Predicting Loss Risks for B2B Tendering Processes Eelaaf Zahid IBM Research - Almaden San Jose, CA, USA eelaaf.zahid@ibm.com Yuya Jeremy Ong IBM Research - Almaden San Jose, CA, USA yuyajong@ibm.com Aly Megahed* IBM Research - Almaden San Jose, CA, USA aly300@gmail.com Taiga Nakamura IBM Research - Almaden San Jose, CA, USA taiga@us.ibm.com Abstract—Sellers and executives who maintain a bidding pipeline of sales engagements with multiple clients for many opportunities significantly benefit from data-driven insight into the health of each of their bids. There are many predictive models that offer likelihood insights and win prediction modeling for these opportunities. Currently, these win prediction models are in the form of binary classification and only make a prediction for the likelihood of a win or loss. The binary formulation is unable to offer any insight as to why a particular deal might be predicted as a loss. This paper offers a multi-class classification model to predict win probability, with the three loss classes offering specific reasons as to why a loss is predicted, including no bid, customer did not pursue, and lost to competition. These classes offer an indicator of how that opportunity might be handled given the nature of the prediction. Besides offering baseline results on the multi-class classification, this paper also offers results on the model after class imbalance handling, with the results achieving a high accuracy of 85% and an average AUC score of 0.94. I. I NTRODUCTION For modern enterprises and organizations to be successful in driving their business objectives, many of their core in- formation technology (IT) operations and infrastructures are often outsourced to external service providers to realize the benefits of the innovations and support systems they provide. Here, multiple service providers must compete against other providers to bid over large-scale IT service contracts through a competitive tendering process, which often comprise of com- plex requirements which can potentially yield high revenue if the contract has been signed [6]. Many of these service providers often bid on multiple tendering process opportunities simultaneously, thus creating and maintaining a pipeline of sales engagements with multiple clients. This pipeline comprises of several different stages, known as sales stages, where each bidding opportunity ad- vances in different phases of the tendering process, as shown in Figure 1. Each of these different opportunities advance through different sales stages depending on the various milestones and objectives cleared at each phase of the tendering process. Fig. 1. Contract Tendering Process Pipeline *Author performed work during employment at IBM Research. In the typical life cycle of a deal tendering process, an opportunity may fall in one of at least five different sales stages which are: 1) opportunity identification, 2) deal validation and qualification, 3) deal pursuit, 4) contract finalization, and 5) closing tendering process. Prior to the initialization of a tendering process, a prospec- tive client typically distributes a Request for Proposals (RFP) to any service providers who will be taking part of the bidding process. An RFP document is a formalized documentation which is used by the client to assess the background of the service providers, their capabilities, and the potential risks involved. A service provider will then provide a response document along with other assets to present the solution they will be implementing. In the first stage of the tendering process, the service providers seek any RFPs that are distributed by the client. In the second phase of the tendering process, the service provider analyzes the RFP to determine whether or not the corresponding deal is a suitable match against their own ser- vice offering portfolios and assess the level of trustworthiness of that client. It is also at this step that a solutioning team is assembled to piece together the preliminary solution, which is then subsequently submitted to the client. Towards the end of this phase, the client begins to filter down the number of potential service providers based on the solutions offered by each of the providers. Next, within the third phase, the provider assembles a sales team to work extensively with the client to define the scope of the solution and formulate the contract and define the contract terms of the deal. During the fourth phase of the bidding process, the client decides which service provider to work with, if any at all. If the client decides to choose a particular service provider, then the contract terms are finalized and agreed upon and the provider and client begin their work engagement to develop the solution. In the last phase, the project is handed off to the team designated to manage the project. For sellers and executives who manage the success of the strategy and execution of these tendering processes, having a data-driven insight of the health of these individual opportuni- ties can provide better visibility on the status of their pipeline. Furthermore, by leveraging the insight gained from knowing the likelihood that a contract will be successfully signed [2], sellers and executives can craft strategies through actionable recommendations provided by these predictive models to bet- ter navigate and optimize time, resources, and labor allocated 1 arXiv:2109.06815v1 [cs.LG] 14 Sep 2021