The single‐sourcing versus multisourcing decision in information technology outsourcing

Abstract

Abstract Although using a single vendor when outsourcing information technology (IT) services (i.e., single‐sourcing) is commonplace, splitting the IT service package among multiple vendors (i.e., multisourcing) is an increasingly popular strategy. However, our understanding of the factors that contribute to this trend is limited. Guided by a thorough review of the extant literature and interviews with industry experts, we focus on the influence of emerging technologies (i.e., cloud‐based services and automation services), vendor location (i.e., offshore or domestic), and experiential learning (i.e., client and vendor experience) on the choice between single‐sourcing and multisourcing for IT services. The trend towards more multisourcing exemplifies a broader movement towards greater disaggregation and distribution of knowledge work. Thus, we utilize a knowledge‐based perspective to develop our hypotheses for how the aforementioned factors relate to the likelihood of multisourcing. Analyzing a large database of IT service contracts, we find that cloud‐based services are less likely to be multisourced, but automation services are not significantly associated with the likelihood of multisourcing. We also find that more experienced client firms are more likely to adopt multisourcing. Conversely, our results indicate that more experienced vendors and vendors located offshore are less likely to be part of multisourcing arrangements. Exploratory post hoc analyses yield additional interesting insights related to client and vendor location. These findings advance our understanding of the drivers of the single‐sourcing versus multisourcing decision in IT and more broadly contribute to the literature on technology sourcing. Our study has normative implications for both client and vendor management teams.

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 — 30 citations exist upstream.

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