I’d look at what share of those extra applicants were actually qualified. If many were strong fits, the cost per good applicant seems pretty reasonable and worth testing at a slightly bigger scale.