Factors Driving Digital Music Purchases: A Study Based on the Theory of Reasoned Action
Factors Driving Digital Music Purchases
Chechen Liao and Yi-Jen Huang — National Chung Cheng University
Pui-Lai To — National Chiayi University
Yu-Ting Lu — National Chung Cheng University
The digital music ecosystem continues to evolve, generating a wide array of products. This study examined the digital characteristics of music as a nonphysical good. Drawing on data from a survey, the researchers built a model based on a refined version of the theory of reasoned action, integrating three groups of factors: risk-related and accessibility elements tied to intangibility; price, variety, and trialability as product-type attributes; and entertainment components such as anticipating consumer needs and perceived playfulness. Results showed that product variety most strongly shaped consumers’ attitudes and intentions, while actual purchase decisions were jointly influenced by attitude and subjective norms. Since trialability had no significant effect on either attitude or intention, lowering the price of digital music may offer a more effective path to reducing perceived risk.
Music is well suited for digital sale because the major costs for record companies are recording and packaging, while selling online eliminates retail overhead and allows individual track purchases.
Rights management in music sales addresses piracy and benefits both buyers and sellers. Some difficulties persist, though earlier research has offered prevention strategies.
The success of Apple’s iTunes store demonstrated that digital music buyers can avoid the common frustration of purchasing an entire album just to get a few tracks. Legal issues arising from early file-sharing declined noticeably after iTunes launched with support from five major labels, signaling both industry demand and consumer willingness to pay. According to the Pew Internet and American Life Project, peer-to-peer usage dropped from 31% in 2004 to 21% in 2005, while paid downloads from iTunes and similar stores rose from 24% to 43% over the same period.
From the customer’s viewpoint, value-added services shape satisfaction and perceived quality. Online shoppers can select individual songs, achieving greater satisfaction at lower cost. Digital stores also offer extra artist information, broader genre selection, current hits, easy consumer feedback, and preview clips. Record companies, meanwhile, can use web stores for advertising and promotion.
E-commerce studies have frequently examined motives behind online shopping — focusing on physical goods, cost, product selection, and ease of use. But research on nonphysical digital purchases remains limited. The present work aimed to close that gap by building on existing knowledge, exploring consumer intention, and identifying key factors in digital markets. Three objectives guided the study:
- Identify factors influencing consumers’ preference for digital music
- Identify which consumer groups are most receptive to buying digital music
- Extend analysis of digital product attributes to help sellers design strategies for boosting digital sales
Research Background and Hypotheses
Literature Review
According to Ajzen and Fishbein’s theory of reasoned action (TRA), a direct link exists between individuals’ attitudes, environments, behaviors, and intentions. The theory holds that a person’s positive or negative stance toward adopting a particular behavior is embedded in that behavior itself.
Moon and Kim described perceived playfulness as an intrinsic source of motivation driving intention, distinct from the extrinsic motivation represented by perceived usefulness. Both types influence how people use a system. Research has shown that playfulness favorably shapes both intention and attitude.
Perceived risk refers to a person’s subjective belief that efforts to achieve an outcome may produce negative side effects. In online transactions, two major forms arise: behavioral uncertainty and environmental uncertainty. Online retailers can exploit the evolving state of e-commerce and the lack of tight transaction regulation, engaging in opportunistic actions that increase behavioral uncertainty.
Digital products differ from physical goods in their cost structure: fixed production expenses are substantial, but copying additional units costs almost nothing. Producing music — a type of information — is expensive, but reproducing it is cheap. Growth rates vary across digital product types, and trading terms may differ — for example, pay-per-use versus pay-per-period. Key attractive attributes include trialability, price, and granularity. Product diversity also influences customer preferences and attitudes. Besides convenience, digital goods offer a strong competitive edge by reducing shopping time.
Hypotheses
The research model rested on TRA and incorporated digital characteristics grouped into three levels: the first covering TRA and product attributes, the second concerning product variety and convenience, and the third dealing with risk issues and preceding factors.
Perceived playfulness has been shown to positively affect consumers’ intention and attitude. Studies on technology acceptance have confirmed that playfulness, as an independent variable, directly influences user attitude and intention. Since music can relieve stress and promote friendliness, the first two hypotheses were:
Hypothesis 1: Perceived playfulness will positively affect consumers’ attitudes toward purchasing digital music.
Hypothesis 2: Perceived playfulness will positively affect consumers’ digital music purchase intentions.
Attitudes and subjective norms from TRA and the theory of planned behavior are significant variables with causal links to intention. A positive relationship exists among purchase intention, subjective norms, and attitude. The study examined how consumers’ perceptions of family and social opinions influence their inclination toward digital music:
Hypothesis 3: Subjective norms will positively affect consumers’ digital music purchase intentions.
Hypothesis 4: Consumers’ attitudes toward making purchases will positively affect their digital music purchase intentions.
Product-Level Characteristics
A broader product selection strengthens a retailer’s competitiveness. The ability to maintain a dynamic, varied selection may be a major reason consumers choose one store over another. Online music retailers can offer over a million tracks — vastly more than any physical store — leading to these hypotheses:
Hypothesis 5: The variety of digital music available will positively affect consumer attitudes toward digital music purchases.
Hypothesis 6: The variety of digital music available will positively affect consumer intentions toward digital music purchases.
Earlier studies indicate that online shopping motives include cost, product categorization, and ease of access — making these factors foundational for a successful e-market. Price is arguably the main element here, and some argue it affects attitude. Consumers’ evaluation of service influences their behavioral intent. Online shoppers can find more cost-related information and compare different retailers, potentially leading them to believe they can get lower prices online. Price-oriented consumers compare prices across channels and choose the cheapest option. The following hypotheses focus on two price types: CD prices and digital music prices.
Hypothesis 7: Compact disc prices will positively affect consumers’ attitudes toward digital music purchases.
Hypothesis 8: Compact disc prices will positively affect consumers’ digital music purchase intentions.
Hypothesis 9: Digital music prices will negatively affect consumers’ attitudes toward digital music purchases.
Hypothesis 10: Digital music prices will negatively affect consumers’ purchase intentions toward digital music.
Some research suggests that consumers who enjoy seeing, feeling, or trying products before buying prefer physical shopping. Studies have found that a majority of respondents like this tactile experience, although digital items like trial software were exceptions. Trialability is seen as a significant means to encourage innovation diffusion and influence consumers’ purchase intentions:
Hypothesis 11: Trialability of digital music will have a positive effect on consumers’ intention to purchase digital music.
Surveys of focus groups across the US showed that convenience was the main reason for online shopping. Other researchers have also cited convenience as a critical factor. Online shopping eliminates noise and personal interaction. Consumers who find offline shopping difficult are more likely to shop online, where they can browse freely. Saving time is a major motivator — especially avoiding the inconvenience of downtown parking. Digital music is available instantly online:
Hypothesis 12: Convenience will positively affect consumers’ digital music purchase intentions.
If products can be sampled, the conversion rate among shoppers tends to be higher. The impact of online reviews on conversion is lower for goods with a sampling option than for those without. Consumers perceive greater risk when buying digital products since they cannot preview them before purchase. Many music retailers offer 30-second samples, yet some research indicates that perceived risk can inversely affect perceived value:
Hypothesis 13: Price will negatively affect consumers’ perceived risk of purchasing digital music.
Hypothesis 14: Trialability will negatively affect consumers’ perceived risk of purchasing digital music.
E-commerce transactions occur in settings of behavioral uncertainty and environmental uncertainty. Behavioral uncertainty arises because online retailers may behave opportunistically, exploiting the immaturity of e-commerce and limited government monitoring. Such uncertainty creates economic risk (monetary loss), personal risk (unsafe products), seller performance risk (imperfect monitoring), and privacy risk (disclosure of private information). Environmental uncertainty stems from the unpredictable nature of the internet — risk of credit card theft, data breaches, and stolen private information. When consumers shop online, perceived risk may lessen their purchase intention. Since online purchasing is an information-technology-based form of direct marketing, consumers see it as riskier than offline shopping, and risk-averse buyers may avoid it entirely. Internet security positively influences purchase intention:
Hypothesis 15: Perceived risk will negatively affect consumers’ digital music purchase intentions.
Method
Design of Measures
Convenience. Convenience encompasses effort, time, and space. This study measured convenience through the time saved in online purchasing using items adapted from earlier research.
Price. Since both CDs and digital music have a cost, price is a significant component in purchase decisions, measured here using a method from prior work.
Trialability. Music categorization was determined based on an established approach. Some digital music products offer previews or temporary use — features that fall within trialability measures.
Perceived Risk. Perceived risk is subjective; the possibility of loss arises from the intended result. This study gauged risk perception using items from prior research. After receiving free samples, prospective purchasers should perceive lower risk concerning hit songs.
Playfulness. Perceived playfulness was measured based on three factors: the person’s focus on the digital track, exploration through interaction, and the belief that the interaction is fun. Items came from earlier studies.
Subjective Norms, Attitudes, and Intention. Items related to attitude and subjective norms from TRA captured respondents’ attitudes toward digital music and the effect on others’ perceptions. Items measuring sentiment and impacts on subjective norms were drawn from established sources.
All items used a 7-point Likert scale ranging from 1 (totally disagree) to 7 (totally agree). Normality and multicollinearity were determined using SPSS before the hypotheses and model were tested with LISREL 8.51.
Pilot Study
Aggregate data were collected using multiple methods. Two academics first reviewed the survey instrument.
The instruments were evaluated by postgraduate students in the College of Management at National Chung Cheng University who had substantial online shopping experience. These individuals reviewed all items and provided recommendations to improve reliability and feasibility. A pilot study was then conducted with 46 participants to test the survey instruments among customers of KKBOX, the largest Taiwanese digital music retailer. In Taiwan, only Yahoo and KKBOX are authorized to sell digital music.
Approximately 70% of pilot study participants were between 21 and 25 years old, with 62% women and 38% men. All pilot participants had broadband Internet access. Cronbach’s α was used via SPSS to assess construct reliability in the pilot.
Main Study
Procedure
For the main study, we tested our model through an online survey using KKBOX as the primary source. Surveys were distributed to KKBOX purchasers, employing a snowball sampling methodology where invitations were forwarded through software. Digital music buyers completed the survey and were asked to invite other KKBOX users to participate.
The invitation described the research purpose and requested participation. Recipients agreed by clicking on Internet links. Each response was automatically stored to protect respondent data, and email addresses were collected for identification. Respondents could only submit the survey after answering all items. As incentives, participants could win one of five iCash cards (worth 200 E-Taiwan dollars [US$6.15]) or one of two MP3 players in a lucky draw. We received 195 completed surveys, five of which were rejected due to overly personal responses.
Participants
The final sample consisted of 190 participants (58% women, 42% men). Most participants (83%) were aged 20 to 30 years, 7% were under 15, and one was older than 35. Regarding education, 46% had university or college education, and 34% held a master’s degree. Occupational distribution closely correlated with age: for example, the proportion of participants under 25 matched the proportion who were students (65%). Aside from students, the remaining participants were employed. Most students (58%) had an annual income of 50,000 Taiwan dollars (US$1,536.65) or less. All but one participant, who used a 56k modem, accessed the Internet via broadband. Additionally, 70% of participants had used the Internet for over 5 years, with 35% using it for 5 to 7 years and 35% for more than 7 years. Daily Internet usage varied: 28% spent 2 to 3 hours, 26% spent over 6 hours, and the remainder reported 19% (3 to 4 hours), 10% (4 to 5 hours), 9% (5 to 6 hours), and 8% (less than 2 hours).
Results
Multicollinearity
Correlations among latent variables were all below .85, with the highest at .66. Variance inflation factor values were less than 10, confirming no multicollinearity among multiple variables.
Measurement Model
Hypotheses were tested using LISREL 8.51 through structural equation modeling (SEM) after verifying normality and collinearity, establishing exact correspondence between latent and manifest variables, assessing variable reliability, and evaluating causal relationships. Except for the goodness-of-fit index (GFI), all other indices—nonnormed fit index (NNFI), incremental fit index (IFI), root mean square error of approximation (RMSEA), comparative fit index (CFI), and standardized root mean square residual (SRMSR)—met required thresholds: χ²/df = 1.71, GFI = .92, NNFI = .92, IFI = .93, RMSEA = .06, CFI = .93, SRMR = .05.
Validity and Reliability
Following confirmation of overall model fit, we analyzed factors and the extent to which construct indicators defined a common, unidimensional latent construct. Values above .5 indicate high composite reliability through standardized loadings (.7). Average variance extracted exceeded .5 for constructs, though this alone does not guarantee validity. Validity, assessed by standard factor loadings greater than .6, was acceptable. Reliability and validity results from confirmatory factor analysis, along with discriminant validity tests, were satisfactory.
Structural Model
The model’s fit indices, except GFI (which was acceptable), met thresholds: χ²/df = 1.72, GFI = .80, RMSEA = .06, NNFI = .92, CFI and IFI = .93, SRMR = .063. Figure 2 illustrates the research model with standardized LISREL path coefficients. Eight hypotheses (H1, H3, H4, H5, H6, H9, H13, H15) were supported; H2, H7, H8, H10, H11, H12, and H14 were not significant.
DIGITAL MUSIC PURCHASES
Figure 2. LISREL path coefficients (standardized). Note: *p < .05, p < .01, *p < .001.
Perceived playfulness → Attitude (0.47*), Perceived playfulness → Subjective norms (0.09). Subjective norms → Intention (0.29). Attitude → Intention (0.47). Price of digital music → Attitude (−0.13), Price of digital music → Intention (0.09). Variety → Attitude (0.31), Variety → Intention (0.25***). Price of CD → Attitude (−0.03), Price of CD → Intention (0.20*). Digital music purchasing intention (outcome).
Discussion
This research identified key factors influencing customer intention to purchase digital music online. Our results, contrary to prior studies (Csikszentmihalyi, 1975; Moon & Kim, 2001), showed that perceived playfulness had an indirect effect (H1 and H2). This may reflect a decline in playfulness associated with digital music purchases compared to other digital goods like games, movies, or software. If Csikszentmihalyi’s (1975) flow theory holds, listening to music is a spontaneous activity while using music programs, so playfulness does not directly drive sales (Kim, Koh, & Lee, 2009). Nonetheless, playfulness remained the strongest influence on attitude in our study, suggesting digital music enhances consumer happiness (Guo & Barnes, 2009; Moon & Kim, 2001; Wang, Yeh, & Liao, 2013).
A similar pattern emerged for price, which affected intention indirectly through attitude (H9) but not directly (H10). This diverges from earlier studies, as consumers may compare prices of physical and digital music (Kim et al., 2009; Kim, Chan, & Kankanhalli, 2012; Wang et al., 2013). However, Taiwan’s oligopolistic online music market limits consumer choices. CD sales and digital music sales are separate products, making it difficult for consumers to compare prices across retailers or formats, reducing purchase inclination. Thus, digital music price did not directly impact intention, but it influenced attitude through objective price (actual cost) versus perceived price (psychological interpretation) (Jacoby & Olson, 1976). CD price showed minimal effect on intention.
Variety was the only factor significantly affecting both intention and sentiment. Prior research indicates that consumers prefer stores with broader product selections (Zhou & Duan, 2009). This applies to physical and digital stores alike, with digital online retailers like iTunes offering over five million categories—a competitive advantage physical stores cannot match.
Trialability proved unimportant in our study, an unexpected finding. We anticipated it would reduce perceived risk and increase purchase intention, but hypotheses H11 and H14 were unsupported. Reasons include the trial period of three days being too brief to provide useful information, despite available expert reviews, artist details, chat rooms, and search tools. Additionally, trial recordings had low sound quality, and many stores offered only 30-second samples, which were insufficient for most consumers.
Many studies highlight how intangibility of digital products shapes perceived risk, as our findings confirm. While sellers can reduce risk by lowering prices to improve consumer attitude (Dai, Forsythe, & Kwon, 2014). Prior research identified ease-of-access as significant for online shoppers (Gehrt et al., 1996), but our results (H12 not supported) suggest convenience did not affect digital music purchase intention. This may be because consumers do not view time, space, or effort as key motivators for intangible products. Convenience is more important for physical products requiring logistics.
Drawing on Kwong and Park (2008), who found subjective norms strongly predict subscription intention, we modeled subscription behavior. Our results supported TRA predictions of attitude and subjective norms positively influencing purchase intention (H3 and H4), confirming that both factors encourage digital music purchases.
Conclusion
Our findings highlight key elements of purchase intention, including that trialability does not apply to digital music. Digital music sellers must reduce perceived risk, for instance by lowering prices, to indirectly boost purchase intentions. Our results also confirm that perceived playfulness significantly differentiates online from physical shopping. Companies offering music downloads should develop software to enhance consumer engagement and interaction.
Variety emerged as a strong predictor of intention and sentiment. It, along with attitude, was a single-dimensional factor influencing subjective norms, intention, and perceived risk. Sellers must expand product offerings and show a wider array. Retailers should leverage subjective norms to increase purchase intentions by emphasizing music companies’ legal authority and discouraging piracy.
Reasons for this may be, firstly, that the 3-day trial was too short to give customers additional information, despite the availability of expert reviews, artist information, chat rooms, and search systems. Second, trial recordings had subpar sound quality, and also, several stores offered just 30 second samples of music, a time period too short for most consumers.
In many existing studies the digital product’s intangibility shapes the perceived risk, as our findings also demonstrate. Although sellers may take steps to reduce perceived risk, they can reduce the price of their products to lower risk and enhance consumer attitude (Dai, Forsythe, & Kwon, 2014).
Our research model was based on the TRA, incorporating elements related to digital music and e-commerce. For example, Hui and Chau (2002) classified music as one of three digital product categories. Future communication research could investigate varying consumer intentions for television, film, and music purchases to generalize our model. Researchers might also explore differences among services like myVideo, iVideo, Hulu, Vudu, YouTube, and Netflix to understand their appeal to different consumer groups.
Cloud computing and technology will continue to influence digital music and video sales by improving ease-of-access and enabling direct purchasing. Early consumer interactions with digital products suggest that global digital item markets will expand, motivating further research to keep pace with evolving trends.
List of References follows.
Adopting an information technology innovation.
Pavlou, P. A. (2003). Consumer acceptance of electronic commerce: Integrating trust and risk with the technology acceptance model. International Journal of Electronic Commerce, 101–134.
Peace, A. G., Galletta, D. F., & Thong, J. Y. L. (2003). Software piracy in the workplace: A model and empirical test. Journal of Management Information Systems, 153–177.
Rogers, E. M. (2010). Diffusion of innovations (4th ed.). New York, NY: Simon and Schuster.
Solomon, M. R. (2014). Consumer behavior: Buying, having, and being. Englewood Cliffs, NJ: Prentice Hall.