1
MODERN DAY MUSIC INDUSTRY:
THE ROLE OF MONEY IN MAKING
SUCCESSFUL ALBUMS
Jeroen Doorakkers (10708073)
Bachelor Thesis in Economics 26-06-2018
2 Statement of Originality
This document is written by Student Jeroen Doorakkers who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Table of Contents
Abstract ...4
1. Introduction ...5
2. Literature review ...7
2.1 The rise of technology ...7
2.2 Talent does not necessarily equal success ...7
2.3 Major labels versus independent labels ...8
3. Conceptual framework ...9 3.1 Factors ...9 3.1.1 Success ...9 3.1.2 Money ...9 3.1.3 Technology ... 10 3.2 Assumptions ... 10 3.2.1 Billboard success... 10 3.2.2 Parent company... 11 3.2.3 Revenues ... 11 4. Methodology ... 12 4.1 Data collecting ... 12 4.1.1 Database ... 12 4.1.2 Sample selection ... 12 4.2 Model specification ... 13
4.2.1 Ordinary Least Squares regression ... 13
4.2.2 Variable determination ... 14 4.2.3 The model ... 15 5. Results ... 16 5.1 Homoscedasticity ... 16 5.2 OLS regression ... 16 5.3 Robustness ... 18 6. Discussion ... 21 6.1 Previous literature ... 21 6.2 Explanations ... 21
6.3 Limitations and further research ... 22
7. Conclusion ... 23
4 Abstract
The music industry experienced an important change over the past few years. With the rise of modern day technology, streaming came along and so a new kind of music revenue was created. Spotify became the largest streaming provider in the world. The effect that this rise has on the impact of money spent on an album regarding the success of the corresponding album has yet to be discovered. In this research, I examine and analyse the effect of this particular rise. In order to do so, during this research, data was obtained through Billboard. From this data, a sample was selected, containing 250 albums from a four year period. This period spans the two years before the launch of Spotify and the two years after the launch of Spotify. The results that the regression analysis, using this data, presented indicate that after the launch of Spotify money began to play a bigger role in making a successful album. To further investigate this topic, extended research is needed.
5
1. Introduction
Every year many music albums are released, but only a few make it into the Billboard charts. These charted albums generate the largest share of the profit for labels (Bhattacharjee et al., 2007, p. 1361). Most of these charted albums are from artists represented by major labels, since about 70% of the total recorded music revenue market share is in the hands of major labels (Music and Copyright, 2014). However, the rise of new technology had multiple effects on the music industry. It is commonly stated that this rise made it easier for independent labels to create a large fan base in a short amount of time. For example, Bhattacharjee et al. (2007, p. 1372) stated that, with embracing the use of new technologies to brand and to attract potential customers, independent labels are able to close the gap with major labels. But to what extent is this true?
It is also often claimed that only artists signed to major labels are able to generate a large fan base and get to sell their music on a worldwide level, because major labels have more capital potential. Because the entry of new firms to the music industry is restricted by entry barriers such as product distribution and promotion (Alexander, 1997, p. 208), this capital potential can be useful. Furthermore, the large-scale promotional activities on a per unit basis are also more costly for smaller firms than for larger firms (Alexander, 1997, p. 208). On the contrary, there are many stories of artists of independent labels with success. Horowitz (2013, p. 21), for example, told the story of hip-hop duo Macklemore & Ryan Lewis, who started their own independent label and released their debut album ‘The Heist’, which peaked at number one on the Billboard 200 chart. This contrast created a debate whether it is necessary to be signed to a major label in order to be successful.
Many independent labels want to know how to fight the high concentration of output among a few major labels. On the other hand, major labels want to preserve this high concentration, because they can earn higher profits by decreasing the diversity of music (Black & Greer, 1986, p. 13). Finally, the music industry is growing. For example, between 2010 and 2015, music revenues increased by 30% (Naveed, Watanabe & Neittaanmäki, 2017, p. 2) and this increase continued: in 2016 it increased by 11.4% (Friedlander, 2017) and in 2017 it increased by 16.5% (Friedlander, 2018). Because of this growth, in 2017 the music revenues account for 8 723 millions of dollars (Friedlander, 2018). Apart from that, the
6
music industry has been heavily impacted by new technology (Bhattacharjee et al., 2007, p. 1359). This means that the impact itself is likely to contribute to economic knowledge.
These topics combined instigated a debate on whether the role of money in making a successful album has changed with the rise of new technology. To investigate this debate, this research has as primary goal to answer the research question: “What role does money play in making a successful album, considering the rise of modern day technology?”
The structure for the remainder of this research will be the following: the research starts with a literature review, wherein existing literature relevant to this topic will be evaluated and integrated. Second, a conceptual framework, where the determinants will be clarified and assumptions will be explained. Third the methodology will be discussed. The methodology will show how the database was obtained, as well as specify the model itself. After that, some results will be revealed and the research will end with a discussion and conclusion.
7 2. Literature review
In this section, an overview will be provided of the existing literature relevant to this topic. First of all the consequences of the development of new technology will be discussed. Subsequently the phenomenon of musicians with a lot of talent but not a lot of success will be considered, along with some possible explanations. Finally, there will be a sub-section on the effects that major labels and independent labels can have on success of an album.
2.1 The rise of technology
In the music industry, the presence of entry barriers such as product distribution and promotion has restricted the entry of new firms (Alexander, 1997, p. 208). However, the rise of new technology may affect these entry barriers. There are various opinions regarding this rise. Bhattacharjee et al. (2007, p.1372) stated that with embracing the use of new technologies independent labels are closing the gap with major labels, by using these new technologies to brand and to attract potential customers. But Bell (1999, p. 83) questioned whether the major labels would allow this and mentioned the possibility of major labels and internet companies selling music together. Alexander (2002, p. 151) had a different opinion and stated that the rise of new technologies facilitate the free exchange of music and, doing so, may undermine the structure of the music industry. But these digital technologies could also have a positive impact on music sales, because consumers use these technologies in their purchase decisions and because of this, they might purchase more music (Fader, 2000, p. 2). Finally, with the rise of new technology came the rise of digital streaming. Streaming allows consumers to listen to music directly from the internet at a fixed monthly payment. Streaming became the largest source of United States music industry revenues in 2015 (Friedlander, 2016).
2.2 Talent does not necessarily equal success
Rosen (1981, p. 846) addressed a situation where small differences in talent result in large differences in success. This shows that there are musicians with a lot of talent but not a lot of success. A lot of literature explaining this situation exists, many adhering to different arguments. Adler (1985 p. 212) explained an argument where consumption requires knowledge, in which this accumulation of knowledge by a consumer contains discussion with other consumers; this discussion exerts fewer difficulties when all participants share
8
common prior knowledge. Secondly, Chung & Cox (1994, p. 771) used an argument of luck, wherein music output is divided among a small number of lucky individuals. Lastly, Alexander (1994, p.96) uses as a third argument that entry barriers may have caused an increased industry concentration and less new product releases than there would be if there was a competitive market structure.
An advantage of already being a successful artist is the ‘spillover effect’, where artists sell more albums because of the performance and knowledge created in prior albums (Hendricks & Sorensen, 2009, p. 366). The trick here is that every album matters, because if one album contains poor quality, it will damage the success of later albums (Situmeang, Leenders & Wijnberg, 2014, p. 1482). The musicians that do not already possess a lot of success are often not discovered by major labels, so to still pursue a career in the music industry they may start their own independent label.
2.3 Major labels versus independent labels
For the official definition of major label, the Association of Independent Music (AIM) was contacted through e-mail. Callum Johnson, Membership Administrator of AIM, stated in a responding e-mail: “A ‘major’ is defined in AIM’s constitution as a multinational company which (together with the companies in its group) has more than 5% of the world market(s) for the sale of records (…)” and that “The majors are (currently) Sony [Music Entertainment], Warner [Music Group] and Universal Music Group”. These three major labels currently account for about 70% of the total recorded music revenue market share (Music and Copyright, 2014). It is commonly stated that a major label has positive effects on the success of an album. This because major labels often have a wide network and high capital which allows easier access to more customers by helping artists record, promote and distribute (Im, Song & Jung, 2018, p. 4). Conversely, independent labels - Callum Johnson stated: “An independent label is defined in AIM’s constitution as ‘not a major nor 50% or more owned by a major’” - have a willingness to take more risks, exploit niche markets, have extensive knowledge of local repertoire and have closer relationships with artists (Stroble & Tucker, 2000, p. 116).
9 3. Conceptual framework
3.1 Factors
In order to construct this research, the determinants must be clarified. In other words, the topics that play a role in the research must be defined. In this research, I try to answer the question: ‘What role does money play in making a successful album, considering the rise of modern day technology?’ This answer must consist of three important factors: success, money and technology.
3.1.1 Success
In traditional terms, the success of an album is determined by the amount of albums sold (Im, Song & Jung, 2018, p. 1). But with the arrival of streaming - which currently represents about 65% of the United States music industry revenues (Friedlander, 2018) - this also must be accounted for. The Billboard 200 chart is a ranking of the most popular albums of the week, based on sales data combined with streaming data. Furthermore, appearing on a Billboard chart has an important impact on the perceptions, awareness and profits of albums (Bradlow & Fader, 2001, p. 368). Therefore, having an album on the Billboard 200 chart can be considered as a primary goal for major labels and independent labels (Stroble & Tucker, 2000, p. 113). The above makes the peak position on the Billboard 200 chart a good indicator for success and will therefore be used as the dependent variable in the model.
3.1.2 Money
Artists who release an album are either signed to a major label or an independent label. Major labels often have stronger capital, which can be used for helping artists record, promote and distribute in order to gain more customers (Im, Song & Jung, 2018, p. 4). On the other hand independent labels often do not have access to enough capital to execute certain ideas (Elberse, 2013, p.74). Because of this difference in capital, label size is an indicator for the amount of money spent on an album and therefore will be used as an independent variable in the model.
10
3.1.3 Technology
In this research, the rise of modern day technology will be considered. The first idea to extend this research was to look at the past ten years and determine how the gap between large and small labels has changed. However the main problem with this idea is that there are many other factors that change over a period of ten years, other than just technology. To prevent the research from this omitted-variable bias, the ten-year period will be reduced to a period of only four years. Furthermore, to really focus on the rise of technology there will be looked at an important shock of technology in the music industry, namely the launch of streaming provider Spotify. In 2017, the United States music industry revenues consisted for 65% of streaming revenues (Friedlander, 2018). This combined with the fact that, in the same year, Spotify was the biggest streaming provider with 100 million customers spread over 60 countries (Datta, Knox & Bronnenberg, 2018, p. 6), makes Spotify an important company in the music industry and therefore the launch of Spotify a good indicator for the rise of technology. The four years will be split into two years before the launch of Spotify and two years after the launch of Spotify.
3.2 Assumptions
In order to create a model with constancy, some assumptions have to be made. The determinants success, money and technology are theoretical. In this research, I take some assumptions about the determinants into account, to successfully convert them into numerical variables.
3.2.1 Billboard success
The peak position on the Billboard 200 chart will be used as an indicator for success, based on sales data and streaming data. Many artists see success in a different way. Singer and songwriter Sia Furler, for example, said that she does not want the fame that charting albums bring, because it caused her burn-outs, so she is not targeting for the charts and the commercial success but for her success is to do what she loves (Furler, 2013, p.21). In this research however, I assume despite these differences in the definition of success, that every artist considers a Billboard 200 charting album as success and that an album with a higher peak position has more success than an album with a lower peak position.
11 3.2.2 Parent labels
Because Callum Johnson stated that an independent label is defined in AIM’s constitution as ‘not a major nor 50% or more owned by a major’, in this research the assumption will be that if a major label acquires at least 50% of a label, the label is being owned and controlled by the major label. This can lead to a chain of parent labels where many labels stand between one minor label and one of the three major labels. In this research, I assume that if a label is owned by a major label, it has the same capital potential as the major label itself.
3.2.3 Revenues
To use label size as an indicator for the amount of money spent on an album, label size must also have its measurement. Label size could be measured in many ways, but since the purpose of label size is to indicate the amount of money spend on an album, in this research label size will be expressed in terms of money. Therefore, I assume that label size is determined only by the revenue of that corresponding label.
12 4. Methodology
4.1 Data collecting
In order to perform my analysis, raw data had to be collected and reshaped into an appropriate form for analysing. To do so a couple of steps had to be taken. The structure of this section is divided into a subsection about how the database itself is collected and a subsection how it is transformed.
4.1.1 Database
The Billboard 200 chart is a weekly updated list. There are 52 editions per year. This research focuses on two years before the launch of Spotify - the 7th October 2008 - and two years
after the launch of Spotify. The last edition before the launch of Spotify is the one on the 4th
October 2008 and the first edition after the launch of Spotify is the one on the 11th October
2008. All together this means that 104 weeks before the launch of Spotify, beginning with the one on the 4th October 2008 and ending with the edition on the 14th October 2006, and
104 weeks after launch of Spotify, beginning with the one of the 11th October and ending
with the edition on the 2th October 2010, will be taken into account. This creates a
time-period beginning on the 14th October 2006 and ending on the 2th October 2010. In each of
these 208 week editions, 200 albums are listed. This creates a total of 41 600 albums. Billboard however does not grant just anyone access into their financial data. To gain access, first of all, Billboard was contacted through e-mail. This contact eventually led to the purchase of a Billboard subscription on ‘Billboard biz’, which allows access into the financial data of Billboard and by this the 41 600 albums with corresponding labels.
4.1.2 Sample Selection
For this research a sample selection is used. From the 41 600 albums, 250 albums were randomly selected. First of all, each album was assigned a number, from 1 to 41 600, and with www.random.org, one by one, 250 albums were selected and alternately converted into an excel sheet. Secondly, some of the 250 had to be removed. This because the Billboard 200 chart also allows albums that are released in past years to remain or re-enter the charts and because this research focuses on a four year period, albums that were released before or after this period had to be removed. The Billboard 200 chart however did
13
not provide the release date of the albums, so from the 250 albums the release date was searched and alternately linked to the corresponding album in the excel sheet. This made it possible to delete the albums that were released outside of the four-year period. After removing these albums, the remaining sample selection consisted of 223 albums. Of these 223 albums, 100 were released after the launch of Spotify and 123 were released before the launch of Spotify. Thirdly, Billboard did link albums to their corresponding labels, but not to every parent label. So from the 223 albums, the parent label of every corresponding label was searched and alternately linked in the excel sheet. From the 223 albums, 160 albums where traced back to one of the major labels, Sony Music Entertainment, Universal Music Group or Warner Music Group, or to a label owned by one of these major labels. The remaining 63 albums were from an independent label. Finally, all labels must be measured in label size. To do so, revenues of all labels were searched and linked to the corresponding labels. If a label had a parent label, the revenue of that parent label was linked to the label, because we assume that if a label is owned by a major label, it has the same capital potential as the major label itself. Because not all labels publish their financial statements every year, estimated revenues are taken into account for those labels that had not published their financial statements through www.owler.com. After these steps were accomplished, the sample selection was completed.
4.2 Model specification
To continue this research, I have to create a model. First the characteristics of the data have to be analysed to determine what kind of model will be used. Second, the variables must be clarified and labelled. These two sub-sections lead to the final sub-section: the model itself.
4.2.1 Ordinary Least Squares regression
In this research, I consider the peak position on the Billboard 200 as the dependent variable and I consider label size and the launch of Spotify as dependent variables. The launch of Spotify is a shock in time and does not directly affect label size. This indicates that, in this case, multicollinearity is out of order. These characteristics of the variables lead to the use of the ordinary least squares regression (OLS regression), but the use of this regression also creates a discussion about its potential problems. For example, if an album reaches a high peak position on the Billboard 200 chart, it possibly increases the revenues of the label that
14
released the album and by doing so also the independent variable label size. If this is indeed the case, than not only changes of the independent variables cause changes in the dependent variable, but also changes in the dependent variable cause changes in the independent variables. This is called simultaneity. Simultaneity is a cause of endogeneity, which means that an independent variable is correlated with the error term. Whenever endogeneity occurs, the OLS coefficient estimates will be biased. Furthermore, an OLS regression shows correlation relationships, since it creates a linear function of the parameters. However, correlation does not necessarily indicate causation. In order to talk about causation relationships, the research design must show causal conclusions. Despite these potential problems, the OLS regression will be used in this research.
4.2.2 Variable determination
The model that will be used in this research consists of three variables. Because the goal of this research is to examine the impact of money spent on an album regarding the success of the corresponding album considering the rise of modern day technology, the success of an album becomes the dependent variable of the model and money spent on an album considering the rise of modern day technology will create independent variables. First of all: the dependent variable “success”. Success is as mentioned above in this research determined by the peak position on the Billboard 200 chart. So the dependent variable in this model will become this peak position and will be labelled 𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖. Secondly, the
independent variable: “money”, which will be determined by label size. A continuous variable will be created, labelled 𝐿𝐴𝐵𝐸𝐿𝑖, which will be expressed in revenue of that label in
millions of dollars. Thirdly, the independent variable: “technology”, which will be determined by the launch of Spotify. A dummy-variable will be created, labelled 𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖,
which in case of the release of an album after the launch of Spotify will have the value of one and in case of the release before the album will have the value of zero. At last, in this research, I want to test the hypothesis whether the relationship between label size and peak position is different after the launch of Spotify than before the launch of Spotify. To do so, an interaction variable will be constructed. This will be done by multiplying the two independent variables and will be labelled 𝐿𝐴𝐵𝐸𝐿𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖. To provide a clear overview of
15 Table I. Variable Overview
Variable Description
BILLBOARDi Peak position on the Billboard 200 chart, measured by the highest
position on the Billboard 200 chart the corresponding album ever reached.
LABELi The label size of the album, measured in revenue of the corresponding
label in millions of dollars.
SPOTIFYi Release date after the launch of Spotify, dummy variable.
LABELSPOTIFYi Interaction variable between the above two variabels LABELi and SPOTIFYi
4.2.3 The model
In the previous sub-sections it became clear that in this research, I will follow the approach of the OLS regression. Also the variables were clarified and labelled. The combination of the OLS regression approach and the determined variables will create the following model:
𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖 = β0+ β1∗ 𝐿𝐴𝐵𝐸𝐿𝑖 + β2∗ 𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖+ β3∗ 𝐿𝐴𝐵𝐸𝐿𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖+ ε𝑖
In this model ε𝑖 is an error term and β0, β1, β2 and β3 are parameters. First of all the
parameters of the model will need to be estimated using a regression analysis. To do so a linear function of the parameters will be created. After that, the model will be used to look whether the impact of label size on the peak position on the Billboard 200 chart significantly changed after the launch of Spotify compared to before the launch of Spotify.
16 5. Results
This section is divided into three sub-sections. The first sub-section will test the assumption of homoscedasticity. In the second sub-section, the results of the OLS regression will be analysed. In the last sub-section, an additional analysis will be performed in order to add robustness to the results.
5.1 Homoscedasticity
In order to perform the OLS regression, first the assumption of homoscedasticity has to be tested, because if this assumption is violated the estimations are less precise and the p-values are lower than they should be. To test this assumption, the White test for heteroscadiscity is used. This test tests the null hypothesis that the model contains homoscedasticity against the alternative hypothesis that the model contains restricted heteroscedasticity. The White test resulted in a chi-square value of 8.50 and a p-value of 0.1309. This indicates that the null hypothesis is not rejected and that the model has homoscedasticity. Thus, the conclusion can be made that the assumption of homoscedasticity is satisfied.
5.2 OLS regression
Because the assumption of homoscedasticity is satisfied, the OLS regression can be run. To clarify, this regression estimates the following model:
𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖 = β0+ β1∗ 𝐿𝐴𝐵𝐸𝐿𝑖 + β2∗ 𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖+ β3∗ 𝐿𝐴𝐵𝐸𝐿𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖+ ε𝑖
The OLS regression resulted in the following table, Table II, which contains the number of observations, the R-squared and the parameter estimates with corresponding standard errors:
17 Table II. Regression Estimation
(1) VARIABLES billboard label 0.000393 (0.00178) spotify 23.07** (9.795) labelspotify -0.00682*** (0.00259) Constant 25.87*** (6.761) Observations 223 R-squared 0.051
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
The R-squared measures the scatter of the observed points around the linear regression line that is estimated. The results show an R-squared value of 5.1%. This means that the independent variables explain 5.1% of the variation of the dependent variable. This low percentage does not necessarily mean that the model is bad. If the independent variables are significant, it is still possible to draw conclusions about the relationships between variables.
In Table II, the asterisks behind the coefficients indicate the results of testing a null hypothesis. This null hypothesis states that the corresponding coefficient is equal to zero. The results show that, except for the variable 𝐿𝐴𝐵𝐸𝐿𝑖, all variables are significant at a
significance level of 0.05. This means that the variables 𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖 and 𝐿𝐴𝐵𝐸𝐿𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖
have a significant effect on the dependent variable 𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖. More specifically, the
variable 𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖 itself has a negative effect on the peak position on the dependent
variable 𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖 and the interaction variable 𝐿𝐴𝐵𝐸𝐿𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖 has a positive effect on
the dependent variable 𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖. The fact that the variable 𝐿𝐴𝐵𝐸𝐿𝑖 in this test resulted
not significant does not mean that the null hypothesis is accepted. This result only means that the null hypothesis is not rejected. The fact that the variable 𝐿𝐴𝐵𝐸𝐿𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖 has a
positive significant effect on the dependent variable 𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖 indicates that label size
has a positive effect if the corresponding album is released after the launch of Spotify. More specifically: if an album is released after the launch of Spotify, an increase in label revenue of, for example, 1 000 millions of dollars would result in a higher peak position of almost 7
18
positions. The differences in revenue between major labels and independent labels are up to 5 200 millions of dollars in this research, which would result in a higher peak position of 35 positions.
To put these results in words, these estimations indicate that an album from an independent label released after the launch of Spotify will reach a lower peak position than an album from an independent label released before the launch of Spotify. This is so because even the positive effect of the interaction variable of the independent label with the largest revenue does not compensate for the negative effect of the dummy variable. If an album is from a major label, the results are different. In this case the total effect depends on how large the major label is. If the positive effect of the interaction variable is larger than the negative effect of the dummy variable, an album from a major label released after the launch of Spotify will reach a higher peak position on the Billboard 200 chart than an album from a major label released before the launch of Spotify. But if the positive effect of the interaction variable is smaller than the negative effect of the dummy variable, an album from a major label released after the launch of Spotify will reach a lower peak position on the Billboard 200 chart than an album from a major label released before the launch of Spotify.
If these results are combined, the main result is that label size has a bigger impact on the peak position on the Billboard 200 chart after the launch of Spotify than before the launch of Spotify. To translate this back to the original research question, this means that because of the rise of new technology, money is playing a bigger role in making a successful album than before this rise.
5.3 Robustness
In order to add robustness to the results, an additional regression has been run. In this regression the model is changed. Instead of the dummy variable 𝑆𝑃𝑂𝑇𝐼𝐹𝑌𝑖, which looked at
the launch of Spotify, another shock in time within the four-year period will be observed. For this additional regression, the dummy variable 𝑆𝑃𝑂𝑇𝐼𝐹𝑌2𝑖 is used. This dummy variable is
used to indicate whether an album is released after the 9th September 2009. This is because
over time Spotify grew to become the biggest streaming provider with 100 million customers spread over 60 countries (Datta, Knox & Bronnenberg, 2018, p. 6). On the 9th September
19
point in time Spotify became a more important company in the music industry compared to the 7th October 2008, when it was launched. This has to do with various characteristics that
can be attributed to a (mobile) phone. First of all, phones offer a new channel to gain customers (Ström, Vendel & Bredican, 2014, p. 1001). Second, a phone differs from other electronic devices in mobility and personal characteristics and is constantly near the customer (Shankar et al., 2010, p. 112), which allows the customer to use Spotify more often. Finally, because of this personal nature, phones offer an opportunity to create a relationship between customers and Spotify (Ström, Vendel & Bredican, 2014, p. 1001). To clarify, the additional regression estimates the following model:
𝐵𝐼𝐿𝐿𝐵𝑂𝐴𝑅𝐷𝑖 = β0+ β1∗ 𝐿𝐴𝐵𝐸𝐿𝑖 + β2∗ 𝑆𝑃𝑂𝑇𝐼𝐹𝑌2𝑖+ β3∗ 𝐿𝐴𝐵𝐸𝐿𝑆𝑃𝑂𝑇𝐼𝐹𝑌2𝑖+ ε𝑖
The OLS regression resulted in the following table, Table III, which contains the number of observations, the R-squared and the parameter estimates with corresponding standard errors:
Table III. Additional Regression
(1) VARIABLES billboard label -0.000995 (0.00142) spotify2 29.09** (12.15) labelspotify2 -0.0104*** (0.00336) Constant 31.17*** (5.440) Observations 223 R-squared 0.062
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
The results of this regression are very similar to the results of the previous regression, wherein the dummy variable for the launch of Spotify is used. What stands out in these results is the coefficient of the interaction variable. This coefficient resulted in this additional regression in -0.0104, while in the previous regression it resulted in -0.00682. This indicates
20
that after the adding of Spotify to the Apple app store, label size had a bigger positive effect on the peak position on the Billboard 200 chart. To translate this back to the original research question, this emphasizes the previous conclusion that because of the rise of new technology, money is playing a bigger role in making a success album than before this rise.
21 6. Discussion
In this section the results of the OLS regression will be discussed and compared to previous literature. Also there will be some possible explanations for the results provided. And at last there will be an option for further research provided.
6.1 Previous literature
Bhattacharjee et al. (2007, p.1372) found the result that albums from independent labels experienced a positive shift after 2000 due to some technological developments. These results are contradictory with the results from this research, since I showed in this research that after the launch of Spotify, label size had a bigger effect on success. This could be translated into a negative shift for albums from independent labels. There are many possible explanations for this contradiction. First of all, the time period used by Bhattacharjee et al. (2007) is a different time period than the time period used in this research. Although in that time period technological changes also occurred, the changes were different from the changes that were analysed in this research. It is possible that various technological shocks cause different reactions. Second, Bhattacharjee et al. (2007) included more variables in their model. For example: gender, debut rank on the chart and whether the artist was in a group or not. Because in this research I only focused on label size and the launch of Spotify, the results could differ. A third possible explanation is that Bhattacharjee et al. (2007) did not used the same indicator for success. Bhattacharjee et al. (2007) did not use peak position on the Billboard charts, but the number of weeks an album spent on the Billboard charts. Albums that reach the same peak position on the Billboard 200 chart do not all stay the same amount of weeks on the chart. So this could also be an explanation for the contradiction.
6.2 Explanations
In this research I examined the role of money in making a successful album, considering the rise of modern day technology. It resulted in the conclusion that because of this particular rise, the launch of Spotify, money began to play a bigger role in this process. A possible explanation for this result is the fact that because of the launch of Spotify it became easier to try to get your music digitally distributed. Because of this, many people tried to do so and thus, an overload of new music was distributed. In this situation, the only way to stand out
22
could be to sign with a major label. Another explanation for this result could be that major labels have more employees and specialists than independent labels. Because of this, these employees and specialist can evaluate possible shocks in the industry and their effects, also the effects of the launch of Spotify. Because of these evaluations, the major labels can react to shocks faster than the independent labels. A third explanation could be that Spotify offers more opportunities for major labels. In other words, an album that is already highly promoted could become even more successful with the help of Spotify, because through Spotify it becomes even easier to listen to the album and share it with others.
6.3 Limitations and further research
Although the research focuses on a time period of four years to prevent the research from an omitted-variable bias, it is still possible that the results are affected by external factors. The success of albums is driven by different factors and is proven very difficult to measure (Franck & Nüesch, 2012, p. 202). For example, measuring only talent accurately is very difficult (Franck & Nüesch, 2012, p. 203) and even then it is arguable whether greater talent results in greater success (Adler, 1985, p. 208). To get rid of these difficulties, this research only focused on the impact of money spent on an album on success of an album considering the rise of modern day technology.
Furthermore, this research focuses on albums that have reached the Billboard 200 chart. Albums that did not reach this chart are not taken into account. This because Billboard did not provide this kind of data and to create it yourself would take a lot of time. In order to rank them, from every album the number of sales and streams must be searched and linked on the same way as the Billboard 200 chart does. Furthermore, the Billboard 200 chart alone contained 41 600 albums in the four year period. So to also take every other released album into account would create a very large database.
Finally, in this research I used the launch of Spotify as a technology shock. The question could rise whether this point in time is the best point to use. Spotify grew over time to become the biggest streaming provider with 100 million customers spread over 60 countries (Datta, Knox & Bronnenberg, 2018, p. 6). Another point in time that could be chosen as technology shock is the point where Spotify became the biggest streaming provider. Yet another point could be the point in time where streaming itself became a large
23
part of the music industry revenues. Streaming became the largest source of United States music industry revenues in 2015 (Friedlander, 2016). So 2015 could also be used.
Streaming is growing every year and every year it takes a larger part of the music industry revenues into account (Friedlander, 2018). This makes it very interesting to see what kind of effect this growing part has. While some studies showed that because of new technology minor labels are closing the gap with major labels (Bhattacharjee et al., 2007, p.1372), in this research I showed the opposite and concluded that after the launch of Spotify, the effect of label size on success of an album increased. In order to validate all statements, further research is needed. The ideal way would be to include all albums, meaning also the albums that did not reach the Billboard 200 chart and focus on several points in time. Furthermore, it might be ideal to use a panel regression instead of an OLS regression and to randomly take an album every year from the same labels. This would take a lot of time but will provide the best results about this topic.
7. Conclusion
The music industry nowadays is dominated by three major labels that use high capital to help an artist record, distribute and promote and only a few albums released by independent labels reach the charts. But the market is changing. Together with the rise of modern day technology, streaming came along. Streaming takes in a larger part of the music industry revenues every year. The primary goal of this research was to answer the research question: “What role does money play in making a successful album, considering the rise of modern day technology?”
In this research I found that the impact of label size, and so also money, increased after the launch of Spotify. So because of this rise of modern day technology, money plays a bigger role in making a successful than before. This result contributes to economic literature, because independent labels and major labels want to know whether it is necessary to spend a lot of money in order to create a successful album, and because of the fact that the music industry takes a notable part of nowadays economy into account. The results of this research indicate that further research is needed to explore this topic.
24 Bibliography
Adler, M. (1985). Stardom and talent. The American Economic Review, 75(1), 208-212.
Alexander, P.J. (2002). Peer-to-Peer File Sharing: The Case of the Music Recording Industy.
Review of industrial organization, 20(2), 151-161.
Alexander, P. J. (1997). Product variety and market structure: A new measure and a simple test. Journal of Economic Behavior & Organization, 32(2), 207–214.
Alexander, P. J. (1994). Entry barriers, release behavior, and multi-product firms in the music recording industry. Review of Industrial Organization, 9(1), 85-98.
Bell, C. (1999). Artists take it to the net. Billboard – The International Newsweekly of Music,
Video and Home Entertainment, 111(46), 63-83.
Bhattacharjee, S., Gopal, R. D., Lertwachara, K., Marsden, J. R., & Telang, R. (2007). The Effect of Digital Sharing Technologies on Music Markets: A Survival Analysis of Albums on Ranking Charts. Management Science, 53(9), 1359–1374.
Black, M., Greer, D. (1986). Concentration and non-price competition in the recording industry. Review of Industrial Organization, 3(2), 13-37.
Bradlow, E.T., Fader, P.S. (2002). A Bayesian Lifetime Model for the “Hot 100” Billboard Songs. Journal of the American Statistical Association, 96(454), 368-381.
Chung, K.H., Cox, R.A.K. (1994). A stochastic model of superstardom: an application of the Yule distribution. Review of Economics and Statistics 76(4), 771–775.
Datta, H., Knox, G., Bronnenberg, B.J. (2018). Changing Their Tune: How Consumers’
Adoption of Online Streaming Affects Music Consumption and Discovery. Marketing
25
Elberse, A. (2013). Blockbusters: Hit-making, Risk-Taking, and the Big Business of Entertainment.
Fader, P.S. (2000). Expert Report of Peter S. Fader, Ph.D. in Record Companies and Music Publishers vs. Napster, July 26. United States District Court, Northern District of California.
Franck, E., Nüesch, S. (2012). Talent and/or Popularity: What does it take to be a Superstar?
Economic Inquiry, 50(1), 202-216.
Friedlander, J.P. (2018). 2017 RIAA Shipment & Revenue Statistics.
http://www.riaa.com/wp-content/uploads/2018/03/RIAA-Year-End-2017-News-and-Notes.pdf
Friedlander, J.P. (2017). 2016 RIAA Shipment and Revenue Statistics.
http://www.riaa.com/wp-content/uploads/2017/03/RIAA-2016-Year-End-News-Notes.pdf
Friedlander, J.P. (2016). RIAA 2015 Year-End Sales & Shipments Data Report.
www.riaa.com/wp-content/uploads/2016/03/RIAA-2015-Year-End-shipments-memo.pdf
Furler, S. (2013). My Anti-Fame Manifesto. Billboard – The International Newsweekly of
Music, Video and Home Entertainment, 125(42), 21-22.
Hendricks, K., Sorensen, A. (2009). Information and the Skewness of Music Sales. Journal of
Political Economy, 117(2), 324-369.
Horowitz, S.J. (2013). One To Grow On. Billboard – The International Newsweekly of Music,
26
Im, H., Song, H., Jung, J. (2018). A survival analysis of songs on digital music platform.
Telematics and Informatics, https://doi.org/10.1016/j.tele.2018.04.2013
Music and Copyright. (2018). UMG and WMG make recorded-music market share gains, Sony outperforms in publishing. www.musicandcopyright.wordpress.com
Naveed, K., Watanabe, C., Neittaanmäki, P. (2017). Co-evolution between streaming and live music leads a way to the sustainable growth of music industry – Lessons from the US experiences. Technology in Society, 50, 1-19.
Rosen, S. (1981). The economics of superstars. The American economic review, 71(5), 845- 858.
Shankar, V., Venkatesh, A., Hofacker, C., Naik, P. (2010). Mobile Marketing in the Retailing Environment: Current Insights and Future Research Avenues. Journal of Interactive
Marketing, 24(2), 111-120.
Situmeang, F.B.I., Leenders, M.A.A.M., Wijnberg, N.M. (2014). The good, the bad and the variable. European Journal of Marketing, 48(7/8), 1466-1486.
Stroble, E.A., Tucker, C. (2000). The Dynamics of Chart Success in the U.K. Pre-Recorded Popular Music Industry. Journal of Cultural Economics, 24(2), 113-134.
Ström. R., Vendel, M., Bredican, J. (2014). Mobile marketing: A literature review on its value for consumers and retailers. Journal of Retailing and Consumer Services, 21(6), 1001-1012.
