Employees, Contractors, or Hybrid: An Operational Perspective

We consider a platform’s problem of how to staff its operations given the possibilities of hiring employees and setting up a contractor marketplace. We aim to understand the operational difference between using contractors and using employees, as well as the potential for using a hybrid staffing solution. We consider an environment where demand is not only stochastic, but also evolving over time, which we capture via a state that determines the demand distribution. The platform controls the number of employee hours it uses for serving demand, as well as the wage paid to contractors per utilized hour. Employee numbers are rigid, in that they cannot be adapted to the state of the world. We consider two model variations which differ in whether contractor wages can adjust to the state. If the contractor wage can be dynamically adjusted, we prove that the problem is reducible to a two-stage newsvendor model with recourse, and the optimal solution has an intuitive interval structure. If the contractor wage must be set up-front, the optimal solution becomes complex and no longer satisfies an interval solution. We show that contractors perform better than employees in two dimensions. First, we can bound the worst-case profit from using contractors, while no similar bound holds for employees. Second, contractors provide great flexibility if, for example, the set of demand distributions are scaled versions of each other. This allows a fast-growing platform to optimally set its contractor wage even if it does not know the probabilities of different growth scenarios. Joint work with Sebastien Martin (Kellogg) and Haotian Song (NYU Stern).