| Application Service Provider (ASP) is a recently emerged software delivery model under which an ASP hosts, manages and delivers software as a service to customers via the Internet or a private network. The ASP model offers benefits from cost savings, specialized expertise, a faster time to market, and a reduced risk due to a lower capital investment. ASP also allows for connectivity within a company and across trading partners and provides financial and technological flexibility. However, customers who are unsure about the value of ASP services and their demands may be reluctant to commit to ASP contracts. Many are also concerned with security, performance and loss of control.; To analyze the ASP model, we build cost and benefit models for ASP and traditional software licensing structures for both software owners and customers. We then propose three real options based approaches to address uncertainty about ASP value, demand, and concerns about security and performance. First, we propose a usage-based pricing structure with an option to switch to a subscription fee. This arrangement allows ASPs to penetrate the low-usage market and to link revenue to the value customers receive, while providing an upper bound on customers' costs. Second, we evaluate an option to bring the software in-house after an initial period of ASP-based access. This option allows customers to manage their uncertainty about the software's value and their demand for it, and to hedge against unacceptable security and performance risks. Lastly, we analyze an option to end an ASP contract prior to its expiration, under a minimum usage requirement and/or a minimum time commitment.; To study these real options, we identify the three key underlying uncertainties to be software value per usage, usage level and the number of users. We propose using interrelated mean-reverting processes with time-varying means and time-varying variances to model these uncertainties, which allow us to capture software cycles, learning effects, company growth and the correlation between software benefit and usage. We utilize a simulation-dynamic programming approach to determine the optimal exercise policies for and values of these real options. Finally, we investigate cost, revenue and risk implications to the seller from providing these options. |
