Every team with a functioning ATS is sitting on the cheapest candidate pool it will ever have access to, and almost nobody works it systematically. A company making 50 hires a year with a typical funnel has processed thousands of applicants; a five-year-old agency CRM holds tens of thousands of records. These people were sourced or attracted at real cost, screened by your team, and in many cases wanted to work for you.
Rediscovery — searching and re-engaging your own database — is unglamorous and has no vendor incentive behind it, which is roughly why it stays undone. It is also the highest-ROI sourcing motion available to most teams, because the acquisition cost is already sunk.
The six segments, ranked by yield
| Segment | Who they are | Expected yield | Message angle |
|---|---|---|---|
| Silver medallists | Reached final stages, lost to another candidate | Highest | "You were close last time and we've been hiring since" |
| Declined offers | Got an offer, said no | High | Something has changed: comp band, role scope, remote policy |
| Withdrew for timing | Pulled out mid-process for personal or timing reasons | High | Timing changes; explicitly reference why they withdrew |
| Strong, wrong role | Good candidates screened out for fit, not quality | Medium-high | A role that matches what they actually do |
| Placed contractors (agencies) | People you've placed before | High | Contract ending soon? New assignment |
| Old applicants (18+ months) | Applied once, no strong signal either way | Low-medium | Treat as warm-cold; re-qualify before contacting |
Silver medallists deserve special attention because they are pre-validated by your own bar: a hiring manager already said yes-but-one-person-was-ahead-of-you. Two years on, they are more experienced, and the reason they lost may no longer apply. Practitioner reports on this segment are consistently the strongest in rediscovery, which is why some ATSs now tag it automatically.
Getting them out of the ATS
The obstacle is usually search, not data. Most ATS search is keyword-based over resume text, which is worse than the tools you use for external sourcing. Three approaches:
Structured queries on your own fields. Filter by stage reached, rejection reason, and date. This alone surfaces silver medallists if your team records rejection reasons consistently — and if it doesn't, fixing that is a two-week process change with a permanent payoff.
Semantic search over the resume corpus. Some ATSs and rediscovery tools rank your database against a role description rather than a keyword string. This is where AI search genuinely helps: your database is a fixed, bounded corpus, so higher recall has no downside risk of pulling in irrelevant markets.
Export and rank outside the ATS. Unglamorous but effective for one-off pushes: export the segment, rank it against the brief, work the top 50.
Whichever route, do the deduplication and re-verification step. Practitioner and vendor estimates put contact-data decay at roughly 20–30% a year, concentrated in work email addresses — so a three-year-old record needs its address re-checked before you send, or your bounce rate will damage deliverability for everything else you send that week. The protocol is in our contact-data guide.
The re-engagement message
Rediscovery outreach fails in one specific way: pretending it's a first contact. A message that reads like cold outreach to someone who interviewed with you twice is worse than cold outreach, because it proves nobody remembers them.
The structure that works acknowledges the history in the first line, states what has changed, and asks for something small:
"You interviewed with us for the senior backend role in early 2025 and it came down to you and one other person. We've just opened a similar role on the payments team — same scope, and the comp band has moved since then. Worth a conversation?"
Three rules. Be specific about the history — role, rough date, and outcome if it was positive. Name what changed; "we're hiring again" is not a reason for someone who was rejected to re-engage. Don't over-explain the rejection unless they ask; if the reason was quality rather than fit, they probably shouldn't be in this segment at all.
For declined offers, the change has to be real. Re-approaching someone with the same package they already turned down is a wasted contact and a small reputational cost.
What to expect
Directionally, from practitioner reports: rediscovery reply rates run well above cold outbound — often two to three times the cold rate for silver medallists and declined offers — because the relationship exists and the sender is recognised. Treat that as a range to test rather than a promise; yields depend heavily on how the relationship ended and how long ago.
Two structural advantages beyond reply rate. Time-to-hire is usually shorter, because these candidates have already been screened and often need fewer interview stages. And cost per hire approaches zero on the acquisition side — no licence, no credits, no job-board spend. When you compute cost per sourced hire (one of the five metrics in our metrics guide), separate rediscovered hires out or they will flatter your cold-sourcing numbers.
The realistic ceiling: rediscovery is a supplement, not a channel you can scale indefinitely. A database of 20,000 records contains a few hundred genuinely re-contactable people for a given role, and once you have worked them you have to wait for the pool to refresh. Teams that treat it as a monthly habit — not a one-off cleanup project — get the compounding version.
Privacy and consent
Re-contacting old applicants sits inside data-protection rules, and "we already had them in the system" is not automatically a lawful basis. Nothing here is legal advice, but four checks belong in the project:
Retention periods. Many privacy notices state how long applicant data is kept. If your retention policy says 12 months and you're mailing four-year-old records, you have a problem that predates the campaign.
Notice and purpose. Did your application notice contemplate re-contact about future roles? Under GDPR this is usually where the argument lives.
Deletion requests. Honour them across every system the record reached, including any sourcing tool that has ingested it.
Opt-outs. Give an easy way to stop hearing from you, and actually record it — re-contacting someone who opted out is the fastest way to turn a warm asset into a complaint.
Which tools help
Rediscovery is a feature of several categories rather than a product category of its own:
Modern ATSs — Ashby, Greenhouse — increasingly include database search, silver-medallist tagging, and talent pools. Check what your existing licence already covers before buying anything; this is the most common accidental duplicate purchase in the stack.
CRM-first tools — Gem, Beamery, Kula — treat your database as the primary asset and build re-engagement campaigns around it. For agencies, Bullhorn and Loxo already hold the records; the gap is usually the outreach discipline, which is why an engagement platform pays off there.
Agents — Noon, Fetcher — can include your existing database as one source alongside external search, screening old records against the current brief. Worth asking specifically: does the agent look at our ATS, and does it respect our stage and opt-out data?
For the wider field, see the 2026 ranking and the pricing benchmark. But start with the free version of this: one afternoon, one segment (silver medallists), one live role. It costs nothing and it is the fastest sourcing result available to most teams.