Professional network

Matching quality is the entire product, and volume actively works against it. A platform judged on applications sent optimises for the thing that makes it worse, because a recruiter receiving a hundred loosely relevant applications reads ten and rejects the rest unseen. The metric worth optimising is not applications but outcomes — interviews, offers, and how quickly a role closes — which is harder to measure, slower to move, and the only version that keeps employers paying.

The candidate side is where trust is lost, almost always in the same way: applying into silence. A rejection that arrives is vastly better than nothing arriving, and a status that updates honestly costs almost nothing to build. This is not a courtesy feature — it is the difference between a platform candidates return to and one they use once and resent, and candidate supply is what employers are actually paying for.

Automated screening is where this product category carries genuine legal and ethical risk. A model trained on who was hired previously learns the preferences embedded in that history, including the ones nobody would defend out loud, and in several jurisdictions automated decisions about employment carry disclosure and explanation requirements. Ranking that a human reviews is a defensible design; automatic rejection on a score is a liability wearing the costume of efficiency.