Your talent pool is leaking — quietly, constantly, and at a rate that most agencies never measure. A candidate you spoke to eighteen months ago has probably changed jobs. Their email may have changed with it. Their career ambitions almost certainly have. Their GDPR consent may have lapsed. The record in your database, meanwhile, looks identical to the day it was created: same title, same employer, same tags. It feels like an asset. It's becoming a liability.
The industry average data decay rate for professional contact databases sits at roughly 25–30% per year according to SHRM's talent research. That means if your candidate database has 2,000 records and you haven't done a systematic refresh, around 500–600 of them are functionally wrong right now. That's not a small problem you can sort later. It's the foundation of everything — every shortlist, every search, every reactivation campaign.
What is a talent pool?
A talent pool is a curated database of pre-qualified candidates — people who have been assessed, categorised by skills and seniority, and kept warm for future roles — enabling recruiters to fill positions faster by searching internally before sourcing from scratch. Unlike a raw applicant database, a talent pool implies an ongoing relationship: the candidates have expressed some willingness to be contacted, and the recruiter has enough context to know whether a new role is worth surfacing to them.
The distinction from a general candidate database matters. A talent pool is selective by design. A database is a collection of records; a pool is a collection of relationships. That distinction is also why pool decay is so damaging — when the relationship context rots, you're left with a database that masquerades as a pool.
The economics of decay: why it's worse than you think
Most agency leaders understand that their database has some stale records. What's less intuitive is how stale records compound into real commercial costs — and how those costs compare to the alternative.
Start with the time cost. A recruiter searching for a Head of Operations role finds 30 candidates in the database matching the profile. She calls through the list. Four numbers are disconnected. Three people tell her they've been in a new job for a year and aren't interested. Two emails bounce. That's nine dead contacts out of 30 — a 30% decay rate in action. The time spent on those nine calls and emails isn't recoverable. And that's before she's identified a single qualified candidate.
Now compare two scenarios over 12 months for an agency with 200 placements/year and a 45-day average time-to-fill:
| Metric | Stale pool (30% decay) | Maintained pool (<10% decay) |
|---|---|---|
| % placements from existing database | 15–20% | 40–60% |
| Time-to-shortlist (database candidates) | 18–25 days | 4–8 days |
| Re-sourcing hours per placement | 12–15 hrs | 3–6 hrs |
| GDPR exposure | High (lapsed consents) | Low (rolling refresh) |
The time-to-fill difference is the commercial kicker. LinkedIn's 2025 Future of Recruiting research found that agencies using active internal databases filled roles 11 days faster on average than those relying primarily on ad responses. Across 200 placements a year, that's 2,200 consultant-days saved — the rough equivalent of adding two full-time recruiters without adding headcount.
"Reactivation beats re-sourcing every time — not because it's easier, but because you already have the relationship. A candidate who met you once is ten times more likely to respond than a cold outreach."
The three types of pool decay — and which is hardest to fix
Decay isn't one thing. It helps to separate it into three categories because the remediation strategies are different.
Contact data decay is the most visible. Emails bounce, phone numbers disconnect, LinkedIn URLs go stale when profiles are deleted or renamed. This is mostly fixable through automated enrichment — tools that periodically verify email addresses and re-match records against current LinkedIn data.
Status decay is subtler. A candidate was open to a move 14 months ago. They've since been promoted and are no longer looking. Or vice versa — they were happy when you spoke, but their company has since had layoffs and they're now actively searching. Your record says "passive, check back Q3" from 18 months ago. That data is worse than no data, because it shapes your prioritisation incorrectly.
Consent decay is the most legally consequential, and it's the one agencies handle worst. Under GDPR, if you have no ongoing legitimate interest or an expired explicit consent, you cannot legally contact that person — and you certainly can't build AI-generated shortlists from their data. The European Commission's GDPR guidance is explicit that consent for one purpose (a specific job application) does not extend indefinitely to future uses. Recruitment agencies that haven't set rolling consent review windows are sitting on a compliance liability.
Decay audit: a practical checklist
Before running any reactivation campaign, run this audit on your database. It's a one-time investment that makes every subsequent campaign more efficient.
Step 1 — Age segmentation. Tag every record by last meaningful interaction: 0–6 months, 6–12, 12–18, 18–24, 24+. Records in the 18-24 and 24+ buckets need enrichment before they're usable.
Step 2 — Email verification. Run every email address through a verification service. Remove hard bounces immediately. Flag soft bounces for a manual check. Unverified addresses in the 12-month+ bucket should be treated as stale until confirmed.
Step 3 — LinkedIn cross-reference. For records with a LinkedIn URL stored, check whether the current employer and title still match. Any mismatch is a status decay flag. Use our free AI resume parser to re-ingest updated CVs when candidates send them in response to a reactivation touch.
Step 4 — Consent review. Pull every record in the 12-month+ bucket. Does each one have a documented legal basis for contact? Explicit consent, legitimate interest with a documented assessment, or active application? If not, that record cannot be included in outreach or AI shortlisting until you've refreshed the legal basis.
Step 5 — Skills taxonomy clean-up. This is the slow one. Free-text skills fields ("strong communicator", "good with Excel") are unusable for AI matching. Convert the top 200 records in your pool to structured tags using your preferred taxonomy. That's the foundation for AI-assisted reactivation.
AI-assisted reactivation: what works and what doesn't
Reactivation campaigns have a mediocre reputation because most of them are generic: "Hi, it's been a while — any new roles I can help with?" That's a lazy message, and candidates have learned to ignore it. AI assistance doesn't fix a bad strategy. It amplifies it — for better or worse.
What AI does well in reactivation: personalisation at scale. If your CRM has clean data — current employer, last conversation notes, skills tags, the roles they interviewed for previously — an AI assistant can draft a re-engagement message that references those specifics. Not "it's been a while" but "last time we spoke, you were about to move into a new Head of Ops role at [employer]. How has that gone? We have a search right now that might be an interesting next chapter." That message gets a 3–5x higher response rate than the generic version, and it takes ten seconds to generate rather than five minutes.
What AI doesn't fix: genuinely lapsed relationships. A candidate who never replied to your last three attempts hasn't become more responsive because you're now using AI to draft the message. For those records, the right move is either a very different hook (a specific, time-sensitive mandate rather than a "staying in touch" message) or a realistic write-off and re-source.
"The best reactivation message is one that makes the candidate feel you actually remembered them — not just that you have their record on file. That distinction is exactly what clean CRM data enables."
Yena's AI search can surface the top-matching candidates from your pool against a new mandate brief — by semantic similarity rather than keyword matching. That means a candidate whose profile mentions "supply chain transformation" and "cross-functional leadership" can surface for a COO search even if they've never used that exact title. The match quality is only as good as the underlying data quality, which brings us back to the audit.
GDPR reactivation: the right sequence
For records where consent has lapsed or the legal basis is unclear, the reactivation sequence has to start with consent renewal — not a role pitch.
The cleanest approach: a short message that explains you still hold their details, gives them a one-click way to update their preferences or opt out, and only proceeds to role outreach once they've confirmed their details are current and they're still open to contact. This feels slower than jumping straight to a pitch, but it's the only defensible sequence under current GDPR interpretation — and it also produces a cleaner list, because you're only talking to candidates who've re-confirmed their interest.
See our full talent pooling guide for consent flow templates and retention window recommendations by record type.
Frequently asked questions
What is a talent pool in recruitment?
A talent pool is a curated, maintained group of pre-qualified candidates who have been assessed for fit, categorised by skills and availability, and kept engaged for future roles. It differs from a raw applicant database in that a talent pool implies an active relationship — the recruiter knows enough about each person to shortlist them quickly when a matching role opens. Effective talent pools typically fill roles 10–15 days faster than sourcing from scratch.
How fast does a talent pool decay?
Contact data in professional databases decays at roughly 25–30% per year — meaning around a quarter of your records will have at least one significant data point (email, employer, title) that's incorrect after 12 months. Status data (availability, motivation) decays faster, often within 6–9 months of the last direct conversation. Without a structured refresh cycle, a pool that looked healthy three years ago may be functionally unusable today.
What's the difference between a talent pool and a talent pipeline?
A talent pool is a broader reservoir of qualified candidates across roles and seniority levels, maintained proactively. A talent pipeline is typically role-specific: candidates actively progressing toward a particular position. Pools feed pipelines. For more on how this plays out in agency recruiting, see our guide to building pools that actually generate placements.
Can you use AI to reactivate a stale candidate database?
Yes, but only after a data quality audit. AI can personalise reactivation messages at scale, match candidates to new mandates by semantic similarity, and flag records most likely to be open to contact based on tenure and career trajectory signals. What it can't do is compensate for missing data, lapsed GDPR consent, or a fundamentally broken CRM schema. The audit comes first.
How often should you refresh talent pool data?
A rolling quarterly enrichment check (email verification, LinkedIn cross-reference) is the practical minimum for active pools. Full consent reviews should happen annually. Any candidate you haven't had direct contact with in 18 months should be treated as unverified until re-engaged — and records past 24 months without contact should either be re-confirmed or deleted, depending on your retention policy and the legal basis on record.
Talent pool strategy for recruitment success — why most databases fail and what to do about it (2024)
The talent pool problem isn't a technology problem. It's a discipline problem. The agencies that treat their database as a living asset — with regular audits, rolling consent management, and structured re-engagement — consistently outperform those that treat it as a filing system they'll sort through when they need it.
If you want to see how AI-assisted matching works against a properly structured candidate pool, explore Yena's candidate search — or start with the free AI resume parser to begin structuring your existing records.