Should boundary-spanners be ambidextrous?

Building on Thompson’s (1967) distinction between technical core and boundary-spanning units, we argue that boundary-spanning units: (a) enhance their chances of survival by developing ambidexterity, and (b) achieve ambidexterity through the development of a supportive relational context (i.e. the set of social relationships with internal and external parties in which they are embedded). We test and find support for these propositions using primary data collected from 95 corporate venture units over a three-year period. Our study contributes to ambidexterity theory by suggesting that the antecedents and consequences of ambidexterity at the intra-organizational level differ according to the type of unit.