Keywords

1 Introduction

Information and communication technologies have developed exponentially in recent years, becoming one of the most important resources for society. This phenomenon has led to a boom in data transmission, exchange and knowledge [32]. Information and communication technologies (ICT) have revolutionized the means of communication since the early ’90s. Since then, the Internet has moved from a specialized instrument used by the scientific community to an easy-to-use network that has transformed social interaction [7]. One of the information technologies that has witnessed the largest growth in this period is streaming, which consists in data compression and distribution of multimedia content (audio and video) by means of a continuous, real-time flow directed to a user’s laptop [35]. The explosive development of streaming has caused several changes in entertainment companies. In 2000, 69 million Americans subscribed to cable television or satellite services; today, that number stands at 49 million [34]. In addition, between 2015 and 2016, the number of TV series with original scripts produced by online streaming services increased substantially from 46 to 93 [19], marking the peak of this industry in recent years.

Netflix is a leading company in the streaming field. With more than 100 million subscribers all over the world, this service has been recognized as one of the most innovative companies in the industry. The latter characteristic is critical for success in the current complex and changing environment, since the survival of companies depends on their adaptive capacity [18, 30]. In this context, we wonder what factors lead people to watch content on Netflix.

2 Literature Review

Several studies have focused on the intention to adopt a particular information system. Due to the nature of Netflix, we conducted a review of the studies that used UTAUT2 or extensions of the same. Factors were identified that influenced the adoption of mostly hedonic information technologies in order to determine in which consumption and entertainment contexts the model was applied, the variations of the same and how these tests may or may not be applicable to this study.

Helkkula [22] carried out research on the intention to subscribe a music streaming service. In his study, the author extended the UTAUT2 model by adding the variable tangibility preference, which refers to the physical properties of the product and the extent to which it can be seen, felt, heard and smelled, among others [14]. The results of this model may be valuable for other innovative and highly hedonic industries like the video game industry. In addition, Baabdullah [5] researched the intention to adopt games in mobile social networks (M-SNG). In this case, the model included the variable trust, which is related to intention to use and, at the same time, is determined by the variables hedonic motivation and social influence. Vinnik [39] studied the adoption of mobile applications, incorporating the variables herd behavior and online rankings and reviews into the model. To understand the reasons that people tag photographs in social networks, Dhir [11] added variables based on the social cognitive theory [6] to the model proposed by UTAUT2. This way, the model comprises the variables social presence, social status and self-efficacy.

These models confirmed that the UTAUT2 model can be used in different hedonic contexts with high reliability and adding single variables related to the context of the study. Therefore, the UTAUT2 model can be applied to the streaming industry. We decided to apply the model developed by Baabdullah [5], as it includes the variable trust. This variable is relevant to prepaid media streaming service.

3 Model and Hypothesis

The extension of the model selected for this research is based on the study conducted by Baabdullah [5], which incorporates the construct trust (TR) into the UTAUT2 model. Trust is a latent variable that within information systems (IS) refers to the perception that one can trust another person [40]. In e-commerce, trust refers to the confidence that clients will be provided with the desired benefits and facilities under safe and trustworthy conditions [16]. According to Gefen [17] trust in a seller is associated with greater behavioral intention (BI), which is crucial for a paid subscription service like Netflix. Additionally, the relationship between TR and BI has not been studied in streaming services. Considering all of this, the model proposed for this study is shown in Fig. 1.

Fig. 1.
figure 1

(Source: own elaboration)

UTAUT2 model extension applied to Netflix

Full size image

3.1 Performance Expectancy (PE)

Performance Expectancy is defined as the degree at which the use of a technology will help consumers perform certain activities [38]. Consumers seem to be more motivated to use new technologies if they perceive that these are more useful in their daily lives [3, 10, 37]. Several studies have determined the importance of this construct out of the workplace [1, 5, 20]. From this, we believe that:

H1: PE has a positive effect on intention to use Netflix.

3.2 Effort Expectancy (EE)

Effort Expectancy is defined as the degree of perceived ease of use of the technology by the individual [38]. The behavioral intention of people varies if they feel that using a specific service is much easier and more convenient [5]. A number of studies have shown the influence of this construct on the intention to adopt new technologies [2, 25, 31]. Therefore, we propose the following hypothesis:

H2: EE has a positive impact on the intention to use Netflix.

3.3 Hedonic Motivation (HM)

HM is the fun or pleasure derived from the use of a technology. This variable has been proven to play an important role in acceptance and use of technology [38]. Netflix is a service classified as highly hedonic because its purpose is to entertain consumers who subscribe to its content. Studies on streaming have demonstrated that entertainment (closely associated with hedonic motivation) is related to intention to use [9, 22]. Therefore:

H3: HM has a positive effect on the intention to use Netflix.

HM increases people’s trust in using this information technology. When individuals are highly motivated by hedonic factors, the trust in using this technology grows [3]. In online purchase systems, when consumers discover that buying can make them experience enjoyment and usefulness, they start to trust and adopt the online purchase system [16], which can be extrapolated to a paid subscription service. Therefore, we believe that:

H3’: HM increases the role of trust of Chilean customers in using Netflix.

3.4 Trust (TR)

This construct refers to the perception inherent to humans of being able to trust another person [40]. Mayer [29] defines it as the will to be in a vulnerable state based on positive expectations of the future behavior of another person. Studies about online services have shown that trust affects intention to use [4, 13, 21, 23]. As a paid subscription service, Netflix needs users to trust that the company is acting in good faith. Therefore, the following hypothesis is proposed:

H4: TR has a positive impact on intention to use Netflix.

3.5 Social Influence (SI)

Social Influence is defined as the extent to which individuals perceive that other people important to them believe they should use the new information system [37]. This construct is also denominated as a social norm in models such as TAM, TAM2 and TRA.

Social influence proved to be significant in music streaming, games and mobile purchase services [12, 24, 42]. This way, we believe that:

H5: SI has a positive effect on intention to use Netflix.

SI plays an invaluable role in the level of trust on a service. Therefore, through different opinions, potential users could voice their intention to use or not a service [27]. When people know their peers and the society prefer to user a technology like Netflix, people tend to believe that the use of such a technology can bring them benefits and similar values. Research has already indicated the importance of social influence to trust [5, 33]. Consequently, it is believed that:

H5’: SI influences trust in Chilean Netflix users.

3.6 Facilitating Conditions (FC)

Facilitating Conditions are defined as the extent to which individuals believe that there is an organizational and technical infrastructure that supports the use of the system [37]. In a consumer context, this variable can be defined as the perceptions users have of the resources and support available to perform a specific behavior [38].

Netflix requires some technological elements such as a laptop, mobile phone or Smart TV, as well as an Internet connection with enough speed for the platform to work properly. Users with access to a favorable set of facilitating conditions will have greater intention to use a new technology, as demonstrated by Zhou [43]. Therefore, it is believed that:

H6: FC has a positive impact on intention to use Netflix.

3.7 Price Value (PV)

Price Value is the cognitive calculation users make between the benefits perceived from an information system and the monetary cost of using it [38]. Price value is positive when the benefits of using an information system are perceived as higher than its monetary cost, which makes users more enthusiastic about adopting a new technology [38]. Netflix is a service for which consumers pay. Therefore, PV is expected to have a significant impact on the selection of the service. According to Venkatesh [38], individual consumers are more sensitive to price than people who use a service paid for by their company, because the cost of the new technology is paid by the same consumer. Several studies on technologies used outside of the workplace have pointed to the importance of price value for intention to use [28, 41]. Therefore, we believe that:

H7: PV has a positive impact on the intention to use Netflix.

4 Methodology

In this study, a confirmatory approach for the structural theory was undertaken. The survey applied to Netflix use in Chile was an adaptation of the questionnaire created by Baabdullah [5]. The instrument was validated by an exploratory text in which 15 people were first surveyed and asked for feedback on the questionnaire. This way, ambiguous questions were removed, while the wording of others was changed for clarity. Subsequently, an expert gave a final opinion about the duration, simplicity and clarity of language the language used in the questionnaire. A 7-point Likert scale was used, in which scores ranged from totally agree (value 1) to totally disagree (value 7) (See Table 1).

Table 1. Survey used in this study (Source: own elaboration based on Baabdullah [5])
Full size table

Afterward, the questionnaire was published through SurveyMonkey on different social networks. Therefore, sampling was by convenience. Four hundred and fifteen surveys were filled in, all of them 100%. In addition, a demographic information section was included, with questions regarding age and gender, as well as questions about the frequency of use and consumption of Netflix.

The software Statistics v24 and SPSS Amos v24 were used to obtain the statistics and conduct the confirmatory analysis by SEM. Additionally, a reliability analysis was conducted using Cronbach’s alpha.

5 Results

5.1 Profile and Characteristics of Respondents

Out of the 415 people surveyed, 53.5% were men and 47.5% were women. The age of the respondents ranged from 15 to 65 years, with a mean of 24 years. The age group 21–29 made up 75.9% of participants, followed by people under 20 years of age, who represented 15.4%, and by the 30–39 age group with 7.0%. In the last place were people 40 or over, who made up 1.7% of the sample.

Regarding the characteristics of the use of Netflix mentioned by participants, 98.6% has used Netflix before, while 87.2% currently use it. As for frequency of use, most participants use Netflix several times per week (35.2%), followed by those who use the service once per week (21.7%). Additionally, most respondents use the service for 1 to 2 h per time of use (43.9%), and 26.0% for 2 to 4 h per time of use.

5.2 Scale Reliability

To ensure the reliability of the model’s constructs, Cronbach’s alpha was analyzed. According to Loewnthal [26], the limit value for this indicator is 0.6, as these constructs have fewer than 10 items. The results are presented in Table 2.

Table 2. Cronbach’s Alpha by construct (Source: own elaboration)
Full size table

As can be seen, all the constructs have a Cronbach’s Alpha value above the cutoff point (0.6), which ranges between 0.610 (FC) and 0.906 (SI). Therefore, the constructs are above the acceptance limit and variables do not need to be eliminated.

5.3 Confirmatory Analysis Results

The results of the model regressions indicate that the \( R^{2} \) of intention to use is 0.493; that is, 49.3% of variance in the errors of this latent endogenous variable is explained by the latent exogenous variables.

Table 3 indicates that 396 degrees of freedom are obtained. This means that the model is over identified and (i.e. more equations than unknown parameters). Therefore, there is no exact solution and more than one set of parameter estimates is possible.

Table 3. SEM degrees of freedom (Source: own elaboration)
Full size table

5.4 Standardized Regression Coefficients

Table 4 shows the standardized regression coefficients of intention to use and its associated exogenous variables. The significance of these relationships is also provided.

Table 4. Standardized regression coefficients and significance for each latent variable (Source: own elaboration)
Full size table

The results indicate that the factors for trust (TR), performance expectancy (PE) and hedonic motivation (HM) are significant in the prediction of the intention to use Netflix, since they have a p-value lower than 0.05. In contrast, the factors for social influence (SI), facilitating conditions (FC) and price value (PV) were not significant. Additionally, social influence (SI) and hedonic motivation (HM) were found to be latent variables that affect trust (TR) significantly. These results are of the utmost importance for the conclusions of this study, as they show which are the variables that predict behavioral intention (BI) with acceptable significance. Thus, hypotheses H1, H3, H3’, H4 and H5’ are accepted, while hypotheses H2, H5, H6 and H7 are rejected. The results of the structural equation model can be seen in Fig. 2.

Fig. 2.
figure 2

(Source: own elaboration)

Results of the structural equation model

Full size image

Goodness-of-Fit Statistics

Below, Table 5 shows the final results of the goodness-of-fit statistics as well as the criteria considered.

Table 5. Goodness-of-fit statistics of final model (Source: own elaboration)
Full size table

The results of the fit indexes indicate that CMIN was 1650.824 with a significance below 0.05, which equals an acceptable model. The other statistics are not within the respective criteria for good adjustment of the model: RMSEA, CFI, NFI, GFI and CMIN/DF are below the acceptance limit. Based on these results, a new model will be proposed for assessment on a new sample. This model will be presented in Sect. 7.

6 Conclusions

In this study, we conducted different analyses to determine the factors that influence the adoption of paid subscription streaming services by Chilean users. The results indicate that the latent variables trust (TR), performance expectancy (PE) and hedonic motivation (HM) are significant for the behavioral intention (BI) of Netflix. Additionally, the latent variables hedonic motivation (HM) and social influence (SI) affect trust (TR) significantly.

The performance expectancy associated with Netflix is the factor that most determines the intention to use this service (0.423). This expectancy is related to the possibility of watching multimedia content faster, as well as to the personal perception that Netflix is useful for everyday life. Considering that Chile is a country with high stress rates in the population [8, 15], the need for distractive and hedonic elements is very high. At the same time, these elements need to be fast to avoid lost or dead time, which is often associated with loading or downloading times for multimedia elements. As such, a fundamental aspect for the intention to use Netflix is the avoidance of elements that delay entertainment; that is, the perceived speed of the service. Another relevant aspect is the importance a person gives to audiovisual media.

The second factor most significant for the intention to use Netflix is trust (TR) (0,325); that is, the perception of the honesty, reliability and quality of the service is crucial for the intention to use this service in Chile. Being a paid service that requires the use of a credit card to subscribe, Netflix needs a high perception of trust for users to be willing to provide their personal information to use the service, especially taking into account the sometimes-unreliable credit card security systems.

Hedonic motivation is the third factor that directly affects the intention to use Netflix (0.281). Therefore, for people to use Netflix, they need to find the service entertaining, fun or amusing, which reassures the hedonic nature of this streaming service. Other studies about hedonic information systems have already demonstrated this phenomenon [22, 36]. This indicates that the content of a streaming service influences the intention of users to subscribe to the same, because if the service catalogue does not satisfy the requirements of users, they will not feel entertained when they access the platform.

In addition, we determined that hedonic motivation influence trust significantly. In other words, when a service is perceived as entertaining, this translates into an increase in the trust of users. Consequently, when users feel comfortable using a service, their trust in the same will grow. This point highlights the importance of digital content within a streaming service, since hedonic motivation not only directly impacts intention to use, but also trust, which is the second major predictor of intention to use.

7 Recommendations

From the results, we can see that the variables performance expectancy, trust, hedonic motivation and social influence were fundamental for the intention to subscribe to Netflix in the Chilean context. This means that Netflix and other companies from the streaming sector should focus on having higher service speed, as speed is directly associated with performance expectancy. Therefore, brands need to be positioned as high-performance services targeting the functions the user needs. In advertisement, it should be underscored that Netflix and other streaming service are much faster than buying, renting or downloading a movie, and that these services allow users to enjoy multimedia content without having to move from home to go to the cinema. Additionally, streaming platforms should emphasize that content can be watched without interruptions, even if the Internet speed of the user is slow.

Another fundamental aspect to bear in mind is the content of the streaming service, which directly and indirectly affects intention to use through hedonic motivation. It is necessary that advertisement reinforces the concept of fun when using the application. At the same time, companies need to continue creating their own content to give users personalized content based on their preferences and needs, creating different libraries for series, movies and documentaries to satisfy the needs of all user segments. Advertising the awards won for original content goes in this same line.

A third point is that companies should invest in security systems to protect user data, since trust directly influences intention to use. Streaming companies should constantly improve their security systems in order to provide maximum confidence to the people who trust them with their personal data. Thus, companies need to effectively communicate their efforts to maintain cybersecurity, as well as to implement high level post-sales service that allows users to reach them quickly and efficiently in case of any questions or problems with the service or payment.

Based on the fit indicators of the model and the significance of factors that affect the intention to use Netflix, we propose a new research model. This model does not include the latent variables whose regression coefficients were non-significant. Consequently, the constructs effort expectancy, facilitating conditions and price value are removed from the relationship between SI and BI.

In the model proposed (Fig. 3), social influence not only affects trust but also influences hedonic motivation and performance expectancy. Furthermore, trust determines the performance expectations of Netflix, because users who trust the company tend to think the service will have better performance than users who do not trust it. Thus, the model proposed is the following:

Fig. 3.
figure 3

(Source: own elaboration)

Model proposed for new study on the intention to use Netflix

Full size image

The new model does not consider the observable variables with regression coefficients below the cut-off value (0.5). Therefore, TR6, PE1 and PE2 are not part of this model. The survey proposed for this new model is presented in Table 6.

Table 6. Survey proposed for new study on the intention to use Netflix (Source: own elaboration based on Baabdullah [5])
Full size table

Another aspect to consider is that although the UTAUT2 model focuses on consumption, its questionnaire is associated to the observable variables and therefore does not completely fit the context of streaming and Netflix. This is because Netflix is a service exclusively devoted to entertainment and some questions do not match 100% what users expect from this type of service. Thus, we propose to adjust these limitations in a new study by putting forward a new acceptance model for new entertainment and streaming technologies.

8 Limitations of the Study

The main objective of this study was to apply a quantitative approach to determine the factors that affect the intention to use Netflix. This could have restricted the capacity of the study to analyze more carefully the issues related to the behavior of Chilean Netflix customers. Therefore, using a mixed method (quantitative and qualitative) could provide a more detailed explanation of the results of this study, specifically regarding the reasons behind non-significant relationships. Another important aspect is the possible existence of bias in the selection of the sample, mostly because the sampling method is non-probabilistic. In addition to this reason, since people had no incentive to answer the survey, they might have responded quickly and without complete awareness of their own answers.