This study attempts to measure how much success non-mainstream genres and artists have had in crossing over onto the mainstream "pop" charts. The author examines Billboard magazine's year-end album charts and determines how well country, Latin, metal/hard rock, rhythm and blues, rap/hip-hop, and other genres outside of pop have fared. The author also looks at how changes in Billboard's chart compilation methods have affected the rankings.

Introduction

Although the popular music industry has tended to cater to certain sounds that fit into the mainstream, divergent genres have managed to "cross over" on to the pop charts from narrower charts. These genres include, among others, rhythm and blues (R&B); disco, which came to dominate the pop charts during the late 1970s; country; Latin; and rap. Until the early 1990s, a few artists from these genres--such as country singers Dolly Parton and Kenny Rogers--had found consistent mainstream success, but, as a whole, the music dominating the charts tended to be generic "pop." By the early '90s, however, it became clear that many artists from non-mainstream genres were reaching gold and platinum status. [1] This study attempts to measure how much success these non-mainstream artists have had in crossing over onto the mainstream "pop" charts.

In May 1991, Billboard, the popular music industry's most influential trade journal, began using SoundScan to track album sales. [2] Instead of relying on reports from retailers, rack-jobbers, and wholesalers, Billboard began monitoring exact unit counts using technology similar to the barcoding used in grocery stores. Only four months later, on 28 September 1991, Garth Brooks's third album, Ropin' the Wind, debuted at No. 1 on the "Billboard 200" album chart to become the first country album to reach No. 1 since Kenny Rogers' Greatest Hits in 1980 (Grein, "Chartbeat," 28 Sept. 1991: 12). Brooks's No. 1 debut was surprising not only because no country artist had ever debuted at No. 1, but also because new releases by mega-selling rockers U2 and Guns N' Roses were eligible to chart during the same week and came in behind Ropin' the Wind, which would remain at No. 1 for 16 weeks. In June 1992, country singer Billy Ray Cyrus's debut album, Some Gave All, began a 17-week run at No. 1, only to be replaced at No. 1 on 10 October 1992, by another Brooks album, The Chase ("Billboard 200," 13 June 1992: 102; "Billboard 200," 10 Oct. 1992: 98). [3]

Since this breakthrough, country artists ranging from the well-established (e.g. Reba McEntire, Alan Jackson) to relative newcomers (e.g. John Michael Montgomery, Trisha Yearwood) to rookies (e.g. Terri Clark, David Lee Murphy) have had albums make respectable showings on the "Billboard 200." Outside country music, artists in areas as diverse as rap/hip-hop (e.g. N.W.A., Salt-N-Pepa), metal/ hard rock (e.g. Alice in Chains, Nine Inch Nails), Latin (e.g. Ricky Martin, Christina Aguilera), New Age (e.g. Enya), and even comedy (e.g. Jeff Foxworthy, the Simpsons) have had major hits.

These successes are significant because the publication considers its "Billboard 200" the "catch-all" or "pop" chart, while country, R&B, rap, New Age, and other genres have separate charts. The way an album, single, or artist is classified matters, and not just to academics. Marketing/promotion decisions, image-creation, and radio station playlists are all built around how artists and their music are perceived. For instance, country singers who are "too country" have problems attaining airplay on pop stations. Rock band Evanescence, in an apparent attempt to shed its Christian image and reach broader audiences, asked that its 2003 album Fallen be deleted from Billboard's Top Christian Albums listing and not be sold at Christian retail stores. The album peaked at No. 3 on the "Billboard 200" and sold 3 million copies (Price 13; Mayfield, "Over" 55).

Despite criticisms of Billboard and other trade journals, their rankings remain important. For example, Tom McCourt and Eric Rothenbuhler noted that these charts are "crucial to the popular music industry" (203-04). Quoting Nancy Lanthier, they said, "Record companies, radio stations, retailers, artists, and songwriters use the charts when making decisions on everything from record marketing strategy to tour planning. Reputations are made--and sometimes marred--by how high individual records are ranked." Jim Melanson admitted that everyone from retailers and radio personnel to company executives and rock stars will criticize the charts--for valid reasons--but "no one goes so far as to ignore them. Looking good on the national charts can be awfully important," he concluded (109). [4]

This paper compares the 10 years preceding and following the conversion to SoundScan in 1991 to examine the differences on the "Billboard 200," the chart that ranks albums. Conventional wisdom before 1991 held that genres such as country, R&B, rap, and alternative music reached only a narrow, specialized audience, but Brooks's Ropin' the Wind, as well as later albums representing many non-mainstream styles, added evidence to the belief that several genres long considered step-children to mainstream pop had wider followings than originally assumed.

Examining Billboard data from 1991 to 2001, the author employs methods similar to those used by Alan Wells to assess which genres saw gains and losses after the advent of SoundScan. [5] The paper is actually trying to measure two phenomena: the successes and problems of "non-mainstream" genres and the impact of SoundScan.

Conceptual Bases and Previous Studies

While many scholars have written about music performed by non-mainstream artists, most have focused on the cultural and historical importance of these genres (see, e.g., Carr, Malone). Few studies have looked at the chart performance of marginalized groups.

Before looking at chart performance, however, it is necessary to look at how the Billboard chart has been compiled over the years. Peter Hesbacher, Robert Downing, and David G. Berger ("Record Roulette") noted that "the methods of measuring ... poses [sic] problems" (75). Although this system has been reformed several times since the article was written, at the time, only the 22 largest markets were monitored for airplay and sales data. Airplay data for singles were determined by 63 stations, and only those "with the highest ratings" were included (78). Each record was assigned points according to its rank on a station's playlist (only the top 30 records earned points), and the station was given a weighting factor based on its market size and rank within that market. Thus the highest-ranked stations in the largest markets had more input. Airplay points counted for one-third of a record's total (79).

Sales points for singles and albums were determined by 110 one-stop and retail stores in the 22 markets; outlets were chosen based on "unit or dollar volume, neighborhood (an indicator of music type), and financial solvency" (78, their parenthesis). Each record could receive points in two ways: sales rated as "very good," "good," or "fair" received 20, 10, or 5 points, respectively, while the top 15 in sales were given additional points according to position (79).

Although Hesbacher, Downing, and Berger did not discuss subjectivity at this point in their article, it is necessary to note that there was room for manipulating the figures at each step of the process. In general, relying on only 22 markets eliminated a large portion of the nation and did not account for non-urban tastes. As far as airplay, program directors determined the rankings, although no one seems to know how; this situation meant certain records could receive more or less airplay than actually reported. Listeners' requests may or may not have made a difference. The sales points were determined in an even more subjective way, as Hesbacher, Downing, and Berger noted that the "very good" ratings were subjective, thus could easily have been inaccurate, while the top 15 rankings might have been based on actual sales or on perceived sales. The fact that only certain stores were chosen also ignored varying tastes across neighborhoods.

More appalling than these subjectivities, however, are those that Hesbacher, Downing, and Berger say existed in the final compilation of the chart. After the total was determined for each record, the points were "subjected to additional conversion procedures, and the actual chart position may not [have] always be[en] that designated by grand total points. Applications of these conversion procedures depend[ed] upon specific characteristics of the recording" (80). While not listing what these "characteristics" were, the authors noted that several techniques, such as "smoothing," were used to "normalize" the life cycle of a record, and that these options were used at the discretion of "'chart' personnel" (80).

The charts themselves, as presented weekly in Billboard, also had the disadvantage of not being able to account for the gaps between the albums or regional strengths. It was impossible to tell, for instance, whether the No. 1 album in a given week outsold the No. 2 album by 100,000 copies, by 12,000 copies, or even by 1 copy. [6] In addition, it was never clear whether a single or album was at the top of the charts because it was uniformly popular across the nation or because it performed particularly well in a heavily weighted metropolitan area.

Nonetheless, Hesbacher, Downing, and Berger concluded in a follow-up article ("Sound") that the charts remain worthy of study not only because they represent "a vast data source" (14), but also because, in spite of their faults, they were--and remain--the best measurements of success in popular music (14-17).

This situation continued for the most part until 1991. Although the 1980s saw Billboard surveying more markets, including small and medium markets, the same subjective reporting and subsequent distortion were possible. However, in May 1991, the albums chart was converted to a new system, and the singles chart followed in November 1991. Billboard now uses computer technology called Broadcast Data Systems and SoundScan to tabulate its rankings. Broadcast Data Systems monitors airplay and provides "gross impressions," also called "audience impressions," both of which are jargon for the number of people listening to a station multiplied by the number of times the record was played. [7] SoundScan is able to compile actual sales data from approximately 11,400 locations (Lander 1; "Monitored Airplay" 5, 79). [8] The initial sample included 15 national and regional chains, including Musicland and Camelot, representing about 2,000 individual locations, about 300 independent retailers, and about 4,000 racked accounts (Lander 88). The conversion initially "caused a commotion in the industry," with complaints of an inadequate number of stores and an inappropriate mix of chains, independent stores, and rack-jobbers (Terry 1, 85). At the same time, however, Billboard itself noted that the changes made it more difficult for retailers and wholesalers to distort sales data (Rosen 14). Billboard has since revised its panel several times to reflect market realities (Mayfield, "Decade" 5). While selection of radio stations and stores is still subjective, it is far more representative than the previous system. All in all, the SoundScan/Broadcast Data Systems method appears more accurate than other methods Billboard has used to compile its charts.

An argument against SoundScan and Broadcast Data Systems comes from McCourt and Rothenbuhler, who fear the technologies "may discourage innovation in popular music" (201). Noting that corporate executives try to limit uncertainty and increase predictability in all facets of the popular music industry, McCourt and Rothenbuhler pointed out that the technologies give specific data and allow almost instant feedback. The data in turn allow quick decisions on who is selling and who is not, leading the authors to fear that "the likely result is that record companies and promoters will exploit 'proven' properties at the expense of fledgling performers in order to reduce uncertainty of demand" (208).

McCourt and Rothenbuhler also challenge SoundScan's choice of outlets, noting that, although SoundScan claims to be able to account for 85% of all sales across the nation, it is heavily weighted toward chain stores and discount/department stores, to the near exclusion of smaller chains without barcoding ability, independent record stores, and stores that sell records in addition to other merchandise (e.g. book stores). The store selection is a disadvantage to independent labels and non-mainstream music, as the authors note: "It is inarguable that independent stores set trends while chains and rack jobbers stock proven sellers. By focusing largely on chain and racked stores, SoundScan tracks retailers who follow the market, rather than those who start the market" (210). The authors conclude that SoundScan measures more consumers but not necessarily the most active record-buyers and that its store sample is one of "convenience rather than representativeness" (214).

To test the idea that diversity would drop as a result of the new technologies, McCourt and Rothenbuhler examined selected charts. An average of 7.42 new records appeared each week on the Hot 100 singles chart for the 200 weeks before SoundScan was implemented, while 6.60 new records appeared on average for the 158 weeks after SoundScan appeared. However, new records did tend to debut significantly higher after the advent of SoundScan. On the country albums chart, the comparable averages were 1.70 new albums per week before SoundScan, and 1.43 new albums per week after SoundScan. Again, more albums debuted in the top 15 after SoundScan was implemented. The same pattern held for the "Billboard 200:" an average of 9.07 new albums debuted each week before SoundScan; 8.48 new albums debuted post-SoundScan. However, Billboard columnist Geoff Mayfield ("Decade" 5), who looked at a lengthier time frame, pointed out that, from 1984 to 1990, an average of 677 albums appeared on the "Billboard 200" per year. That number rose to 780 albums per year for 1992 through 2000. He also noted that many debut acts have had No. 1 albums, including Billy Ray Cyrus, Snoop Doggy Dogg, the Spice Girls, Alanis Morissette, Christina Aguilera, and Nelly. [9]

Despite these problems in chart methodology, it is worth briefly reviewing the other literature that has examined chart performance of marginalized groups. In his 1988 book, Nelson George, a critic for Billboard, Village Voice, and other publications, discussed the implications for Black artists who achieve mainstream success. The book, The Death of Rhythm and Blues, related the history of R&B music and its performance on the pop chart in several ways, noting that "what developed was a deadly feast-or-famine syndrome" (157). The years from 1967 to 1973 were very good for Black cross-over acts, as "the average pop peak of each year's No. 1 soul hits was at least No. 9. In other words, a typical No. 1 hit on the soul chart could be expected to, at least, reach the top 10 on the pop chart" (157). From 1974 to 1976, the average peak for a No. 1 soul hit was No. 15 pop, and by 1978, it was No. 22 (157). The situation worsened in the early 1980s, and in 1983, only one No. 1 soul hit, Donna Summer's "She Works Hard for the Money," made the pop top 10 (182). [10]

Only eight studies have focused on chart performance of specific marginalized groups. [11] Serge Denisoff and John Bridges found that about 14.8% of all pop artists were Black, and Wells ("Women in Popular") found that Black females have historically been able to match the chart performance of their White counterparts, while Black males have lagged behind White males. Wells's method is especially important to this study, as he traced the history of Black male and female performers by examining the top 50 singles for the years 1955 through 1985 to determine which years were best and worst. The point system he used is discussed more fully in the next section of this paper.

A conference paper by Jim Sernoe ("Performance") examined the weekly Top 10 on the singles and albums charts, comparing the two years before and after the implementation of SoundScan and BDS. He concluded that music by Black performers did considerably better on the singles charts after the new technologies were adopted; on the albums charts, however, music by Black performers did only slightly better.

Continued from page 2.

Two articles by Hesbacher et al. and one by Wells examined the chart performance of women. Hesbacher, Nancy Clasby, H. Gerald Clasby, and Berger studied the No. 1 hits from 1940 through 1974, finding that female solo vocalists accounted for 12% of all No. 1 hits, while female vocal groups accounted for 5%, and mixed-gender groups/ duets accounted for 7%. Hesbacher and Bruce Anderson found that female singers did the best on the charts during the time from 1970 to 1978. Wells ("Women on the") again looked at the top 50 singles each year, this time covering from 1960 to 1986. The best years for women performers were 1962, 1963, 1964, 1967, and 1985; the worst years were 1969, 1972, and 1981.

A more recent study by Wells ("Nationality") looked at the album chart performances of women, Black musicians and foreign musicians in the late 1980s and the 1990s. He concluded that White American women performed better on the charts than Black American women, and, even in their best years, total scores for women were lower than total scores for men. African-American men fared considerably better than African-American women, although totals for all African-American artists continued to rise. Scores for foreign artists in general were low, with a few Canadian women, including Shania Twain, having successes.

Another study by Sernoe ("Here") examined country music's ability to cross over to the pop charts. Examining the year-end charts from 1955 to 1996, he documented country music's uneven track record of mainstream successes and failures. The best years for country were 1955 to 1960, 1974, 1975, and 1981. [12]

It is important to note that all of these studies--with the exceptions of McCourt and Rothenbuhler's examination of SoundScan, one study by Wells, and half of Sernoe's paper on Black performers--have considered only the singles charts. This study looks specifically at the pop albums chart, where the financial stakes are considerably higher.

Research Strategies

The year-end pop albums charts from Billboard magazine for the years 1981 to 2001 were studied. Billboard was chosen because it is generally regarded as the leader among trade journals catering to the popular music industry. Because Billboard changed its compilation methods in 1991, the years 1981 to 1990 and 1992 to 2001 provided 10 years for study on both sides of the conversion. Totals were computed for 1991 as well, but they were not used in any averages (see Tables 1 through 8). [13]

Each album that appeared on Billboards listing of the top 100 pop albums was given a point value based on its position in the year-end rankings. One hundred points were given for the No. 1 album of the year, 99 points were given for the No. 2 album, 98 points were given for the No. 3 album, down to 2 points for the No. 99 album, and 1 point for the No. 100 album of the year. This is the system Wells used in his studies, although he looked at the top 50 entries and thus counted down from 50 instead of 100 (see earlier section on previous studies). Sernoe used a similar system in his examination of country cross-over. As noted before, one criticism of using these charts is that they do not give distances between entries. Thus, the No. 1 album of the year receives 1 point more than the No. 2 album, regardless of the difference in final sales totals. This criticism is valid on one level, but, because this study is testing the idea of cross-over from other charts, the specifics are not as important. For example, the fact that albums by Reba McEntire or Enya sell enough to make an appearance on the pop chart (as opposed to being marginalized to the country or new age charts, respectively) is enough to obtain a measure of their popularity against other types of music. If a researcher were attempting to gauge sales strength of the various genres, the actual piece counts would be more important. However, this study looks at chart performance, not sales. [14]

Each non-mainstream album was initially classified into one category: pop, country, Latin, metal/hard rock, R&B, or rap/hip-hop. Unfortunately, an "other" category had to be created due to the fact that certain artists did not fall easily into one of the original categories (among others, Grover Washington Jr., Devo, the Royal Philharmonic Orchestra, Jane Fonda, Sade, Ziggy Marley, Weird Al Yankovic, Enya, the Three Tenors, Beavis and Butthead, the Simpsons, the Benedictine Monks of Santo Domingo De Silos, Kirk Franklin, and B.B. King). [15]

The determination of which genre a particular album represented was based on marketing, not musical, factors. While it could be argued that purists care more about musical styles, Billboard measures marketing/business success, regardless of the music itself. [16] So, for example, if country listeners began buying music by a rock group such as the J. Geils Band and the country radio stations monitored by Billboard played it, it would presumably appear on Billboard's country charts. [17] For purposes of this study, the album's genre was determined using the following criteria:

* If the album appeared on a non-mainstream chart and crossed over to the "Billboard 200" later, it was classified according to the chart on which it initially appeared (e.g. Kenny Rogers' Greatest Hits reached No. 1 on the country album chart in 1981 before crossing over to the mainstream chart). This type of cross-over was far more prevalent before the advent of SoundScan than after.

* Similarly, if the album appeared on a non-mainstream chart at the same time it appeared on the "Billboard 200," it was classified according to the chart on which it also appeared (e.g. Alicia Keys's Songs in A Minor appeared on the R&B albums chart and the "Billboard 200" simultaneously in 2001). This type of chart action was far more prevalent after the advent of SoundScan than before.

* If the singles from an album appeared on a non-mainstream chart and crossed over to the pop singles chart later, the album was classified according to the chart on which it initially appeared (e.g. singles from Whitney Houston's first album, in 1985-86, received airplay first on R&B stations and thus appeared first on Billboard's R&B singles charts; thus, the album Whitney Houston is classified as R&B; most of her later albums--with the exception of I'm Your Baby Tonight in late 1990--were serviced to R&B and pop radio at the same time, and thus are considered pop in this study).

* If an album appeared on the "Billboard 200" and did not appear on any specialized chart, for purposes of this study it is considered mainstream pop and not part of a specialized genre.

Total points for each genre were determined for each year, and a yearly average was determined for each genre before and after the advent of SoundScan (see Tables 1 and 2). Each genre was then examined more closely to determine how many albums crossed over and how high those albums ranked on the year-end top pop albums listing (see Tables 3 through 8).

In addition to a year-end ranking of pop albums, Billboard compiles a year-end ranking for several genres, including country, R&B, new age, jazz, and several others. These rankings for each year were also examined for cross-reference and are referred to occasionally in this paper.

Discussion

Results

General trends. Totals for country, Latin, and rap/hip-hop jumped dramatically over the 20-year period (see Tables 1 and 2). Of a possible 5,050 points in each year, country averaged 156.1 points per year before the implementation of SoundScan (1981-90). However, from 1992 to 2001, after the implementation of SoundScan, country's average was 599.6 points per year. Comparable averages for Latin music were 31.5 before SoundScan and 115.5 after; for rap/hip-hop, the averages were 167.7 before SoundScan and 650.5 after. Rhythm and blues dropped slightly, from 785.6 points per year before SoundScan to 736.1 after. Metal/hard rock took a slightly larger fall, from an average of 671.2 points before SoundScan to 594.8 points after. "Other" remained basically stable, with pre- and post-SoundScan averages of 238.3 and 232.6, respectively.

A more detailed examination of each genre follows.