Prescribing Response Strategies to Manage Customer Opinions: A Stochastic Differential Equation Approach

• 成果类型: Article
• 署名作者: Yang, Mingwen; Zheng, Zhiqiang (Eric); Mookerjee, Vijay
• 署名单位: University of Washington; University of Washington Seattle; University of Texas System; University of Texas Dallas
• 刊物名称: INFORMATION SYSTEMS RESEARCH
• ISSN/ISSBN: 1047-7047
• DOI: 10.1287/isre.2018.0805
• 发表日期: 2019
• 页码: 351-374
• 关键词: word-of-mouth; reviews; reputation; sales; BEHAVIOR; IMPACT

摘要

Today, the reputation of a firm is profoundly influenced by user opinions expressed in online consumer reviews. Managing these opinions is, therefore, critical for the success of firms. We study the problem of devising an appropriate opinion management strategy (or response strategy) for a firm to respond to online customer reviews. To unravel the underlying mechanics of the problem, we develop a stochastic differential equation model that describes the evolution of review ratings over time for a given response strategy employed by the firm. This model is validated using data on online customer reviews and firm responses from two of the world's largest online travel agents. When pitted against popular benchmark models, such as autoregressive moving average, generalized autoregressive conditional heteroscedasticity, moving average, exponential smoothing, and naive method, our approach not only achieves comparable (often better) predictive performance, it is also able to incorporate the response strategy into the data-generation process underlying the review ratings. Our approach, therefore, is not just predictive, but, more importantly, one that can be used in a prescriptive sense, namely to prescribe a response strategy that controls review ratings in a desired manner. We operationalize the theoretical response strategy in our stochastic model to an operational prescription that a firm can implement and show the applicability of our approach for different business objectives, such as mean control, mean-variance control, and service-level control. Finally, we demonstrate the flexibility of the stochastic differential equation model by extending it to encompass multiple state variables.