A Behavioral Segmentation of the Arts Market

Donald E. Sexton, Columbia University
Kathryn Britney, Columbia University
ABSTRACT - This study describes a behaviorally based segmentation of the audiences of performing arts. A similar segment structure was found for samples consisting of respondents from both ballet and opera audiences. The segment differences are examined and marketing implications discussed.
[ to cite ]:
Donald E. Sexton and Kathryn Britney (1980) ,"A Behavioral Segmentation of the Arts Market", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 119-120.

Advances in Consumer Research Volume 7, 1980     Pages 119-120

A BEHAVIORAL SEGMENTATION OF THE ARTS MARKET

Donald E. Sexton, Columbia University

Kathryn Britney, Columbia University

ABSTRACT -

This study describes a behaviorally based segmentation of the audiences of performing arts. A similar segment structure was found for samples consisting of respondents from both ballet and opera audiences. The segment differences are examined and marketing implications discussed.

INTRODUCTION

While there have been several studies focussed on attendance at a specific cultural institution (DiMaggio, et. al., 1978) there has been less consideration given to the overall patterns of attendance at various cultural institutions. This study is an attempt to examine these attendance profiles to identify market segments as a basis for developing marketing strategies for cultural institutions.

DATA BASE

The data were collected by National Research Center of the Arts, Inc., in 1974, under a grant from the New York State Council on the Arts. Self-administered questionnaires were distributed at performing arts events including ballet, opera, symphony, theater, modern dance, jazz, and poetry and literary readings. Questionnaires were completed by over 30,000 audience members and included questions concerning demographic information, frequencies of attendance at various cultural events over the last twelve months, attitudes toward event attributes affecting their choices, and behavior with respect to marketing variables such as price and source of information.

Demographic characteristics of the total sample indicate a generally up-scale profile. Over 50 percent of the total audience sample had a college degree and about 50 percent had household incomes of $15,000 or more.

RESEARCH PLAN

The main thrust of the study was to determine if there was a parsimonious segmentation scheme that would identify audience members with differing interests in various cultural events and, if so, to describe the members of these segments with respect to demographic characteristics, attitudes, and behavior with respect to marketing variables such as sources of performance information. Initially, the reported attendances of respondents at several cultural events or institutions were utilized to segment the sample by attendance patterns. Them differences among the resulting segments were investigated.

TABLE 1

FACTOR LOADINGS FOR EVENT ATTENDANCE FREQUENCIES

METHODOLOGY

Twelve-month attendance frequencies reported by respondents for the 10 performing arts events listed in Table 1 were factor analyzed to group events with similar attendance patterns. The principle component factor analysis (with orthogonal rotation) suggested three factors which accounted for the total variance in the attendance frequency data. As shown in Table 1, the first factor could be interpreted as attendance at classical performing arts events. The second factor could be interpreted as attendance at museums, and factor three suggests an attendance pattern that focuses on dancing (modern and ballet).

The purpose of the factor analysis was to find a subset of relatively independent event variables to be used in the cluster analysis since the Howard Harris clustering algorithm assumes independence among the variables on which the clustering is based. Based on the factor analysis results, attendance frequencies at the following four events were selected to represent the ten original events: ballet, opera, art museum, and history museum.

Since some of the questions were situation specific (i.e. asked with respect to the specific performance being attended) the analysis was done using respondents who were attending one type of event (i.e., ballet performances) and replicated on a different group (i.e., opera respondents) to determine the robustness of the segmentation scheme. Due to software limitations, for both the ballet and opera samples, a subsample of approximately 1,000 respondents was randomly chosen for analysis.

SEGMENTATION RESULTS

Two to ten-group cluster solutions were derived. After the four group solution the smaller clusters splintered rather markedly while the largest segment remained unchanged. Based on the rules of thumb that the segments be interpretable and that subsequent splits not substantially reduce the size of the largest existing sample, the four-group solution was selected as the most informative partitioning.

TABLE 2

BALLET SAMPLE: MEANS OF ATTENDANCE FREQUENCIES IN PAST 12 MONTHS

TABLE 3

OPERA SAMPLE: MEANS OF ATTENDANCE FREQUENCIES IN PAST 12 MONTHS

As shown in Table 2, the ballet respondents appeared to fall into groups one might describe as: "The Lights" (segment 1)--relatively low attendance at all events studied. "The Museum Fans" (segment 2)--relatively heavy attendance at museums. "The All-Rounders" (segment 3)--relatively frequent attendance for most events. "The Dance Specialists" (segment 4)--relatively frequent attendance at dance performances as compared to other events.

For the opera respondents (Table 3), a similar pattern emerged--Lights, Museum Fans, All-Rounders, and Opera Specialists. (Preliminary results for symphony respondents also appear in harmony with this four-group segmentation scheme.)

DIFFERENCES AMONG SEGMENTS

As shown in Tables 4 and 5, household income was not individually useful as a variable to identify segment membership. For the ballet sample, the key variables (as indicated by an analysis of variance) were whether or not the respondent lived in the city, whether or not he or she was married and had a child below 16 years of age, and the respondent's age and sex. In particular, as a group the Lights included more people living outside the city and living with children below 16 than any other group (a not surprising result). However, the high proportion of women and the lower average age among the Ballet Specialists are notable.

Among opera respondents, the Lights show similar average characteristics to the Lights in the ballet sample--a low proportion living in the city and a higher proportion with children younger than 16. However, the Opera Specialists seem distinct from the Dance Specialists. They include a majority of males and the average age for this segment is about the same as those of the other segments.

TABLE 4

BALLET SAMPLE: MEANS OF DEMOGRAPHIC CHARACTERISTICS

TABLE 5

OPERA SAMPLE: MEANS OF DEMOGRAPHIC CHARACTERISTICS

Similar comparisons among the segments were made for other variables available. Results were generally similar for both the ballet and opera samples. For example, the Specialists on the average spent the most for the entire evening while the tickets of the Lights were on the average the most expensive. The Lights were less likely to have heard of the performance through media including newspapers, critics' reviews or articles, and mailings, in part because a relatively higher proportion of the Lights received their tickets as gifts.

Ballet and Opera Specialists differed in that the majority of Opera Specialists purchased their tickets through mail while nearly 50 percent of the Ballet Specialists bought their tickets at the box office. However, both types of Specialists were, on the average, more inclined to weight critics' reviews more heavily in their attendance decisions than the members of the other segments.

CONCLUSIONS AND IMPLICATIONS

For two independent samples--members of ballet and opera audiences--a similar four-segment partitioning emerged, suggesting four different profiles of attendance at cultural events and institutions: Lights, Museum Fans, All-Rounders, and Specialists. For both samples, three demographic variables (live in city, sex, and married with children younger than 16) were useful to distinguish the segments. Members of the segments displayed different behavior with respect to marketing mix variables and with respect to their decision process for selecting events to attend.

These results can be utilized to design and direct communications tools for performing arts institutions. For example, mailings to the All-Rounders might include offers of joint subscriptions to several different types of performing arts events, while those to the Specialists might include offers for various companies working in the same art form.

Future research will consist of enlarging the sample on which these results were based and on formulating a more specific model of attendance at cultural events.

REFERENCE

DiMaggio, Paul, Useem, Michael, and Brown, Paula (1978), Audience Studies of the Performing Arts and Museums, Washington, D.C.: National Endowment for the Arts.

----------------------------------------