What is source of hire and how do I track it accurately?

The channel that produced a person who was actually hired, as distinct from the channel that produced the most applications. Those two answers are frequently different, and the gap is the point of the metric. A large aggregator may supply most of your applications and almost none of your hires, while referrals or your own careers site produce a small number of applications that convert repeatedly. Tracked properly over several quarters, source of hire tells you where to put budget, where to put recruiter hours, and which relationships to invest in. Tracked loosely, it produces a chart everyone quotes and nobody trusts, usually because two people are counting it differently.

Because candidate journeys are not linear. Someone sees a post on a social platform, checks the company on an aggregator, hears from a friend who works there, then applies through your careers page a month later. Every one of those channels can claim the hire. Self-reported answers to how did you hear about us are approximate at best, since people forget. Agency and referral claims add a commercial dimension, because money often depends on the answer. The fix is not perfect attribution, which does not exist. It is one written rule applied consistently, so the trend is comparable over time even if any single record is arguable. Perfect attribution is not available here, and pretending otherwise is what makes people distrust the whole report.

Decide first-touch or last-touch and write it down. First-touch credits whatever originally brought the person into your orbit and rewards awareness activity. Last-touch credits the channel through which the application arrived and is easier to capture automatically. Neither is correct in the abstract, so choose one and keep it. Then capture the source at the point of application through tagged links and defined job board integrations rather than asking later. Set a rule for referrals and agency introductions, including how long a claim lasts, and write it into the agency terms. Configure the source list in your hiring software as a controlled set rather than free text, because free text becomes unanalysable within a quarter.

Compare cost and conversion per source, not volume, using recruitment analytics rather than a manual export. For each channel, look at applications, interviews, hires and the money and recruiter hours it consumed. A board that supplies plenty of applications at low cost per application can still be the most expensive per hire, and a channel with few applications and several hires deserves more investment. Segment by role family, since a technical community that works for engineering hiring is irrelevant for sales. Then review quarterly and actually reallocate. The most common failure is not the measurement, it is collecting the data for a year and never moving a budget line in response to it. Put the reallocation decision into the same quarterly meeting that reviews the numbers, with a named owner.