Impact of platform cues and algorithmic listings on B2B vs B2C gig success: analysis with random forest and panel regression

Purpose This study aims to investigate the importance of platform cues on project success using signaling theory (ST) in using B2B or B2C artificial intelligence/machine learning (AI/ML) service providers on gig platforms. In addition, it uses locus control theory (LCT) to argue that success may variably affect gig players’ revenues, influenced by algorithmic positioning. Design/methodology/approach The study identifies key value-based and performance-based signals impacting project success and revenue prediction using Panel Regression (PR) and Random Forest (RF) analysis on 10-month data from 28 gig agencies and professionals. Findings The result of PR found three important value drivers (automation, innovation and personalization) and four significant performance-based signals (number of jobs, number of hours, average review and job completion rate), which measure the B2B project success score in the gig platform. RF analysis shows that value-based signals, such as personalization and automation, have greater predictive power for gig success than performance-based signals, regardless of whether the gig is B2B or B2C. However, PR suggested that value-based signals are more important for B2B gig providers. Originality/value This study contributes to the online service industry by emphasizing the relative importance of platform signals (in time, quality and forecasting) through the lens of signaling theory. This study also suggests that the signal may not yield equitable outcomes for gig agencies. Instead, it depends on external factors, such as algorithmic listings, as explained by locus control theory. Hence, the dual application of ST and LCT indicates the transient nature of signals that vary over time, the nature of enlistments (B2B vs B2C) and uncontrollable external factors beyond the control of the seller and buyer in a B2B or B2C transaction.