cc_byKaya, Muhammed-FatihSchoop, Mareike2025-08-052025-08-052025https://doi.org/10.1007/s10726-025-09920-5https://hohpublica.uni-hohenheim.de/handle/123456789/18012The exchange of information is an essential means for being able to conduct negotiations and to derive situational decisions. In electronic negotiations, information is transferred in the form of requests, offers, questions and clarifications consisting of communication and decisions. Taken together, such information makes or breaks the negotiation. Whilst information analysis has traditionally been conducted through human coding, machine learning techniques now enable automated analyses. One of the grand challenges of electronic negotiation research is the generation of predictions as to whether ongoing negotiations will success or fail at the end of the negotiation process by considering the previous negotiation course. With this goal in mind, the present research paper investigates the impact of information load on predicting success and failure in electronic negotiations and how predictive machine learning models react to the successive increase of negotiation data. Information in different data combinations is used for the evaluation of various classification techniques to simulate the progress in negotiation processes and to investigate the impact of increasing information load hidden in the utility and communication data. It will be shown that the more information the merrier the result does not always hold. Instead, data-driven ML model recommendations are presented as to when and based on which data density certain models should or should not be used for the prediction of success and failure of electronic negotiations.engMachine learningInformation growthNegotiation outcomeClassificationPrediction performanceModel selection650Article2025-07-25