Abstract

We use the concept of means–ends decoupling to examine why companies continue to be major contributors to environmental and social problems despite committing increasingly to corporate social responsibility (CSR). Specifically, we ask: How do companies restrain (versus fail to restrain) means–ends decoupling? We answer this question through a comparative case study of four multinational companies with different levels of means–ends decoupling. Based on interviews and secondary data, we inductively identify two distinct approaches to CSR implementation: experimental vs. consistency-oriented CSR implementation. Experimental CSR implementation means that companies (1) produce CSR knowledge about what is happening in specific CSR contexts and use this knowledge to (2) adapt CSR practices to local circumstances – an interplay that restrains means–ends decoupling. Consistency-oriented CSR implementation lacks this interplay between knowledge production and practice adaptation, which fosters means–ends decoupling. Our model of experimental versus consistency-oriented CSR implementation advances two streams of research. First, we advance research on means–ends decoupling by highlighting the importance of experimentation for restraining means–ends decoupling. Second, we advance research on the impact of CSR activities by questioning the widespread assumption that consistency should be at the heart of CSR implementation.