Predicting negotiation behavior to support decision making in civil dispute resolution

Despite the advocacy of leveraging data analytics to improve operational efficiency, there is a paucity of research on how analytical technologies afford professional service innovation and enhancement. We propose a data-driven decision support framework for civil litigation negotiation, which is a routine business activity in legal service firms. It is typically conducted in a traditional manner with the conflicting parties drawing on their past experiences and prior knowledge to guide decision-making. This model predicts human negotiation behavior based on historical records and incorporates the behavioral insights into the decision-making process. We introduce a sequential directed acyclic graph to characterize the causal relationships between offers and employ different approaches to predicting the opponent’s next moves. By integrating utility analysis, each player can decide whether to accept the opponent’s offer or counter back. The proposed framework is illustrated through a field experiment based on the UK MoJ Portal for handling low-cost injury claims and 88 cases with complete negotiation history. We find that better outcomes for both parties can be delivered by implementing the proposed model. The analysis result also represents convincing evidence that low-cost cases should ideally be settled out of the court via negotiation to maximize shared benefits. This paradigm could be easily generalized to other types of civil dispute resolutions negotiation to enhance both operational efficiency and service quality.