1 Introduction

Recent technological advancements, especially those related to the evolution (and ubiquity) of the internet (Bouncken and Kraus 2022), digitization (Nambisan 2017; Tagscherer and Carbon 2023), and digitalization (Nambisan et al. 2019; Kraus et al. 2019b, 2023; Tagscherer and Carbon 2023), have contributed to the emergence of digital businesses, new business models (Bouwman et al. 2019; Ghezzi and Cavallo 2020; Guo et al. 2022; Calderon-Monge and Ribeiro-Soriano 2023; Cavallo et al. 2023), digital entrepreneurial opportunities (Bouncken et al. 2021; Felicetti et al. 2023; Galindo-Martín et al. 2023), expanded entrepreneurial ecosystems (Muldoon et al. 2023), and digital business ecosystems (Calderon-Monge and Ribeiro-Soriano 2023; Fernandéz-Portillo et al. 2024). In the digital context, the importance of the freemium business model is well-known (Hamari et al. 2017; Holm and Günzel-Jensen 2017; Guo et al. 2022), and digital businesses particularly thrive when conversion occurs from free to premium (Mäntymäki et al. 2020).

While several businesses providing digital services have prospered based on the freemium business model – encouraging new as well as existing digital services providers to use this business model (The Economist 2022) –, others have failed (Holm and Günzel-Jensen 2017), increasing the appeal of research focused on the conditions for success using the freemium business model.

Notwithstanding the importance of digital businesses and digital entrepreneurship and the factors that support its success (Nambisan 2017; Berman et al. 2023; Galindo-Martín et al. 2023), existing research is still scarce (Kraus et al. 2019a, b, 2023). Likewise, previous research highlights the scarcity of studies focused on the freemium business model (Rietveld 2018; Roslan et al. 2022). A third and related research gap is understanding what motivations drive consumers to distinguish between competing platforms. Knowing the motivations that drive consumers to use digital services can be a source of strategic advantage, while knowing the motivations for upgrading from free to premium is vital to help entrepreneurs and companies strategize on their service’s versions (Camilleri and Falzon 2021).

In this research, we focus on the music industry, a paradigmatic example of the importance of digital businesses and the use of the freemium business model. Stimulated by recent developments (Calderon-Monge and Ribeiro-Soriano 2023), the size of the global digital music market is growing at an average annual growth rate of 8.8% and is expected to reach a global revenue of US$34.8 billion by 2025 (Statista 2021). Several on-demand music streaming services use the freemium business model (Statista 2020; Roslan et al. 2022), turning their service into a platform where free and premium users coexist. These companies are keen to know what consumers might be looking for in a music streaming platform and develop features to make it stand out (Morris and Powers 2015; Statista 2021) and stimulate conversion.

Therefore, this study aims to explore consumers’ motivations for selecting and adopting different platforms within the digital music industry, and then converting, or not, into premium users based on the insights from Uses and Gratifications Theory (Katz et al. 1974). Based on an online questionnaire focusing on users’ motivations and characteristics, we collected data from 231 users of music streaming platforms. Using nonparametric tests and the binary logistic regression, our results highlight the role of satisfaction, perceived value, and ubiquity in choosing a platform and, together with users’ age and occupation, in subscribing to the premium version. These results may be valuable for managers and entrepreneurs defining the strategy for their digital businesses.

2 Literature review

In this section, we will address the freemium business model, which is followed by a review of the literature based upon motivations and research mostly under the Uses and Gratifications Theory (Buzeta et al. 2020). This builds the set of six motivations and the research model presented later.

2.1 Freemium business model

From the digital innovation (Felicetti et al. 2023; Trischler and Li-Ying 2023) and management innovation (Lin et al. 2023) perspectives, one common and prominent research topic is business model innovation, considering the relevance towards businesses’ success (Guo et al. 2022). The freemium business model assumes a relevant role in this context, considering the presence in numerous digital content services (Hamari et al. 2017; Rietveld 2018; Niemand et al. 2019; Guo et al. 2022), and should be approached from the customer perspective (Rietveld 2018; Guo et al. 2022). Considering that a business model describes how businesses create value to capture value from customers (Rietveld 2018; Bouwman et al. 2019), the aim of the freemium business model is to attract consumers with low acquisition costs, but with high lifetime value (Zhang et al. 2010; Segal 2021). Since free users can use the service without financial commitment, businesses increase the total number of users (Segal 2021).

To ensure the success of the freemium business model, services must first attract and retain free users, and then convert them into premium users (Kumar 2014; Roslan et al. 2022). Having a solid user base is a key priority. But free users evidently do not provide the same level of revenue as premium users do (Guo et al. 2022), which makes conversion rates a prime concern (Niemand et al. 2019). To attract more consumers to the premium service, some companies offer a free trial of the premium version (Statista 2021), more effective on medium-to-high usage consumers (Reza et al. 2021). Additionally, services like Spotify, for example, offer premium alternatives with different costs (Statista 2021).

2.2 User motivations and characteristics

On a broad approach to social media use, Buzeta et al. (2020), based on the Uses and Gratifications Theory, focus on entertainment, integration and social interaction, personal identity, information, remuneration, and empowerment motivations. Camilleri and Falzon (2021), in the specific case of the video streaming context, focus on ritualized and instrumental use. However, focusing on paying for online content, the generic motivations are convenience, essentiality, added value, perceived service quality, usage frequency, perceived fairness and safety concerns (Wang et al. 2005). In a video game context, characteristics such as assurance, empathy, reliability and responsiveness are related to play intention, but none directly affect the intention to buy the premium option (Hamari et al. 2017). Additional constructs, such as intrusiveness of advertising, social connectivity, discovery, ubiquity, price value of the premium subscription, enjoyment and intention to upgrade or keep the premium, were also studied (Mäntymäki et al. 2020). According to Mäntymäki et al. (2020), the constructs enjoyment and price value are determinants of the intention to upgrade, while keeping the premium version was influenced by the constructs discovery and ubiquity.

In the context of digital music services, motivations such as enjoyment, discovery, ubiquity, and social connectivity lead to platform adoption and usage, with different levels of importance between free users and premium users (Mäntymäki and Islam 2015). The psychological characteristics are also relevant and play a key role in evaluating features and prices (Niemand et al. 2019). On the other hand, loyalty is also a factor when using these platforms because from a utilitarian point of view, a consumer invests time in creating playlists and recognizing the user-friendliness of the service, causing reluctance to switch providers (Sinclair and Tinson 2017). Empirically, satisfaction and user loyalty were found to be linked (Voss et al. 2010). Lastly, since advertisements are present for free users, ad intrusiveness may come into play, whereas some users may disregard them, as suggested by Mäntymäki et al. (2020). If perceived as intrusive, advertisements are likely to elicit negative attitudes in consumers (Li et al. 2002), although consumers that are highly adapted to advertisements aren’t motivated to convert (Z. Li and Cheng 2014). Curiously, there was no significant negative effect of intrusiveness of advertising impacting satisfaction among free users (Mäntymäki et al. 2020).

While not neglecting that some users may adopt the premium version directly, our research focuses on the path starting with the adoption of the free version and the subsequent conversion (or not) to the premium version. Based on the above literature, we formulate the following research hypotheses:

H1: The motivations influence platform adoption.

H2: The motivations influence premium conversion.

H3: The user characteristics influence premium conversion.

Based on Krause et al. (2014), to understand perceived gratifications in online applications of the music industry, this research will focus on Discovery (the ability to discover new music), Satisfaction (how much one enjoys using the service), Ubiquity (the ability to listen to music wherever and whenever one wants), Social and Personalization (the importance of personalization of the avatar, playlists and the opinion of friends), Exclusivity (what premium perks are more desired) and Perceived Value (how both versions of the service are perceived).

2.2.1 Discovery

Music streaming platforms can serve as a channel for music discovery (Aguiar 2017) through the platforms’ search and recommendation functionalities (Prey 2017; Mäntymäki et al. 2020). This leads consumers to follow recommendations because they often wish to discover something new (Kamehkhosh et al. 2020), which is something that individuals may value. In the context of music platforms, the recommendations are coherently associated with personal preferences revealed, for instance, through previously listened to and liked songs.

Generating the playlists customized for each user to discover new music poses a big challenge (Schedl et al. 2018). Playlist coherence and prediction accuracy not only lead to more approval but also increase quality perception by its users. Discovering new music, that is appreciated, is related to satisfaction (Garcia-Gathright et al. 2018), which is also related to artist diversity or track homogeneity in a playlist (Cunningham et al. 2006; Lee et al. 2011). Discovery was deemed a key factor contributing to premium users keeping their subscription (Mäntymäki et al. 2020).

Considering the above, we formulate the following hypotheses:

H1a: Discovery influences platform adoption.

H2a: Discovery influences premium conversion.

2.2.2 Satisfaction

Although satisfaction is key to choosing music to listen to, when addressing the satisfaction with the music platform, this is most valued among free users (Mäntymäki and Islam 2015) and strongly influences continuance intention (Kim et al. 2018). However, in this scenario, it has the inverse effect on purchase intention of the premium version (Hamari et al. 2020; Rahmansyah and Hati 2020). To offset this, if there are significant differences between the free version and the premium version, satisfied customers are converted (Kim et al. 2018). For free users, enjoyment plays a dominant role in predicting intention to upgrade, which may be understood as contrary to the perspective presented by Hamari et al. (2020) and Rahmansyah and Hati (2020), but poses no effect on the premium users in keeping the subscription (Mäntymäki et al. 2020).

Based on the literature, we formulated the following research hypotheses:

H1b: Satisfaction influences platform adoption.

H2b: Satisfaction influences premium conversion.

2.2.3 Ubiquity

One of the challenges for music enthusiasts that was difficult to address in the past relates to the precise choice of tracks to listen to at any specific moment. Nowadays, thanks to technological developments, listeners have more control over the music they listen to (North et al. 2004; O’Hara and Brown 2006) and are allowed to select freely, at any given moment, what they listen to, thus strengthening their sense of self-efficacy (Nambisan and Baron 2009; Mäntymäki and Islam 2015).

Ubiquity highly affects user-friendliness, time convenience and enjoyment in mobile services (Tojib and Tsarenko 2012). Furthermore, technologies that give power of choice to the user probably lead to a bigger sense of psychological ownership (Kirk et al. 2015), resulting in a more positive listening experience (Krause and North 2017a, b). For premium users, ubiquity is seen as a paid benefit, and important for their retention (Mäntymäki and Islam 2015). Additionally, freemium services can differentiate their free and premium versions in terms of ubiquity (Mäntymäki et al. 2020). Ubiquity may create a lock-in effect among premium users, as there were differences in both levels of ubiquity between basic and premium and the effect on converting and keeping the premium subscription (Mäntymäki et al. 2020).

Based upon the above, the following hypotheses are presented:

H1c: Ubiquity influences platform adoption.

H2c: Ubiquity influences premium conversion.

2.2.4 Social and personalization

Music streaming platforms allow users to create and share content with each other, even extending to other social media (like Instagram). This personalization enables users to tailor their own service experiences (Hamari et al. 2017; Sinclair and Tinson 2017), which may ultimately allow users to show and express an extended or enhanced self in the digital world (Belk 2013). Personalization takes the form of increased service options, interface choices, user avatars (Morris and Powers 2015; Hamari et al. 2017), user playlists and even personal uploads.

Most of the perks of becoming a premium user aren’t directly linked with community aspects (Oestreicher-Singer and Zalmanson 2013). Social connectivity is deemed the weakest factor to continue to use a service (Mäntymäki and Islam 2015) and has no effect on conversion intent, but showed a small negative effect on remaining a premium user (Mäntymäki et al. 2020). However, friends who purchased the premium strongly influence the propensity to buy the premium version while having non-subscribing friends has a small negative effect on converting (Oestreicher-Singer and Zalmanson 2013). This is where individual judgment is influenced by others, in the sense that these judgements are taken as more or less trustworthy in the reality that all are participating (Deutsch and Gerard 1955). Research found that on Last.fm, the presence of an affective community may be ultimately related towards monetary payment, along with the fact that more active users in the community will convert to premium sooner than less active or even non-active users (Oestreicher-Singer and Zalmanson 2013).

Considering the contributions from previous literature, we propose the following research hypotheses:

H1d: Social and Personalization factors influence the platform adoption.

H2d: Social and Personalization factors influence premium conversion.

2.2.5 Exclusivity

To distinguish both sides of the freemium spectrum, premium users are offered extra benefits. Thus, buyers have a relative advantage, defined as the benefit of premium features (Kim et al. 2018). However, balance must be taken into account for designing and sustaining high value offerings for both free and premium (Dörr et al. 2013; Palazon and Delgado-Ballester 2013; Holm and Günzel-Jensen 2017; Niemand et al. 2019).

The premium users, besides having avoided advertisements, are (generally) able to download and store tracks on the device – allowing for offline listening on the app – and stream at a higher sound quality, have access to the full music catalogue, the ability to choose any song on any device – or streaming mode (Kim et al. 2017) – and unlimited song skips (Soundcloud 2021; Spotify 2021).

Considering the insights on exclusivity, we formulated the following hypotheses:

H1e: Exclusivity influences platform adoption.

H2e: Exclusivity influences premium conversion.

2.2.6 Perceived value

Finally, in terms of motivations to adopt and convert, based on Krause et al. (2014), we focus on perceived value. One’s perceived value is a general assessment of the utility of a product, based on a trade-off between benefits and sacrifices (Zeithaml 1988). It is relative by virtue of its comparative, personal, situational nature, defining it as preferential, perceptual and cognitive-affective (Sánchez-Fernández and Iniesta-Bonillo 2007). Value for money corresponds to whether acquiring the premium version offers value for the money spent (Kim et al. 2018). If the perceived benefits surpass perceived sacrifices, conversion happens (Z. Li and Cheng 2014).

Usually, consumers will experience a zero-price effect (characterized by having a positive affective evaluation of the free option). Therefore, more value is perceived in the free version (Niemand et al. 2019). Companies could offset similarity between versions by changing availability of functions and features in both versions (Lin et al. 2013; Gu et al. 2018) and inform users on the gained benefits for adhering to premium (Shi et al. 2015). Providing additional features for free users creates perceptions among users of the value of the premium version (Wagner et al. 2014), increasing the probability of free users upgrading. However, price value has no influence on maintaining the premium version (Mäntymäki et al. 2020). This finding suggests that when presented with sufficiently good price value, decisions upon remaining premium are based on other constructs, such as discovery of new music or ubiquity.

Based on the previous literature, we propose the following research hypotheses:

H1f: Perceived Value influences platform adoption.

H2f: Perceived Value influences premium conversion.

2.3 Conceptual model and hypotheses

Based on the literature review, the first part of the research addresses the importance of each motivation for service adoption, while the second part evaluates if a premium conversion occurs, depending on how users evaluated their experience for each construct alongside user characteristics. In short, this study seeks to comprehend what motivates users regarding the adoption and use of different music streaming platforms, according to the conceptual model (Fig. 1).

Fig. 1
figure 1

Conceptual Model

Each construct is compiled by a set of items, allowing the measurement of how users evaluated their experience. Additionally, user characteristics encompass demographic attributes of the user, what content is preferred on music streaming platforms, how bothersome are ads, the importance given to music and listening habits. The research hypotheses are systematized in Table 1.

Table 1 Research Hypotheses

3 Methodology, data collection and data analysis

3.1 Methodology and data collection

In order to obtain data, the survey strategy uses a questionnaire. The study’s population are individuals who live in Portugal and who are at least 18 years old and are a user – free or premium – of one or more music streaming platforms. To acquire the target sample, the non-probability sampling technique of convenience sampling was used (Saunders et al. 2016). A pre-test was made to check if the participants would understand the study’s context, given questions and scales presented. The questionnaire was administered online from the month of June till July 2021 and was shared on different social media, including Instagram, LinkedIn, Facebook and WhatsApp by the researchers, totaling around 800 individuals.

We acknowledge that the use of a non-probability sampling technique (Saunders et al. 2016; Sarstedt and Mooi 2019) and the focus on a single national context can be considered limitations of our data collection. Furthermore, we recognize that a qualitative approach could provide more in-depth insights on the motivations of users while selecting the music platform and whether to become a premium user. However, we believe that the replication of this research in different contexts and complementing it with qualitative insights may constitute valuable avenues for future research. Figure 2 summarizes the methodology and data collection, while considering and identifying the previous and subsequent activities.

Fig. 2
figure 2

Summary of the methodology and data collection

To measure each motivation, various scales were adapted (translations to Portuguese and generalization of the topic in the questions) and items created based on the literature. Discovery was adapted from Mäntymäki et al. (2020), with the aim to understand how the recommended songs are perceived and its effectiveness. Satisfaction, also based on Mäntymäki et al. (2020), analyzes user satisfaction, and two factors were incorporated: loyalty towards the music streaming platform, and advertisement intrusiveness. Ubiquity endeavours to uncover the importance of the qualities of music streaming, which was studied by Krause and Brown (2019). Social and Personalization focuses on both social identity in streaming and the impact of a community, as studied by Mäntymäki et al. (2020). Perceived Value seeks to evaluate the perception of both versions of the freemium spectrum, previously studied by Wagner et al. (2014). The exclusivity factor checks if there were any preferred premium benefits among the identified list (Kim et al. 2017).

Besides these scales and items, additional items were included to assess the music streaming platform’s version and look (Morris and Powers 2015), the importance of music, the type of content listened to, and others. The degree of agreement or disagreement was measured via a seven-point Likert scale (Albaum 1997), ranging from Totally Disagree (1) to Totally Agree (7).

3.2 Data analysis

The data was analyzed via the SPSS IBM Statistics software. A total of 291 answers were harvested, being refined into 231 valid answers. For the hypotheses testing, six constructs were created from a set of questions via an index (Saunders et al. 2016), corresponding to each motivation. This creation demanded a preliminary data analysis of the new indexes, or construct validity (Saunders et al. 2016), through a Principal Components Analysis (PCA) (Table 2).

Table 2 Principal Components Analysis

The KMO values of the index variables oscillate between 0.640 up to 0.828 (satisfaction not included) and the total explained variance is at least superior to 46% for the presented indexes, close to the suggested 50% (Sarstedt and Mooi 2019). Regarding Bartlett’s Test of Sphericity, all variables are significantly correlated (p = 0.000). The values for communalities reach at least 0.5 (Sarstedt and Mooi 2019), with very few exceptions. Additionally, to measure construct internal consistency and assess reliability, the Cronbach’s Alpha coefficient was calculated (Table 3).

Table 3 Cronbach’s Alpha coefficient

All presented index variables with a Cronbach’s Alpha greater than 0.7, except satisfaction, are granted internal consistency and reliability for the created variables of this study (Sarstedt and Mooi 2019). According to the preliminary data analysis, the satisfaction construct failed to show correlation and consistency. Thereupon, each item composing this motivation will be analyzed separately and only one item measured satisfaction itself. Additionally, items on other scales that turned a construct’s Alpha Cronbach value to lower than 0.7 were also studied in isolation.

The nonparametric test Kruskal-Wallis H (Sarstedt and Mooi 2019) was applied, considering normality issues and the low number of users of other platforms, to measure the relation between the importance of motivations and the adoption of different music streaming platforms (H1). The Binary Logistic Regression (Sarstedt and Mooi 2019) was used to study the influence of each motivation and user characteristics on premium conversion (H2 and H3, respectively).

4 Results

4.1 Descriptive statistics

Of the 231 participants who use at least one music streaming platform, 66.7% are female, 86.6% are aged between 18 and 25 years, 56.3% have a bachelor level, and 62.8% are students. 57.6% don’t have income (Table 4). Regarding the importance of music, this is high (mean = 6.07), and the listening frequency is also high (mean = 6.21). Daily use stands at between 3 and 4 h (mean = 3.62). The most common devices on which people listen to music, in descending order of preference, are the mobile phone, computer, radio, tablet, CD/Vinyl record player.

Table 4 Sample characteristics: gender, age and education (%)

From the various platforms, YouTube is the most used service (93.1%), followed right after by Spotify, while the remaining services share a relatively small number of users. The preferred music streaming platform is Spotify. The users consider that platforms possess a good look (mean = 6.23). The type of content that was most sought after was music (100%) followed by podcasts.

Focusing on the type of user, about half of the sample are free users (51.1%). 7.4% stated they were premium in the past, but not anymore. Reasons as to why the subscription was cancelled were “I no longer want to pay”, “The cost was too high”, “I use Spotify on the computer, …, since I can freely choose what song I want to listen to”, “the premium trial was over, and I did not feel the need to upgrade”, “Not worth it” and the increase in price for “I’m no longer a student” and a curious answer stated that “With so much supply, might as well change to the one that is the cheapest”. The idea of no longer being able to use the student discount, the end of the trial and not seeing value in premium were the most stated answers.

When asked to define the importance for each motivation while using music streaming services, the most important motivation is Ubiquity (Mean = 6.48), followed by Satisfaction (Mean = 6.19), Discovery (Mean = 5.86), Social and Personalization (Mean = 5.22), Perceived Value of the free and premium (Mean = 4.91) and lastly, Exclusivity (Mean = 4.23).

The Inter-Item Correlation Matrix for Discovery was created (Table 5) to address previous literature, regarding the connection between high prediction accuracy or high track similarity into high user satisfaction or an increased adoption of the service.

Table 5 Inter-Item Correlation Matrix for Discovery

According to the Matrix for Discovery’s items, liking the new recommended tracks, tracks fitting one’s musical taste and the recommendation system making good choices achieved a positive correlation. In the Correlation Matrix of the model, Satisfaction is positively correlated to Discovery, but had a negative relation with Ubiquity. Usage frequency and number of hours listening to music were positively related to the Social index.

When asked if users were satisfied, the answer is positive (mean = 6.27), with premium users agreeing more than free users. Users somewhat agree that the services they use exceed their expectations (mean = 4.76). Both types of users also agree on advertisement disturbing the music listening experience (mean = 6.24), with free users being slightly more vocal on the matter. On being loyal to the music platform, users somewhat agree (mean = 5.31) and the more hours listeners stream, the more evident this loyalty is. Finally, 61.5% of premium users stated they used other music streaming platforms less frequently, 16.7% remained using different services and the remaining 21.9% stopped using the alternatives.

4.2 Hypotheses testing

The first Research Hypothesis, H1, hypothesizes a connection between motivations and the adoption of different music streaming platforms. The analysis was supported on the nonparametric test Kruskal-Wallis H (Sarstedt and Mooi 2019). Table 6 shows results on the motivations.

Table 6 Kruskal-Wallis H – Motivations in the Adoption of a Platform

The results (Table 6) show that the differences are statistically significant when it comes to the influence of Satisfaction (K-W (2) = 11.708; p = 0.003), Ubiquity (K-W (2) = 9.503; p = 0.008) and Perceived Value (K-W (2) = 6.083; p = 0.048) motivations on choosing different music streaming platforms. Due to a low user distribution in the studied sample, Soundcloud, Tidal, Apple Music and YouTube Music were integrated into a single group called “Other” platforms, allowing for a simpler interpretation of the results. Satisfaction manifests itself with the highest value for Spotify, being followed by YouTube and then Other (almost sharing the same mean rank). Ubiquity scores the highest value with Other, lowering progressively for Spotify and lastly YouTube. Perceived Value tops with Spotify, Other (being closely tied) and then YouTube. This way, H1 is not rejected for Satisfaction, Ubiquity and Perceived Value, while not supported for the remaining motivations.

The second and third Research Hypotheses, H2 and H3, hypothesize a connection between motivations and user characteristics in influencing premium conversion. The analysis is supported by a Binary Logistic Regression (Sarstedt and Mooi 2019). The Cox and Snell R Square and Nagelkerke R Square values indicate 67.3% and 90.6% of the variability is explained by the total set of variables of the model and the Hosmer and Lemeshow Test suggested the model is fit (Pallant 2016). It classifies with a 95.2% accuracy the cases of subscribing to the premium service. Table 7 presents the significant variables in the Binary Logistic Regression model.

Table 7 Significant Variables in the Model

The results reveal that only few motivations and user characteristics influence premium conversion. The motivations are Perceived Value (B = 3.212: p = 0.000), Ubiquity (B = 5.720; p = 0.000), Satisfaction (B=-1.182; p = 0.047) and user characteristics Age (B=-1.841; p = 0.015) and Occupation (B = 0.872; p = 0.019). Therefore, H2 is supported for Perceived Value, Ubiquity and Satisfaction, while not supported for the remaining motivations. Additionally, H3 is supported by Age and Occupation, and rejected for the other user characteristics.

5 Discussion and conclusion

5.1 Discussion

Based on our results we argue that music streaming users are influenced by satisfaction, ubiquity and perceived value, defining these as significant motivations to distinguish music streaming services. When adopting a new platform and before considering going premium, usually individuals start as a free user. In this context, satisfaction can be interpreted in the way that every consumer expects to be pleased when trying out a new service. Therefore, it’s possible to theorize a connection with the previous findings by Mäntymäki and Islam (2015), Kim et al. (2018), and Mäntymäki et al. (2020) regarding satisfaction being important for free users. Furthermore, it may influence choosing a new platform, taking into consideration the strong influence satisfaction has on continuance intention (Kim et al. 2018).

Companies strategize in a high value free version and premium version with distinct characteristics (Lin et al. 2013; Wagner et al. 2014; Gu et al. 2018; Kim et al. 2018), suggesting perceived value as another motivation in choosing a platform that has benefits new users notice, which was also supported in the context of this research. This aligns with Hamari et al. (2020) because the service must be attractive enough for newcomers, but not perfect, so that consumers want to upgrade, which is freemium’s ultimate goal (Kumar 2014; Niemand et al. 2019).

Moreover, ubiquity, which was also found to be statistically significant, comes in as a stimulus that music streaming services have when compared to other formats to listen to music, reinforcing the findings about usability, functional utility, flexibility and playback diversity by Krause and Brown (2019). Companies can manipulate ubiquity of their service in different ways. For example, a free user on YouTube can’t continue to stream the music video while the cell phone is blocked, and Spotify instead disables the ability to choose any song for a free user on mobile, somewhat limiting the service’s ubiquitous quality.

Some studies were dedicated to what led platform users to convert (Mäntymäki and Islam 2015; Mäntymäki et al. 2020). In this regard, results of this research suggest that the motivations perceived value, ubiquity, satisfaction and user characteristics age and occupation influence premium conversion, thus expanding the findings of Mäntymäki et al. (2020) on enjoyment and price value.

Regarding satisfaction, the results replicate previous literature (Hamari et al. 2020; Rahmansyah and Hati 2020): satisfaction does impact a premium conversion, but negatively (B=-1,182). In other words, the greater the satisfaction of a free user, the less likely premium conversion should occur. If the free user is satisfied with the service, there are no reasons to upgrade.

The fact that perceived value presents itself as a positive influence in conversion suggests that consumers view the premium service as a high-value offer, successfully differentiating the service in accordance with the business model’s strategy (Gu et al. 2018; Lin et al. 2013; Wagner et al. 2014). Creating distinct and attractive premium features, as recommended by Kim et al. (2018), was proven to originate an offer with value for money because 41.6% of the questionnaire’s sample are premium users. The main goal of freemium is premium conversion (Kumar 2014; Garrahan et al. 2015; Mäntymäki et al. 2020) and in this study’s findings, it is proposed that Spotify’s users perceive premium highly, in contrast to YouTube whose users are mainly free, thus revealing clear differences in the rate of conversion to premium between the different platforms.

The discovery motivation is statistically non-significant regarding conversion, indirectly posing a contradiction with the findings of Mäntymäki et al. (2020) on keeping the subscription, especially for premium users. The same conditions apply to Social and Personalization motivations. This could be easily justified by the fact that there are no changes or upgrades on how music is recommended, personalization and social options, regardless of the type of user (Oestreicher-Singer and Zalmanson 2013), but may open room for new approaches of the platforms in this area.

Finally, our results reveal that ubiquity positively influences conversion. Since exclusivity (a list of premium perks) was assigned as statistically non-significant, ubiquity seems to be the main reason for upgrading the service, confirming previous results and retaining premium users (Mäntymäki and Islam 2015; Mäntymäki et al. 2020). This motivation is key in defining music streaming services (Hagen 2016) and differentiation strategies deployed by businesses regarding the service’s version (Mäntymäki et al. 2020). In practical terms, premium users value their ability to choose any song, at any point in time. According to the study, Spotify premium users most likely value listening to their preferred song in that moment or via offline access, while free are unable to on mobile devices or offline stream said track.

As for user characteristics, age negatively affects the premium purchase, which translates to the older the user, the more likely it’s a free user. This may constitute a specific challenge for platforms aiming to enlarge the penetration among older users, but this result might have shown up due to the high percentage (86.6%) of young individuals in the sample. On the other hand, occupation has a positive effect, which may be explained by the same comment on age and a possibility that justifies this result is the existence of the Spotify Premium Student alternative aligning with the student respondents (62.8%). These findings find support from what was suggested in previous literature, regarding both types of users having distinct demographics (Anderson 2009; Pujol 2010).

Besides music streaming platforms adoption and premium purchase motivations, this study also endeavors to address other events that take place when using these services. The answers in the unique question for the users who stopped being premium greatly reflect on the importance of offering a limited time trial of the premium service and the effectiveness of different pricing models like the student discount for Spotify (Statista 2021), further emphasizing premium sales promotions (Palazon and Delgado-Ballester 2013). In the same question, participants commonly reported that the service was not worth acquiring after the trial expired, which may be interpreted as a lack of perceived value motivation by these users. This provides answers to Mäntymäki and Islam (2015) on why premium is discontinued.

A challenge faced by these services is the generation of personal playlists (Schedl et al. 2018), because it is uncertain to what extent prediction accuracy leads to higher user satisfaction (Lee et al. 2011). The findings suggest that music platforms indeed succeed with their algorithms, due to the positive mean values in the items regarding new music discovered by the recommendations and a positive Inter-Item Correlation between these items. In practical terms, this suggests that good recommendations lead to increased satisfaction with the service, confirming previous literature (Cunningham et al. 2006; Lee et al. 2011; Garcia-Gathright et al. 2018). Ubiquity was found to be negatively correlated to satisfaction: despite having limited streaming options, compared to their premium counterparts, free users remain satisfied. The conclusions from Krause and North (2017a, b) regarding positive listening experience would seem contradictory at first, but since satisfaction plays a negative role in conversion and ubiquity is important, an unsatisfied free user would be motivated (by ubiquity) to upgrade and be more satisfied as a result. The social motivation found accordance with Oestreicher-Singer and Zalmanson (2013) regarding the effect of judgment on the free/premium version and the weakest positive correlation was the sense of belonging in a community. Personalization features recorded positive levels of importance by the respondents, especially highlighting the ability to personalize playlists and checking others’ music activity. However, it is important to remember that neither social nor personalization motivations are critical in the adoption or premium conversion, suggesting that these features are welcome, but actually not that important for consumers.

As Roslan et al. (2022) argue, modelling the behaviour of users is challenging, which is mitigated in this research, as we reach 95% accuracy classifying the premium users with the proposed model. The competing services were distinguished by ubiquity, perceived value and satisfaction motivations. Satisfaction needs to be balanced, because free users won’t convert into premium if their needs are met. This research also brings some insights on why premium users return to the free version, which interconnects with previous research (Mäntymäki and Islam 2015).

Both previous literature (Lin et al. 2013; Wagner et al. 2014; Gu et al. 2018) and the results reinforce that creating a high-value free and premium version of the service poses relevancy. Freemium’s unique nature allows businesses to strategize and balance the number of features available for both versions (Wagner et al. 2014; Mäntymäki and Islam 2015; Hamari et al. 2017) and ubiquity is one of the drivers that motivates consumers. Likewise, promotion and trials play a decisive role in attracting consumers to go premium (Palazon and Delgado-Ballester 2013; Statista 2021). The strategies involved in how users control their music (ubiquity) can also be worked on, as these are vital for premium users (Mäntymäki et al. 2020). A different point of view is suggesting companies develop new premium features for music discovery, social and personalization aspects, as these lack differentiation for free or premium. Ultimately, this poses as an opportunity to further increase the reasons to upgrade. Lastly, satisfaction must be balanced carefully, offering a music experience that any user can enjoy and still generate the desire to convert to premium.

5.2 Conclusion

The recent evolution of technology enabled the development of new digital businesses, new business models, and new digital entrepreneurial opportunities, including in the music industry which we explore in this research as an example of freemium business model use in the context of digital businesses. The technological advances lead to new alternatives to access several services, for example listening to music, allowing individuals to satisfy their needs in an optimized way, without needing specific devices as in the past, and facilitating their access to the desired contents, mostly without incurring a financial cost.

The literature review revealed that, in a broad context where knowledge of digital businesses and digital entrepreneurship success determinants is still scarce, it is especially relevant to clarify the dynamics associated with the freemium business model, towards the premium version adoption, and the motivations leading consumers to distinguish between freemium services. Based on these research gaps, in this paper we simultaneously and distinctively approach the importance of consumers’ motivations to adopt freemium services and conversion to premium versions.

Our analysis highlights ubiquity, perceived value and satisfaction as driving motivations, alongside the user characteristics age and occupation, as significant to choosing a platform and subscribing to the premium service. These findings suggest that the companies who own these services should maintain a high-value, balanced offer, for both free and premium.

The freemium business model, despite its merits, brings several management challenges. First of all, the characteristics of the platform need to surpass the competitors’ platforms. But, if the free version is too good, it may hinder the users’ motivation to convert to the premium version. This is supported by these research results, showing the importance of ubiquity and perceived value both for the adoption of platforms and the conversion to the premium version; but, as for satisfaction, platforms need to realize that while this positively influences the adoption, it negatively influences the conversion to premium, which should lead platforms to carefully assess their decisions leading to high adoption but poor conversion, impacting their survival chances. The next step is to maintain the premium users, which can also be challenging, as the analysis of the answers of the quitting users also reveals.

From an academic perspective, this paper further increases knowledge of digital businesses, especially focusing on the importance of motivations in the music streaming context, as an example of the freemium business model. The managerial contributions provide insights to companies about what drives consumers in the differentiation of these services and on the motivations that lead to premium subscription.

5.3 Limitations

While carrying out this research, a few limitations were noted. The use of a non-probability sampling technique is prone to bias and influence beyond the researcher’s control, which may result in not yielding a representative sample. As described, the Satisfaction is solemnly measured by 1 item, possibly conditioning the results, which is particularly relevant given the positive influence of satisfaction for the adoption of the platform and, contrarily, to the conversion to premium user. Finally, our analysis is focused on music streaming platforms, which may be insufficient to draw insights for other types of businesses.

5.4 Future research lines

Regarding future research, a construct absent in this research is price, and this could be incorporated in the future, especially considering that the perceived value was present and statistically significant in our research. Since age was found to negatively lead to premium conversion, it would be interesting to endeavor to examine if younger consumers were more keen to stream, while older consumers preferred more traditional ways of listening to music.

Future studies could research the studied motivations in another freemium context, such as free-to-play video games, or different music formats. Finally, while we present some insights on the motivations to abandon the premium versions and return to the free versions, a more substantive analysis of these motivations to abandon would also be relevant for practitioners.