Fans like what they hear: using social media analytics to improve value creation in the music industry.

The Internet has changed the music industry. Yet, its digitization impacted the music industry in more ways than just in form of declining sales as propagated by the media. This paper is concerned with the issue of how the digitization impacted the music industry in regards to marketing and how music labels can use social media analytics to adapt their value creation process to the new business conditions. Thus, it first outlines the music industry and its digitization. In a second step, the impact of the industry’s digitization on marketing is elicited. Concepts and applications of social media analytics are introduced in chapter three. Conclusively, possible applications of social media analytics are analyzed.

2. The music industry

The following refers to a very narrow definition of the music industry’s market. Gmelin, Günnel and Steinkrauß (2008) define A&R management, marketing and distribution as core competencies of businesses within this market. Yet, it is worth noting that the music industry also includes numerous upstream and downstream markets such as merchandising, advertising or concerts. Having said that, these upstream and downstream markets will go unheeded in the following remarks.

2.1 Impact of digitization

Since the turn of the millennium the music industry has been confronted with a recessional phase caused by the industry’s digitization (Tschmuck, 2008). The Internet abrogates the capitalistic logic of the market since the value of an offer does no longer increase proportional to its scarcity but rather to its demand (Poster, 2001). The ongoing digitization of music fosters a feeling of non-excludability and non-rival consumption that leads to a perception of music as a public good and therefore to a decline in music sales (van Dyk, 2008). For instance, Rob and Waldfogel (2006) estimate that each unpaid music download of an album reduces its sales by .1 to .2 units. Albeit the music industry has yet to find ways to benefit from its digitization rather than understanding it as a burden, Freedman (2004) has already recommended that record labels should change their response to the culture of music downloads from lawsuits to a marketing and promotion orientation.

2.2 Marketing in the music industry

The marketing mix is traditionally divided into the four areas product, price, distribution, and promotion (Meffert, Burmann & Kirchgeorg, 2008). These areas are elucidated in regards to the music industry hereinafter.

2.2.1 Product policy

Product policy is an essential part of the marketing mix since attractive music contents are vital for a music label’s economic success. The repertoire’s strategic orientation can be based on criteria ranging from geographic selections (e.g. local pop and international pop) to genres and sub-genres such as Mainstream Pop or Gangster Rap (Mahlmann, 2008). Furthermore, product related decisions include the configuration of individual products. This includes the selection of music contents, formats and packaging as well as the specification of meta data, technical details as well as tools for marketing and promotion (Mahlmann, 2008).

2.2.2 Price policy

Decisions in regards to pricing particularly involve the evaluation of a music product’s status, lifecycle and value in order to classify it into a standard price category such as deluxe-price, full-price, budget or development, from which an appropriate catalogue price is then derived (Mahlmann, 2008). Catalogue prices develop dynamically in accordance with the product lifecycle and therefore often feature a typical development from full-price to upper-mid-price to mid-price and possibly to budget as lowest price category (Mahlmann, 2008).

2.2.3 Distribution

Distribution policy deals with support of the various sales-relevant channels such as wholesalers or retailers (Caspar, Mucha & Wustlich, 2008). One of the main objectives of distribution is to ensure that “the right product is available at the right place at the right time, in a way the interested customer is able to perceive it” (Mahlmann, 2008). Most music labels have been outsourcing the physical distribution to service providers since the 1990s. This is not only caused by the labels’ increasing focus on promotional functions but also a constantly decreasing understanding of distribution as a competition-relevant criterion (Mahlmann, 2008). Therefore, in-house distribution departments focus on conceptualizing and implementing promotion campaigns for the different distribution channels. For example, these could be decorations on the shop floor or a magazine editorial (Mahlmann, 2008).

2.2.4 Promotion

Promotion in the music business involves advertising and promotion, whereby advertising is understood as the employment of paid means of advertising while promotion is defined as unpaid advertising through the use of music in media (Mahlmann, 2008). Classical examples of advertising are TV spots, radio and print advertising as well as out-of-home-advertising such as billboard and outdoor advertising (Mahlmann, 2008). Radio and television are exemplary channels for promotion. As an experience good, music bears consumption risks for its purchasers since the product’s quality cannot be evaluated prior to consumption (Albers, Clement & Schusser, 2008). Therefore, music played on television and the radio serves as a product sample that helps customers to reduce their consumption risks.

2.3 Impact of digitization on marketing

There is rapidly growing literature (Oberholzer-Gee & Strumpf, 2007; Zentner, 2006; Hong, 2007) on the impact of file sharing on sales and therefore on the music industry’s role as a “victim of the Internet revolution” (Preston & Rogers, 2013). Yet, the impact of the industry’s digitization acts beyond sales. And while the current state of research does not allow for apodictic predictions in regards to the digitization’s impact on sales, its effect on other areas of the industry can be identified. The digitization’s impact on the music industry´s marketing mix is to be analyzed in the following.

2.3.1 Product policy

In regards to product policy, the digitization predominantly leads to sales slowdowns (Mahlmann, 2008). Physical offerings such as CDs or vinyl records are produced less and less often while there is an increasing number of singles or albums that are at first or – in some cases – exclusively released digitally. This development has been encouraged by a rule change regarding the single-charts in 2007, according to which digital-only singles were now enabled chart entry (Mahlmann, 2008), thus making CD singles de facto obsolete.

2.3.2 Price policy

Changes in regards to labels’ price policies primarily correlate to the specific costs structure of music downloads as digital goods. As such they cause considerable costs upon production of the first copy while the variable costs of reproduction tend towards zero. This particular constellation facilitates a myriad of price strategies. Yet in most cases, music downloads underlie a homogeneous pricing. For instance, it is common practice to offer individual songs for $.99 on iTunes (Buxmann, Pohl & Strube, 2008). In consideration of the initial costs for deployment, licenses, margins and taxes, the economic sustainability of this pricing can be qualified as doubtful. Yet, most music labels lack the necessary negotiating power to set against iTunes’ prices. With a market share of more than 70% for single downloads in the United States, Apple’s iTunes platform is hard to ignore as a business partner in the music industry (Gmelin, Günnel & Steinkrauß 2008). Therefore, most music labels have to accept the dictated prices.
For this very reason music labels try to serve customer segments with diverse payment reserves when releasing full-length albums. This is often done by means of a qualitative price discrimination also referred to as versioning. In this instance, the basic version of an album is modified in respect of various features and then sold at a higher price as a new product variant (Buxmann, Pohl & Strube, 2008). Bundling constitutes another price strategy of equal importance. In this case, multiple individual tracks are bundled into a new product to deplete the existing variance in payment reserves for the individual songs (Buxmann, Pohl & Strube, 2008).

2.3.3 Distribution

With regards to distribution, the biggest impact of the digitization was the imposed necessity for music labels to establish non-physical structures of distribution. Unlike physical distribution, digital distribution does not entail costs for warehousing or returns. However, it requires a complex technical infrastructure that is most commonly offered by external service providers (Mahlmann, 2008).

2.3.4 Promotion

Promotion is the part of the market mix that has experienced the most intense reorganization in consequence of the music industry’s digitization. Analogous to traditional promotion, online promotion is defined as the unpaid presentation of artists and music on websites. Was the presentation paid for, it is to be considered online advertising or digital advertising (Mahlmann, 2008).
While television and radio have become increasingly less relevant as channels for promotion, the Internet has continuously gained importance. The Internet’s interactivity, the ability to engage in a dialog with customers, is of particular importance and democratizes the promotion of artists and music, thus facilitating whole new ways of promotion. Whereas consumers had been used to just listening to music for decades, a new and increasingly participatory consumer has been begotten by the digitization (Diederichsen, 2001). For music labels, this participation is a fundamental feature of nearly every digital marketing operation. The average consumer progressively refuses traditional marketing while proclaiming his own opinion about brands and products online (Schwerdt, 2005). Music labels can make the consumers’ willingness to participate and express work to their advantage. Internet promotion enables music labels to integrate consumers into the promotion and therefore to create advertising that has the potential to overcome the consumers’ frustration with advertising. Katz and Lazarsfeld (1955) proved as early as 1955 that personal recommendations, or word-of-mouth, was much more effective in influencing consumers’ behavior than mass media or personal selling. Particularly in view of the consumption risk associated with music described in 2.2.4, it is obvious why the effectiveness of personal recommendation and thus the integration of consumers into promotion campaigns are of enormous interest for music labels. It is essential for music labels and every other advertiser to know what is being said about their products and who the most influential users within a network are. This can be achieved by resorting to social media analytics.

3. Social media analytics

Social media has experienced exceptional growth in their user base in recent years. There is a myriad of existing social media applications and platforms and new ones are routinely brought to market. In general, social media applications can be categorized as weblogs (e.g. WordPress) and microblogs (e.g. Twitter), social networking sites (e.g. Facebook), location-based social networks (e.g. FourSquare), wikis (e.g. Wikipedia) discussion forums, podcast networks, picture and video sharing platforms such as Instagram or Youtube, ratings and review communities like Yelp, social bookmarking sites (e.g. Digg) as well as avatar-based virtual reality spaces as, for instance, Second Life (Stieglitz et al., 2014). Zeng et al. (2010) define social media broadly as “a conversational, distributed mode of content generation, dissemination, and communication among communities”.
Social media and its users generate large quantities of data of a highly dynamic and complex nature. Social media data exhibit both structured and unstructured data. Structured data, also referred to as meta-data, are composed of user profile characteristics, temporal, geographical and thematic data as well as attention-related data such as the number of likes, comments or retweets. On the other hand, unstructured data involve user-generated textual content of diverse complexity, for instance Facebook comments or audiovisual material.
Complexity is not the only challenge companies face when working with social media data. Many social media platforms do not provide standardized ways to access their data, such as APIs (application access interfaces). Therefore, different social media platforms require different methods of access and data collection (Stieglitz et al., 2014).
It has become a challenge for companies to monitor and interpret what people post on social media. Traditional content analysis methods quickly reach their limits when being faced with the tremendous amount of data produced within social media on a daily basis. Therefore, automatic methods to quickly analyze such amounts of data are increasingly needed by companies (He et al., 2015).
These required automatic methods are supplied by social media analytics. Social media analytics can be defined as methods, technical frameworks and software tools for tracking, modeling, analyzing, and mining large-scale social media data. In a business setting, social media analytics might be understood as a subset of business intelligence that is concerned with methods, processes and technologies that transform raw data from various social media sources into useful information for business purposes (Stieglitz et al., 2014). Two particularly important areas of social media analytics are text mining and social network analysis.

3.1 Text mining & sentiment analysis

Text mining is an emerging technology aimed at extracting meaningful information from unstructured textual data. Other than conventional content analysis, the principal task of text mining is to automatically excerpt insights, useful patterns or trends from a given set of textual content (He et al., 2015). Text mining is commonly used for categorization, information extraction as well as cluster and link analysis (Hung, 2012). Of particular interest for companies is a special application of text mining known as sentiment analysis. Sentiment analysis, also referred to as opinion mining, is concerned with the automatic extraction of positive or negative opinions from texts (He et al., 2015). For instance, sentiment analysis is used to study people’s opinions in regards to entities, individuals, products, events or topics. Technically, sentiment analysis can be performed based on two different approaches. The traditional approach resorts to a dictionary-based classification of sentiment orientation. In this case, the software employs to dictionaries of words, each annotated with their sentiment orientation. As a result, textual content can be evaluated on a dichotomous scale of positive and negative sentiment orientation. More detailed information can be extracted from textual content by using an approach based on machine learning, where the evaluation can resort to three classes: negative, positive and neutral. Notwithstanding the advantages of automated sentiment analysis, manual text analysis is still needed to test the findings due to the informal nature of social media textual content that often includes acronyms, slang or emoticons (Stieglitz et al., 2014).

3.2 Social network analysis

Social network analysis analyzes the structure of connections between persons, organizations, interest groups, states, etc. in order to study their relationships. Freeman (2011) characterizes social network analysis as an approach involving four defining properties: the intuition that links among social actors are important, its foundation of collecting and analyzing of data that record social relations that link actors, its drawing on graphic imagery to reveal and display the patterning of those links and its development of mathematical and computational models to describe and explain these patterns. Social network analysis focuses not on the attributes of individual users, but on relationships between individuals based nodes and links, whereas nodes are abstractions for individuals, organizations or communities and links represent various types of relationships (You, Liu, Xia & Lv, 2011). Social network analysis can help to identify influential users or opinion leaders. There is a myriad of measures for the influence of an actor. Within the context of social media, data is commonly used as a proxy for influence. For instance, a user’s amount of followers and mentions could be used as an indicator of his influence within his network.

4. Possible applications for music labels

While the digitization of the music industry imposed numerous changes and challenges as outlined in 2.3, music labels run a chance to turn these changes into opportunities. Social media analytics can be used to a special degree to counteract the negative development in regards to product and promotion policies imposed by the digitization.

4.1 Product policy

As outlined in 2.3.1, in regards to product policy, the digitization leads predominantly to a decrease in sales and the concomitant decline in production of physical sound recording mediums such as CDs or vinyl records. But the increasing distribution of digital sound recordings also simplifies the artists and repertoire management as well as the discovery of new artists. While it used to take an expensive demo recording to evaluate an artist’s potential, music labels can now appraise an artist’s acceptance in the digital value chain before they physically distribute and promote him (van Dyk, 2008). Resorting to text mining and sentiment analysis, they can collect and assess opinions expressed by social network users, discussion board members and visitors of music websites. Depending on the elaborateness of the feedback and the level of knowledge of the users (e.g. YouTube comments in comparison to comments on the Billboard Magazine website), music labels may soon be able to use the results to improve a recording according to the listeners’ feedback. The ability to release inexpensive demo recordings of newcomer bands online, to automatically assess their feedback and utilize it to improve an artist’s performance or image before distributing him physically, would implicate tremendous economic advantages. Since most record labels loose money on many new releases, a small number of successful artists and releases need to finance a large number of less successful ones. For example, only 20 per cent of all CD-releases in the United States were profitable in 2006 (Altig, Clement & Papies, 2008). Therefore, being able to make a projection of an artist’s success based on automatically assessed feedback of hundreds or thousands of listeners, could reduce development costs and thus entrepreneurial risks dramatically.

4.2 Promotion policy

Music labels can employ social media analytics in different ways to adapt to the changes in promotion policy imposed by the industry’s digitization. Due to the interactive nature of web 2.0 and social media in particular, music labels and artists nowadays are in constant dialogue with their customers and fans. Social media analytics by means of text mining, sentiment analysis and continuous social media monitoring enable music labels to track and engage in online discussions related to their artists and repertoire. Moreover, social media analytics can be used to track the spread of links and content and thus allow music labels to evaluate the effectiveness of certain promotion actions. For instance, the navigation path of a website visitor can be traced through click-stream data and used to evaluate the outcomes of exposure to a certain marketing action (Wilson, 2010).
The most fundamental and effective area of application of social media analytics for music labels in regards to promotion is the identification of influencers by means of social network analysis. Influencers can be defined as users that have found a way to channel their popularity and reputation within a network into collective action. Consumers’ declining faith in traditional media as well in traditional advertising and promotion creates the need for new means of promotion. Friestadt and Wright (1994) detected customers develop comprehensive knowledge about advertising practices with advancing media consumption and therefore reject traditional advertising. Music labels can bypass this dismissal by working with influencers that spread their promotional content. Contrary to traditional advertising, word-of-mouth-activities allow for the exchange of product-related information on the basis of mutual trust. The interactivity fostering structure of social media and web 2.0 simplifies the distribution of word-of-mouth-activities. Once a message is composed, it can be published to and shared by the digital following of an influencer at the click of a mouse. That way, music labels can make fans perform certain parts of the value creation process of promotion and inexpensively potentiate their content reach. Music labels ought to focus on influencers when seeking to generate awareness for a new artist or release since there is a positive correlation between the size of an influencer’s network and his connectivity and the effectiveness of his word-of-mouth (Katona, Sarvary & Zubcsek, 2011).
Furthermore, influential users should receive advance copies of new releases and be asked to review them. Product reviews are a specific kind of user-generated content. In general, user-generated content is defined as “content made publicly available over the internet, […] which reflects a certain amount of creative effort, […] and which is created outside of professional routines and practices” (OECD, 2007). Product reviews as user-generated content are not only representative for the newly acquired consumer-power but also a very subtle way influencers can take over parts of the value creation process of promotion. Even though the consumption risk described in 2.2.4 can be minimized by the possibility of listening to audio clips prior to a purchase, the process of discovering and selecting new music is very time consuming while only offering a small chance of success. This situation is aggravated by the decrease of consulting competence of retailers. With this in mind, the effectiveness of product reviews by influencers is evident. Marketing literature abounds with examples for the positive correlation between product reviews and sales (Chevalier & Mayzlin 2006; Chintagunta, Gopinath & Venkataraman 2010; Duan, Gu & Whinston 2008; Liu 2006). Product reviews are perceived as trustworthy. This perception is even unimpaired when consumers are explicitly pointed towards the possibility of manipulating product reviews (Bambauer-Sachse & Mangold, 2013). This high level of trust compared to traditional advertising as well as the verified positive correlation with sales demonstrate the importance of product reviews for music labels.
Social media analytics applications that combine sales data with social media data offer music labels even more opportunities to optimize their decision-making. Online mentions of an artist’s appearance at a late night show combined with sales data allow music labels to evaluate and predict the impact of such appearances. Online chatter about an artist’s guest appearance on another artist’s song combined with sales data allows music labels to evaluate the economic sustainability of the investment involved in such collaborations.

5. Conclusion

The available evidence suggests that while the music industry’s digitization had a significant impact on its value creation processes, it also facilitates music labels to improve and innovate certain processes. Social media analytics can be employed in various ways to make these changes more effective. In particular the identification of influencers by the means of social network analysis as well as the automatic assessment of opinions by the means of sentiment analysis were proven to be of tremendous importance in this paper. Continuous improvement of the available technology as well as the increase in quantity and wealth of detail of social media user data will allow music labels to optimize their value creation processes, minimize entrepreneurial risks and maximize reach in the future.