We use online job board data to document novel facts regarding unemployment and on-the-job searches. We define a relevant set of job ads using the bipartite network between ads and applicants and show how applications are affected by demographics such as gender, age, and marital status as well as timing variables like unemployment duration, job tenure, and cyclical conditions. We also study the selective margin, i.e., how the posted-expected wage gap and the applicant fit into ad requirements such as education, experience, location, and occupation affect applications. This evidence can help discipline current and future search theories.
JEL Codes: E24, J40, J64