A significant strand of research on the job-person matching process has focused on the role of social networks in the hiring process. Past empirical studies of the comparative effectiveness of recruiting sources for getting jobs have started with a group of hires. But starting with hires is likely to compress differences between network recruits and other employees. In contrast to past studies, we treat hiring as a process rather than an event. We look at a pool of applicants to four job categories in a large retail bank who first inquired about jobs between 1993 and 1995 to identify the effects of referral networks at each stage of the hiring process. We trace this cohort from the point of initial contract through the interview, offer, and hire stages. Results of probit models indicate that network referrals are advantaged at both the interview and offer stages compared to external non-referral applicants after controlling for other factors. Consistent with theoretical arguments that referrals are pre-screened by current employees, our results show that referrals present more appropriate resumes than non-referral applicants. Also, referrals candidates are more likely than non-referrals to apply when market conditions are more favorable. However, resume quality and application timing cannot explain referrals’ advantages at the interview and hire phases.