Is candidate rediscovery just database search with better marketing?

It is if the underlying records are left alone, which is why most attempts disappoint. The difference is whether anything has been done to the free-text notes first. Matching against the same thin, stale structured fields produces the same results regardless of how the matching works — the gain comes from making the prose searchable, not from a better algorithm.

How old is too old for an ATS record?

It depends entirely on which part of the record you mean. A work email is dead the day someone leaves. A mobile number is often good for years. A job title is meaningless after about two years, and a stated preference — would not commute, wants remote — is frequently still true long after. Grading a whole record as fresh or stale throws away that distinction.

Can this tell us who is available right now?

No, and be wary of anything that claims it can. Availability is not recoverable from a record of any age. What a well-structured database gives you is a much better-ordered list of who to ask, and a specific reason to open the conversation — which is a different and more honest claim.

Do we need to replace our ATS to do this?

No. The work reads existing records and writes back structure and a shortlist. Migrating an ATS costs a year and is dominated by data history and retraining rather than licences, and it would not address the reason the search fails, because the same unstructured notes would move across with everything else.

Guide

Your database is a better source than your job board

Most agencies pay a job board every year to find strangers, while sitting on thousands of people who already know them. The reason that list goes unused is specific, and it is not that matching is hard.

In short

A staffing firm's own database is usually a better source than the job board it pays for, and it goes unused for a specific reason: the data is unusable before matching begins. Records decay within months, contact details rot at different speeds, and the valuable material — what someone wanted, and why they declined — sits in free-text notes that keyword search cannot reach.

Who this is for: Staffing firm owners and delivery leads deciding whether to renew a job board subscription, or wondering why the last database re-mining push produced nothing.

11 min read · Last reviewed 25 September 2026

There is a trade every staffing firm makes without quite deciding to. You pay a job board tens of thousands a year for access to people who have never heard of you, whose details were last confirmed by somebody else, and who are being called this week by every other agency running the same search. Meanwhile your own system holds thousands of people who have spoken to you, been placed by you, or turned you down for a reason one of your recruiters wrote down at the time.

The second list is better on every dimension that matters. It is also, in most firms, completely unused. This is about why, and what would have to change.

What you are actually buying

Job boards have got worse in a way that is not the boards' fault and is not going to reverse. Generative tools made applying free, so postings that drew a hundred applicants now draw a thousand, and the ratio of serious to speculative has collapsed. Profiles sit unchanged for years while the person behind them moves twice. Some boards supplement their counts with data scraped from elsewhere, so a proportion of the people you call never registered with them at all.

None of that is a reason to cancel tomorrow. Boards still fill roles, and for a genuinely cold search on an unfamiliar skill they may be the only option. It is a reason to notice that the price per useful conversation has been rising for years while the invoice has too.

Why the better list goes unused

Ask why nobody re-mines the database and you will usually be told there is no time, or that the search never returns anything good. The second answer is the real one, and it has a specific cause that is worth stating plainly:

The bottleneck is not matching. It is that the data is unusable before matching begins. Records decay within months. The structured fields are thin and stale. And everything genuinely valuable — what a person wanted, what they would move for, why they said no — is in free-text notes that no keyword search can reach.

This is why database projects stall in the same place every time. A firm buys a tool that promises better matching, points it at the existing records, and gets results no better than the search it replaced — because the tool is matching against the same thin fields, and the part of the record that would have made the difference is prose.

It is also why the vendors are not lying, exactly. The matching layer generally works. It is being asked to work on material that will not support it.

Different things rot at different speeds

Treating a record as uniformly stale, or uniformly fine, is the mistake underneath most bad reactivation lists. The useful model is per-aspect:

What	How fast it goes	Why
Work email address	Dead the day they leave	It belonged to an employer, not to them. The single fastest-decaying field in your system.
Mobile number	Often good for years	People keep numbers across jobs. Usually the most durable contact you hold.
Job title on the record	Meaningless after about two years	It describes a job they have left. This is what filters them out of your own search.
Rate expectation	Roughly a year	Moves with the market, and the market has moved.
What they told you they wanted	Slow, and partially durable	Someone who would not commute two years ago usually still will not.
Availability	Immediately, and permanently	Not knowable from a record of any age. Ever. See below.

The last row is the one to be strict about. No record tells you whether somebody is free today, however recent it is. A reactivation list that implies otherwise gets found out on the first call, and a recruiter who is embarrassed once will not use the list again.

The notes are the asset

A recruiter note is written in thirty seconds between calls. It is abbreviated, ungrammatical, full of internal shorthand, and it contains things no job board will ever hold:

Why they turned down a role — the single most valuable field in the system, because a requisition that differs on exactly that point is a reason to call today.
What they are actually doing now, which is frequently two levels above the title on their record.
What they said they wanted, in their own words rather than as a dropdown value.
When a current assignment ends, which turns a cold call into a diarised one.
Client feedback from a completed placement, which is evidence no board can supply at any price.

Structuring this material is unglamorous and it is the whole job. It is also the part AI is genuinely good at, as opposed to the parts it is usually sold for: reading prose, pulling out stated facts, and keeping them attached to the sentence they came from so a person can check them.

See it working

We built a worked example: six ATS records as they actually are — thin fields, inconsistent tags, notes typed between calls — against one live requisition. It shows what can be recovered from each, what has rotted, and why one of them is worth a call that no job board could have told you about. Synthetic data, real logic.

See it on six decayed records

On the statistics, including ours

This subject attracts confident numbers, and they deserve more scepticism than they usually get. Two worth knowing, with their provenance attached:

Bullhorn's 2026 GRID Industry Trends Report, based on a survey of roughly 2,300 recruitment professionals conducted in late 2025, reports that 85% of firms with a redeployment plan place in under 20 days. That is a reasonable sample and a plausible finding. The same report attributes 47% higher redeployment rates to firms using AI and automation inside their ATS — which is a software vendor measuring its own customers, and should be read accordingly.

Beyond that you will find figures circulating about engagement rates and cost-per-reactivation, most of them published by companies selling reactivation tools and traceable to no primary source. We are not going to repeat them. We also have no figures of our own, because we have not run this on your data or anyone else's.

What we would suggest instead is that this is unusually easy to measure yourself, and the answer is specific to your firm rather than to an industry average.

Take the last twenty roles you filled. For each, check whether the person placed was already in your database before the search started.
Count how many were. That fraction is what you paid a board to find twice.
Take your last ten placements that came from a board. Work out the fully loaded cost — subscription share, sourcing hours, screening hours — per placement.
Pull fifty records nobody has touched in two years and read the notes. Not the fields — the notes. Count how many contain a stated reason for declining, a rate, or an end date.

That fourth step takes about an hour and tells you more than any vendor statistic. If the notes are empty, reactivation is not your next move and no tool will change that. If they are full of exactly the material described above, you are sitting on something.

What this does not fix

An empty database. If nobody recorded anything, there is nothing to recover, and the honest answer is that enrichment comes before reactivation.
A discipline problem. If notes are not being written now, the same gap will exist again in two years.
Genuinely cold searches. A skill you have never recruited for is not in your database at any quality, and that is what boards are for.
Availability. Worth saying twice. It is the first question on every call, not an output of any system.

The realistic claim is narrower than the marketing around this subject, and it is still worth acting on: a meaningful share of the people you are about to pay to find are already yours, and the reason you cannot see them is fixable.

Common questions

Is candidate rediscovery just database search with better marketing?: It is if the underlying records are left alone, which is why most attempts disappoint. The difference is whether anything has been done to the free-text notes first. Matching against the same thin, stale structured fields produces the same results regardless of how the matching works — the gain comes from making the prose searchable, not from a better algorithm.
How old is too old for an ATS record?: It depends entirely on which part of the record you mean. A work email is dead the day someone leaves. A mobile number is often good for years. A job title is meaningless after about two years, and a stated preference — would not commute, wants remote — is frequently still true long after. Grading a whole record as fresh or stale throws away that distinction.
Can this tell us who is available right now?: No, and be wary of anything that claims it can. Availability is not recoverable from a record of any age. What a well-structured database gives you is a much better-ordered list of who to ask, and a specific reason to open the conversation — which is a different and more honest claim.
Do we need to replace our ATS to do this?: No. The work reads existing records and writes back structure and a shortlist. Migrating an ATS costs a year and is dominated by data history and retraining rather than licences, and it would not address the reason the search fails, because the same unstructured notes would move across with everything else.

Related guides

Where a recruitment desk actually loses its week

Ask a recruiter where their time goes and they will say sourcing. Watch the desk for a week and it is rarely sourcing. It is the packaging, chasing and re-keying around it.

Screening CVs against a brief without a black-box score

Most screening tools answer the wrong question. They tell you how well a CV matches a brief, when what a recruiter needs to know is what the CV actually evidences and what it simply does not say.

If this describes a process you recognise — we are a small company in Vadodara that builds exactly this kind of thing. No obligation and no sales sequence: get in touch or try the automation finder, which will tell you when the answer is not to automate.

+91 93131 52136 · hello@twishivplatforms.com