Sourcing Yield

Pick one downstream stage, then divide the candidates from a channel who reached it by all candidates that channel produced in the same period. Do this per stage rather than only at hire, because channels fail at different points. A job board might deliver plenty of applicants who clear screening but stall at technical assessment. Outbound sourcing might convert superbly once someone replies while producing very few replies at all. The stage-level view tells you which. Yield to hire is the headline, but yield to first interview is the number that lets you act inside a live search rather than after it. A candidate sourcing platform that tags origin at the point of entry keeps this honest, because retrofitted tags are guesses.

Because the channel a candidate arrived through is often not the channel that convinced them. Someone sees a conference talk, follows the company for months, reads a post from an employee, then applies directly when a role opens. Last-touch tagging records that as direct traffic and the referral-shaped influence disappears. The reverse also happens: an agency submits someone already sitting in your database from a prior search. Two habits help. Ask candidates in the application where they first heard about the company, and keep that answer separate from the system-recorded source. Then treat source of influence and source of hire as two columns rather than arguing about which single label is correct. Both are useful; conflating them produces confident, wrong budget decisions.

Slow channels look worse than they are when you measure too early. Referrals, silver medalists and rediscovered candidates from your own database often convert on a timeline measured in months, because the person was not looking when you first spoke. Compare that to a paid campaign that produces applicants within days and the fast channel wins every dashboard, even where the patient channel produces the better hires. Fix it by matching cohorts on entry date and letting each cohort mature the same length of time before comparison. Report the maturity window openly, because a chart labelled only 'last quarter' silently penalises everything slow. Where a channel's yield is genuinely strong but slow, the decision is to start it earlier, not to cut it.

Reallocation, and it should be specific. A channel with weak yield to screen usually has a targeting problem: wrong boards, wrong search string, wrong geography, a posting written for the internal audience rather than the candidate. A channel with strong screen yield but weak interview yield is producing plausible people who do not survive contact with the hiring manager, which is a calibration issue as much as a sourcing one. Cutting a channel outright is the last move, not the first, because volume you remove today does not come back quickly. Pull the stage-level numbers from recruitment analytics rather than a spreadsheet rebuilt each quarter, and keep definitions stable so this quarter is comparable to the last.