Nonlinear Determinants of Time-to-Hire for Retail Sales Clerks: an Empirical Analysis of Applicant Tracking System Data

<jats:p>Background. Time-to-hire (TTH), defined as the number of calendar days from a recruiter’s first contact with a candidate to the candidate’s hiring, is a key efficiency metric in high-volume recruitment. In Russian retail, the factors affecting TTH have not previously been examined systematically using quantitative methods. Objective. To determine which recruitment-process factors affect TTH for sales assistants/cashiers and to what extent, and to quantify the observed nonlinear effect of the number of recruiter contacts with a candidate. Materials and methods. We used anonymized event logs from the corporate Applicant Tracking System (ATS) of a large retail chain for 2018–2020 (861,762 event records). The analysis included statistical hypothesis testing (chi-square tests, rank-based tests, analysis of variance, and logistic regression) and ordinary least squares regression modeling of log-transformed TTH with heteroskedasticity-robust standard errors. Results. Median TTH was 7 days. The regression model explained 45 % of the variance in log-transformed TTH. The duration of the paperwork stage was the strongest contributor to longer hiring times. A nonlinear relationship was identified and quantified: beyond 5–7 recruiter contacts, each additional contact was statistically associated with longer TTH. This association is likely due to reverse causality, as more challenging candidates require more contacts. Conclusion. Based on the estimated model coefficients, practical rules for reducing hiring time were formulated. An illustrative calculation of the potential economic effect of their implementation is provided.</jats:p>