Can artificial neural networks predict lawyers’ performance rankings?

Purpose: The purpose of this paper is to propose a predictive model that could replace lawyers’ annual performance rankings and inform talent management (TM) in law firms. Design/methodology/approach: Eight years of performance rankings of a sample of 140 lawyers from one law firm are used. Artificial neural networks (ANNs) are used to model and simulate performance rankings over time. Multivariate regression analysis is used to compare with the non-linear networks. Findings: With a lag of one year, performance ranking changes are predicted by the networks with an accuracy of 71 percent, over performing regression analysis by 15 percent. With a lag of two years, accuracy is reduced by 4 percent. Research limitations/implications: This study contributes to the literature of TM in law firms and to predictive research. Generalizability would require replication with broader samples. Practical implications: Neural networks enable extended intervals for performance rankings. Reducing the time and effort spent benefits partners and lawyers alike, who can instead devote time to in-depth feedback. Strategic planning, early identification of the most talented and avenues for tailored careers become open. Originality/value: This study pioneers the use of ANNs in law firm TM. The method surpasses traditional static study of performance through its use of non-linear simulation and prediction modeling.