A Knowledge Graph-Based Target Recommendation Approach for Mergers and Acquisitions

Abstract

Selecting the right merger and acquisition (M&A) target is a critical yet challenging endeavor, as the success of these strategic initiatives depends mainly on identifying compatible firms. This study draws upon the theoretical perspectives of strategic fit and organizational search to propose and validate a novel, two-stage M&A target recommendation approach designed as a managerial decision support system. Initially, the method facilitates a focused search in which acquirers define explicit criteria for identifying a highly relevant initial target within a knowledge graph (KG). It then employs a similarity-based search expansion using advanced KG embedding models to recommend additional targets that exhibit latent structural similarities. The efficacy of this approach is validated on a large-scale U.S. M&A dataset (2010–2022). Our key findings are threefold. First, our model demonstrates statistically significant superiority over benchmarks, confirmed through robustness checks, including tenfold cross-validation and temporal validation. Second, in an experiment on deals by experienced acquirers, our model is more effective at identifying these targets, quantitatively demonstrating its superior recommendation quality. Third, our analysis uncovers counterintuitive patterns, revealing that machine-identified structural similarities can be more potent predictors of fit than traditional human-centric filters, such as geography. It further explores the tool’s boundary conditions, showing that it is more effective in complex, high-tech sectors. This KG-based methodology offers a more informed, strategically refined, and empirically validated tool to enhance the quality of M&A decisions.