Multitask Agency, Modular Architecture, and Task Disaggregation in SaaS

  • Authors:

  • Anjana Susarla;Anitesh Barua;Andrew Whinston

  • Affiliations:

  • Department of Information Systems and Operations Management at the Foster School of Business, University of Washington, Seattle;Department of Information Systems, Risk and Operations Management at the McCombs School of Business, University of Texas at Austin;McCombs School of Business, University of Texas at Austin

  • Venue:
  • Journal of Management Information Systems
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

We examine contract choices in the provision of "software-as-a-service" (SaaS), which is a business innovation that transforms information technology (IT) resources into a continuously provided service. We draw upon agency theory and modularity theory to propose that one of the central challenges in service disaggregation is that of knowledge interdependencies across client and provider organizations. The resulting lack of verifiability of certain tasks results in a multitask agency problem. Our key research questions involve (1) the suitability of high-versus low-powered incentives in SaaS contracts when the outsourced tasks involve business analytics that are difficult to verify, and (2) how such contract choices are affected by the modularity of interfaces between the client and the provider. Analysis of data collected from 154 providers of SaaS offering a range of IT services supports our contention that when contracting for business analytics characterized by knowledge interdependencies across clients and providers, incentives should be "low powered." Modularity in the interfaces of the service provider increases the desirability of high-powered incentives in such situations. Our results are robust after accounting for endogeneity issues arising from unobserved matching between service providers and the nature of IT services outsourced by clients. With the increasing importance of information systems in services, this paper suggests that arm's-length relationships and high-powered incentives may be ineffective in incentivizing providers to perform on complex business analytic tasks, unless accompanied by the modularization of interfaces.