The Complete Talent Market Guide for 2026

Talent mapping is structured research into a defined talent market: who holds a given capability, which organisations they sit in, how those teams are structured, where the people are, what they are paid and how likely they are to move. The output is evidence about a market rather than a shortlist for a vacancy, which is why it is commissioned before a hiring plan, a location decision or a pay band is committed to.

It is normally a fixed project fee rather than a percentage of salary, and the fee is a function of researcher days. The variables that move it are population size, number of geographies, the verification standard applied to each data point, how deep the compensation layer goes, and whether you want a one-off snapshot or a refreshed map. Any firm quoting a single headline price for "a talent mapping project" either sells a fixed template regardless of your question or will re-quote once it sees the scope.

Four to eight weeks from signed scope to delivered report is typical. A narrow single-market map can land in two to three weeks; a multi-country map covering several functions runs to ten or twelve. The largest single cause of overrun is not research difficulty but scope change — the target organisation list moving in week three.

A target organisation list with inclusion rationale, the organisational topology of each target team, the named population with title, location, tenure and career history, total compensation bands by level and location, geographic distribution, movement analysis showing who hires from whom, availability signals, a representation view, and a written set of findings answering the original question. Ask any provider for a redacted sample of all of these before commissioning.

Recruitment answers "who can we hire for this role now". Talent mapping answers "what does this market look like" — a question asked before you commit to a plan. The practical test is what you receive if you decide not to hire: a genuine mapping engagement still delivers in full, because the deliverable is the research. If the answer is "nothing", you bought recruitment with a research label on it.

A lighter, vacancy-adjacent form of talent mapping: establishing who in the market could do a specific job that is open or imminent. It usually produces named individuals without the full market-structure, compensation and movement layers that a strategic talent map carries. Candidate mapping is narrower again — identification against a brief, with no market-structure layer at all.

It replaces assumption with evidence at the point where mistakes are most expensive. Concretely: it tells you whether a hiring plan is deliverable from the available pool before you commit to it, whether your pay band is competitive against the organisations you actually lose candidates to, where a capability is concentrated geographically, whether a representation target is achievable, and how deep an external succession bench really is. It also produces a reusable asset — the same research supports the next pay review and the next location decision.

Ask how they are paid, who owns the output, whether you can see a redacted sample from a comparable engagement, who actually does the research, how every compensation figure is evidenced, what you receive if you do not hire, whether they have mapped this market before, how they handle off-limits restrictions, and when the underlying data was gathered. The off-limits question matters structurally: a search firm cannot map a market it is contractually barred from approaching, and one that places from its own map has an incentive for the map to conclude that a hire is the answer.

No single tool produces a finished map. There are four categories — people-data platforms, labour market analytics, compensation data tools and org-chart visualisation — and each solves one part. Platforms are strong on breadth and weak on truth: titles are self-reported and non-comparable, reporting lines are almost never present, and a meaningful share of profiles describe a job the person has already left. Teams that get good results use tools to establish market shape, then verify structure, seniority and compensation by research.

Yes, and it is a reasonable architecture. The API gives you profile records at volume with employer, title, location and tenure. What you still have to build is the hard part: title normalisation onto a levelling scheme, which does not generalise across sectors; reporting-line inference, since topology is rarely in the data; staleness handling; a compensation layer, which is not in the API at all; and a documented lawful basis and retention policy for processing personal data at scale in every jurisdiction you touch.

Talent mapping applied to small, senior populations — CXO, board and functional leadership — where public compensation data is thin or absent. It leans more heavily on organisational structure, tenure patterns and ownership-event timing than on profile data. For private-equity-backed and high-growth companies the question is usually bench depth across a portfolio rather than a single seat, which makes it a maintained asset rather than a one-off exercise.

Yes — Europe, Canada, India, the Gulf and Asia Pacific, including Dubai, the Philippines and Thailand. The method does not change between markets; data availability does. Gulf markets are high-mobility and heavily expatriate, with packages where allowances can exceed base pay. Philippines and Thailand have deep shared-services, engineering and clinical capability but thinner public profile coverage, so mapping leans harder on direct research and local-language sources.

Mapping applied to the scarcest and least title-comparable population we work with. "AI engineer" spans research scientists, ML platform engineers and applied developers at very different prices, so levelling by scope rather than title matters more here than anywhere. Compensation moves fast enough that data older than a quarter is stale, and equity is frequently the majority of the package — which makes total-compensation normalisation essential rather than optional.

Talent mapping produces the evidence; talent intelligence interprets it against a specific decision — hiring, workforce planning, competitor analysis, location strategy, compensation or organisational design. Talent analytics usually refers to the quantitative layer within that: supply and demand modelling, density by location, attrition and movement rates. In practice they are stages of the same work rather than separate products.

A succession bench is the set of people, internal and external, who could credibly take a critical role. Most organisations build the internal half properly and leave the external half as a half-remembered name. An external bench built from a talent map is an evidenced, maintained list — who they are, where they sit, how long they have been there, what it would cost and how that has changed since the last refresh. It is scoped around two or three critical roles and refreshed on a cycle rather than delivered once.

Sixteen states plus the District of Columbia have a pay transparency law in force as of October 2026. Fourteen states plus DC require the range in the job posting: Colorado, California, Washington, New York, Hawaii, Maryland, Illinois, Minnesota, New Jersey, Vermont, Massachusetts, Virginia, Maine and Connecticut. Nevada and Rhode Island require disclosure on request or at offer rather than in the posting. Delaware's posting law takes effect on 26 September 2027.

Fifteen jurisdictions require a pay range in the posting itself: California, Colorado, Connecticut, the District of Columbia, Hawaii, Illinois, Maine, Maryland, Massachusetts, Minnesota, New Jersey, New York, Vermont, Virginia and Washington. Employer-size thresholds vary from no floor at all in Colorado, Connecticut, Maryland, Virginia and DC, up to 50 employees in Hawaii.

The same fifteen. Three were added during 2026: Virginia on 1 July, Maine on 29 July and Connecticut on 1 October. If you are reading a guide that lists twelve or thirteen states, it was written before July 2026 and is out of date.

Eight jurisdictions also require a general description of benefits: Colorado, Washington, Maryland, Illinois, Minnesota, New Jersey, Connecticut and Delaware from 2027. DC requires healthcare benefits to be disclosed before the first interview. Colorado adds promotional-opportunity notice duties. Maine, Vermont and Delaware require commission-only and tipped roles to be labelled as such rather than given a range. Several states also reach internal postings, promotions and transfers.

In the 34 states with no posting requirement in force today (the 33 with no rule at all, plus Delaware until September 2027), yes. But no state permits an employer to prohibit employees from discussing their own pay with each other — Section 7 of the National Labor Relations Act protects that conversation for most private-sector employees nationwide, which makes pay-secrecy clauses unenforceable against them in all 50 states.

Generally yes. Most state guidance treats a role that can be performed from a covered state as covered by that state's law, so a national remote posting inherits the strictest rule it touches. Illinois expressly reaches roles performed outside the state that report to an Illinois supervisor or office, and Vermont covers remote roles predominantly serving a Vermont office. Geo-fencing a posting is fragile because aggregators repost adverts without your location logic.

The range you genuinely expect to pay the person you hire, set before the role is advertised and supported by a band or a benchmark. California's SB 642, effective 1 January 2026, defines the pay scale as a good-faith estimate of what the employer reasonably expects to pay for the position upon hire — which is aimed directly at implausibly wide ranges posted to preserve flexibility. In practice the test applied is whether your published range matches what you actually paid recent hires into that role.

Pay transparency is a disclosure duty: publish or provide a range. The violation is the missing number, and the risk sits in the advert. Pay equity is a substantive duty: pay people doing substantially similar work the same, with only a defined list of lawful reasons for a differential. The violation is a gap you cannot justify, and the risk sits in your actual pay data. Transparency makes equity exposure visible, which is why bands should be audited before they are published.

In several states, expressly. Virginia requires a range in every public or internal posting including promotions and transfers. Connecticut and Delaware cover internal postings. New York, Maryland, Massachusetts and Colorado also reach internal moves, and Illinois and New Jersey add a duty to notify existing staff of openings. Hawaii is the main jurisdiction that carves internal transfers out.

Indicatively, from $300 for a first violation in New Jersey up to $20,000 per subsequent violation in the District of Columbia, with Colorado, California and Illinois reaching $10,000. Washington allows a private right of action with statutory or actual damages plus costs and fees. Most regimes begin with a notice or cure period, so the meaningful risk is a pattern across many adverts rather than a single omission. Confirm current figures with counsel — several states have amended their penalty provisions since enactment.

Define the job by scope rather than title, name a documented peer set of 15 to 30 comparators, pull live posted ranges from the disclosure jurisdictions, separate posted pay from accepted pay, normalise everything to total target compensation rather than base, then set the band and test whether a reachable population actually exists at that number. The last step is the one usually missed, and it is the difference between a compliant range and a fillable role.

Running your existing population against an externally validated band to find unexplained differentials — by grade, tenure, location and demographic group — and mapping each one to a lawful reason. It is the inward-facing half of compensation benchmarking, and it is what stops a newly published range from becoming a documented comparator that your own employees measure themselves against.

Equity benchmarks are the normal grant sizes, vesting terms and refresh practices for a given level and sector — the figures you need to compare a lower cash offer plus equity against a higher cash offer. They matter more than they used to for compliance: California's SB 642 expanded the definition of wages to take in bonuses, equity, benefits and allowances. Most statutes require only a general description of benefits rather than a precise grant figure, but a range that ignores equity entirely will understate your competitiveness for senior and technology roles.

Capabilities for which demand exceeds the supply of qualified, experienced people in a given market. Scarcity is relative to location and time, not absolute — a skill can be plentiful in one metro and scarce in another, and scarcity shifts as training pipelines and demand change. It is measured by time-to-hire across a long run of requisitions, reliance on external search, offer-decline rates, and the size of the addressable population against open demand. Scarce talent is the same concept expressed as people rather than capabilities.

Because the floor rises fastest where supply is thinnest. A scarce-skill population can now read competitor bands in published adverts, is the most able to act on that information, and is the hardest to replace if it leaves. Scarce roles should be benchmarked separately and quarterly, priced against the specific comparators who compete for that exact population, and checked for pool size before a band is committed to.

Not a general one covering private employers. Federal contractors face their own disclosure obligations, and bills have been introduced in Congress repeatedly without being enacted. The state-by-state patchwork should be treated as the permanent operating environment rather than a temporary stage before federal harmonisation.

Quarterly for scarce-skill and senior roles, annually for the rest. Published competitor ranges change continuously now that 15 jurisdictions mandate them, which makes a quarterly re-pull cheap and a stale band expensive. Add a legislative review each January and July, when most effective dates land.

The statutes require the range to be a good-faith estimate, not a binding offer, so an offer below the top of the band is normal and lawful. What creates exposure is a pattern of offers falling outside the published range, or well below it, because that is the evidence used to argue the range was never genuine. If the role has changed enough to justify a different number, repost it with the revised range.