Sourcing Lab- Hands-On Strategies for Smarter Talent Discovery

Sourcing has changed shape over three decades, from dial-up bulletin boards to AI-driven search, but the underlying skill has stayed the same: knowing how to ask the right question of a machine. What follows is a set of concrete tactics for finding candidates who aren't sitting in the obvious places, and for stopping the search once enough good leads exist.

• Start with the end in mind. Reverse-engineer the process by working backward from how many people need to interview, how many get submitted, how many conversations that requires, and what response rate typical outreach gets. Across three decades of building sourcing teams, the average number of passive leads needed to make one hire lands around 147. Once that number of leads is reached, the search stops; if it isn't enough, another batch gets pulled rather than searching endlessly.
• Surface hidden talent by ignoring job titles. Titles are inconsistent and personal — six recruiters in the same room can hold six different titles for the same work. Searching for people by group membership, alumni association, or professional society (rather than title) turns up candidates who never show up under an obvious keyword search.
• Look before and after the target title. The role someone held right before their current one, and the role they moved into after leaving a similar job, both point to strong candidate pools. Company size matters too: moves from small companies to large ones, or large to small, tend to be easier to make than lateral moves between similarly sized employers.
• Target the "almost right" candidate, not the perfect match. Someone who has already done the exact job for years is often burned out on it and ready to move to something adjacent. Someone one step away from the requirement is frequently more motivated and more available.
• Stack filters deliberately. Every filter applied is a decision, so filters on skill recency, education ceiling, geography, and function should each be set with intent rather than piled on loosely. Advanced tools that allow time-boxed skills or capped education levels make this kind of precision possible.
• Break Boolean searches into small chunks rather than building long nested strings, running several narrow searches and merging the results into one candidate list.
• Go where the tribe gathers — conferences, discussion groups, meetups — and participate before reaching out, since candidates in those spaces are far less likely to be buried in generic automated outreach.