The longer definition
A fuller definition has four parts.
Systematic. Research works to a written scope with defined fields and defined exclusions. Opportunistic name-gathering is not research, and the difference shows when someone asks how complete the dataset is.
Identification. Finding the employers first and the people second. A list built by searching for job titles finds people with the right label; a list built from the employer universe finds the market.
Verification. Corroborating each material finding against an independent second source, and recording confidence where that was not possible. This is the difference between a dataset and a guess with names in it.
In a defined labour market. Scope is written down before fieldwork: geography, employer set, levels, capability criteria, language requirement. Every number in the deliverable depends on it.
Research and what is built on it
| Service | What it adds to research | Output |
|---|---|---|
| Talent research | Nothing — it is the raw material | A verified dataset |
| Talent mapping | Organisation and analysis against a defined role or function | A named population with analysis |
| Talent intelligence | Interpretation against a business decision | Evidence and a recommendation |
| Talent sourcing | Approach and qualification | Interested, qualified candidates |
| Executive search | Approach, assessment, referencing and offer management | A completed appointment |
| Compliance support | Screening, credentialing and administration | A process that runs without your team doing the admin |
What a researcher actually does
Translate the brief into capability criteria
- What happens
- The requirement becomes a set of observable, testable attributes rather than a job title.
- Why
- Titles are not comparable between employers or borders. Capability is.
Build the employer universe
- What happens
- Every organisation that could hold the capability, from registers, filings, association membership and sector knowledge.
- Why
- This is the step that determines completeness, and the one most commonly skipped.
Identify individuals inside them
- What happens
- Named people located against the criteria, using public sources where coverage is good and telephone research where it is not.
- Why
- In the German Mittelstand, Austrian industry and Gulf operations, phone research is the only route to most of the market.
Verify and grade
- What happens
- Each material finding corroborated against an independent second source; confidence recorded where it could not be.
- Why
- An unverified dataset presented as fact is a liability attached to whoever presents it — which, in white-label work, is you.
Record provenance and hand over
- What happens
- Source and collection date against every material field, delivered in a format your systems ingest.
- Why
- Provenance is what lets a finding be defended when someone challenges it, and what lets you see later what has aged.
Accuracy and verification
The useful question to ask any research supplier is not “how accurate is it?” but “how do you know, and what do you do when you cannot corroborate?”
Our standard is corroboration against an independent second source for every material finding. Where that is not possible — common with privately held mid-sized employers — the record is delivered labelled single-sourced with the basis stated, rather than quietly upgraded into a fact.
The trap worth knowing about is derived sources. Two datasets that both originate from the same underlying feed are one source wearing two hats, and a great deal of aggregated workforce data has exactly that problem. We treat derived sources as single-sourced and say so.
How fast it goes out of date
| Field | Decays | Practical consequence |
|---|---|---|
| Current employer | With every job move in the population | An 18-month-old dataset in a mobile function has a substantial error rate |
| Job title and level | On promotion as well as on moves | Level is the field most often silently wrong |
| Contact details | Fast, with moves and with provider changes | Re-verify immediately before any outreach |
| Team structure | Slowly, except after reorganisation | Usually holds a year or more |
| Compensation | Within two quarters in fast-moving functions | Benchmarks need refreshing more often than populations do |
| Employer universe | Slowly | The most durable part of any research asset |
Because every field carries its collection date, you can see precisely what has aged rather than refreshing everything or trusting all of it. Re-verifying an existing dataset is usually the cheapest useful thing we do.
Can a database replace it?
For part of the job, yes. A database licence gives breadth, speed and a usable first pass, and in well-covered markets — UK professional services, US technology, Dutch commercial roles — it may be most of what you need.
What it cannot do is verify currency, reach employers that disclose nothing, or cover populations with low professional-network presence. In Germany, Austria and the Gulf that is most of the interesting market, which is why research there costs more per name and is worth more.
The honest framing is that databases and commissioned research answer different halves of the question, and most mature firms buy both. We set the comparison out in full on the platforms compared page.
The legal position
Talent research processes personal data, so data protection law applies in full. What it does not do is require consent: GDPR requires a lawful basis, and for professional research that basis is normally legitimate interests under Article 6(1)(f), supported by a documented balancing test.
What makes a specific project compliant is discipline rather than paperwork: professional fields only, a source recorded against each, a defined retention period, no special category data against named individuals, and a working route for someone to object and have their record removed.
Outside Europe the regime changes — the federal PDPL plus DIFC and ADGM in the UAE, the Privacy Act in Australia, state laws in the US — and we scope to whichever applies. The full position is on our GDPR compliance page.
What research cannot establish
- Internal performance data. Ratings, rankings and succession status are not observable, and any supplier offering them is either inferring or doing something you would not want attached to your name.
- Confidential individual pay. Except where it is disclosed — listed-company executive pay, published public-sector scales — compensation is reconstructed at population level, not verified per person.
- Undisclosed plans. Restructuring, acquisitions and hiring freezes that have not been communicated are not researchable.
- Security clearance. In most markets this is not public. We report it only where it is genuinely evidenced.
- What an individual will do. Research establishes that a population exists and how it is structured. Predicting one person is not research, it is guessing.
- Completeness in an undefined market. A dataset can only be complete relative to a written scope. Without one, ‘complete’ is not a meaningful claim.
Talent research FAQs
What is talent research?
Talent research is the systematic identification and verification of people and employers in a defined labour market. It builds the employer universe, identifies named individuals inside it, confirms employer, level and location by primary research, and records the source and date against every field.
It is the raw material. Talent mapping, talent intelligence, sourcing and executive search are all things you do with talent research once you have it.
Is talent research the same as talent mapping?
No. Research is the activity; a map is one of the things you can produce with it. A talent map is research organised and analysed as a named population for a defined role or function.
You can buy research without buying a map — most recruitment and search firms do, because they want the capacity rather than the presentation.
What is the difference between talent research and market research?
Market research studies markets, products and customers. Talent research studies labour markets: who works where, in what structure, at what level, and how that is changing.
The methods overlap, but the sources, the legal basis and the verification problems are different enough that general market research firms rarely do it well.
What does a talent researcher actually do?
Defines the capability criteria, builds the list of employers that could hold it, identifies named individuals inside those employers, verifies that each is current, records the source and date, and flags what could not be established.
In markets with thin public records — the Gulf, the German Mittelstand, Austrian industry — a large share of that is telephone research in the local language.
How accurate is talent research?
It depends on the verification standard, which is why you should ask about it. Our standard is corroboration against an independent second source for every material finding, with confidence recorded where that was not possible.
No dataset is perfectly accurate, and anyone claiming otherwise is describing a sales position rather than a method. What matters is that you can see which records are solid.
How quickly does talent research go out of date?
Faster than most buyers assume. Individual records decay with every job move in the population, so an eighteen-month-old dataset in a mobile function can have a substantial share of records wrong.
Structural findings — how a competitor is organised, where capability is concentrated — hold for a year or more. Because every field carries a collection date, you can see precisely what has aged.
Can research be done entirely from a database?
Parts of it. A database gives breadth and a fast first pass, and for well-covered markets that may be enough.
What it cannot do is verify currency, reach employers that disclose nothing, or cover populations with low professional-network presence. In Germany, Austria and the Gulf that is most of the interesting market.
Is talent research legal?
Yes, when scoped and documented properly. It processes personal data, so data protection law applies in full — but GDPR does not require consent for this, it requires a lawful basis, which for professional research is normally legitimate interests with a documented balancing test.
What makes a project compliant is the discipline: professional fields only, a source per field, a defined retention period and a working route to object. See GDPR compliance.
Should we build a research function in-house?
If you have continuous demand in markets your team knows, yes. The marginal cost per brief falls to near zero and institutional memory is valuable.
It struggles with languages and markets the team does not cover, with capacity spikes, and with the verification discipline, which takes longer to build than most firms expect. Many recruitment and search firms run a small internal team and buy a seat for overflow and for markets they do not know.
What can talent research not establish?
Anything that is not observable: internal performance ratings, confidential individual pay, undisclosed restructuring plans, and in most markets security clearance.
It also cannot tell you what an individual will do. It establishes that a population exists and how it is structured, not how any one person will respond.
What is white-label research?
Research delivered unbranded so that a recruitment firm, search firm or consultancy can present it to their own client under their own name. It is a standard arrangement in this industry and we support it explicitly rather than tolerating it.
You own the dataset, there is no restriction on resale, we never contact your client, and we do not take direct work from a client we have researched for you without your agreement.
How much does talent research cost?
A defined research project typically costs $3,000 to $8,000. A multi-market or programme engagement typically costs $9,500 to $30,000. A dedicated researcher seat typically costs $5,200 to $9,500 a month, with GBP, EUR and AUD equivalents on the cost page.