Recognizing and Selling Good Ideas: Network Articulation and the Making of an Offshore Innovation Hub

Abstract

Organizations have been offshoring work to companies in other countries for a long time. But in the past decade, the practice of offshoring has been changing in two important ways. First, an increasing diversity of work is becoming digitized such that it is becoming relatively easy to send knowledge-intensive work abroad. Second, companies fearing the loss of intellectual property and uneven talent are establishing captive offshore centers that often place offshore workers at structural holes in a company’s global network. Given that much organizational research has linked digitization and network structure to innovation, we conducted a longitudinal, multi-method study of offshoring at a large automotive engineering firm to explore whether and how the changing nature of offshore work might lead offshore employees to innovate, rather than do low-value work. Our findings show that some offshore workers can take advantage of their position at structural holes in social networks by creating new work practices that allow them to recognize good ideas coming from across the company, whereas others can take advantage of their position in cohesive networks to sell those good ideas to upper management. The findings also show the creation of a certain set of practices that allow offshore workers to articulate their different networks (characterized by structural holes and cohesion) by sharing and seeking ideas from one another in ways that turn an offshore center into an innovation hub. In particular, we found that the workers at the center under study, who received low-status and low-value work, were able to outpace every other engineering center in the company in institutionalizing good ideas for process improvement. We discuss the implications of these findings for our understanding of the changing nature of offshore work and for theories of social networks and innovation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 32 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.