-country comparisons within the Latin American region.