Different strategies for different fields? Exploration, exploitation, ambidexterity, and the performance of self-employed musicians

ABSTRACT

We contribute to the advancement of a contingent view of ambidexterity by examining exploration and exploitation across product and market domains in a sample of Canadian self-employed musicians. We find that (1) groups of musicians are more likely to benefit from ambidexterity than individual self-employed musicians, (2) group performance is more positively related to cross-functional rather than within-functional ambidexterity, and (3) a single strategic emphasis based on either product exploitation or market exploration is more positively related to the individual performance of classical musicians, that is, the performance effects of single strategic emphases depend upon the institutional field at the individual level.

Notes

¹ A few other studies have investigated how individual managers balance behaviors of exploitation and exploration, without systematically observing individual differences or performance (e.g., Mom et al. 2007, 2009; Papachroni et al. 2016).

² We thank an anonymous reviewer for emphasizing this point.

³ We thank one anonymous reviewer and the Associate Editor Norris Krueger for pointing out this limitation.

⁴ As pointed out by one anonymous reviewer, one could wonder about the reliability of our group measures on the basis that in many cases musicians control their individual careers but do not have direct control over their groups’ exploration and exploitation strategies. However, in our research design one does not need to have control over such strategies so long as such strategies are accurately reported. Thus, the reliability of our group measures depend on the accuracy of the respondents’ perceptions of their groups’ strategies. Several studies suggest that the shared cognition of a group is much more complex than the sum of its members’ cognitions (e.g., Bagozzi 2005; Brännback et al. 2018) and, although there is a lack of consensus around the measurement of cognitive constructs at the collective level, several studies suggest that average scores or the median response among team members provide more accurate measures of the group constructs (Brännback et al. 2018; Shepherd and Krueger 2002). These suggestions imply that individual perceptions might be biased towards the individual. However, since several of our hypotheses assume a difference of ambidexterity effects at the individual and the group level, gathering data from a single respondent actually runs against our hypotheses and provides a more robust test for them.

⁵ Detailed results of the factor analyses conducted on these scales are available from the authors.

⁶ Available from the authors upon request.

⁷ We thank one anonymous reviewer for raising this issue.