The Number Everyone Cites — and Misreads
When SHRM published its 2026 Top Five Workplace Issues, one figure dominated the discussion: nearly 70% of organisations reported difficulty filling full-time positions, reflecting persistent labour shortages and widening skills gaps that challenge US competitiveness and productivity. The statistic was picked up by every HR publication and cited in every board presentation. And almost everywhere it was cited, it was misread.
The instinctive reaction was to treat this as a volume problem. Post more jobs. Engage more recruiters. Expand the candidate pool. Move faster through the interview process. Offer sign-on bonuses. These are the tools organisations reached for — and the persistence of a 70% difficulty rate across three consecutive years of applying them tells you whether they worked.
In Talenbrium's Q1 2026 Workforce Pulse Survey, 63% of HR respondents reported that the primary constraint on hiring was not sourcing volume — it was candidate specification misalignment. The roles they were posting for did not match the talent pools that actually existed in their markets.
The 70% figure is not a recruitment failure. It is a workforce intelligence failure. Organisations are making hiring decisions without knowing the supply side of the equation — without knowing whether the role they have defined is achievable in their market at the seniority and skill specification they have set, at the salary they have approved.
Three Structural Mismatches Driving the Crisis
Talenbrium's analysis of hiring activity across 12,400+ employer organisations identifies three specific structural mismatches that explain the persistence of hiring difficulty across markets, industries, and role types. These are not strategic failures. They are information failures — decisions made without the data to make them correctly.
Mismatch 1 — Specification vs. Market Reality
Over a quarter of organisations — 28% according to SHRM's 2026 benchmarking — reported that filling full-time positions required new skills; among them, nearly half (47%) were existing roles that had been updated with new skill requirements. A data analyst role from 2022 has become a data scientist role with AI/ML requirements in 2025. The title stayed the same. The specification changed entirely.
The problem is that the labour supply did not keep pace with the specification change. BLS Employment Projections confirm that computer and information research scientists — the occupational category most closely tracking AI/ML engineering demand — are projected to grow at 23% through 2033. But that is a projection of future supply. Today's supply is the talent that already exists in the market. Talenbrium's job postings analysis shows that 78% of AI/ML engineering postings in the United States require five or more years of experience. Only 22% of current AI-focused professionals have that tenure. The specification and the available pool are structurally misaligned.
Mismatch 2 — Geography vs. Talent Distribution
Talent is not distributed evenly across geography, and hiring decisions frequently ignore this. When Talenbrium maps AI/ML engineering talent supply across US metropolitan markets, a single fact dominates: 33% of all AI engineer job postings are concentrated in California. If an organisation is hiring an AI engineering team in Ohio, they are competing for a talent pool one-third the size of what exists on the West Coast — at similar compensation expectations driven by the national salary benchmark, not the local one.
This matters not just for technology roles. Talenbrium's analysis of healthcare clinical talent supply finds that the ratio of registered nurse openings to qualified professionals in rural markets is 2.3 times more acute than in urban markets for identical roles. Organisations making location-agnostic hiring plans using national benchmarks are systematically over-estimating the talent available to them.
"The hiring difficulty score for a cybersecurity engineer in Northern Virginia is 9.1 out of 10. For the same role in Austin, it is 8.4. That single data point changes the entire staffing plan — and most organisations don't have it before they open the requisition."
— Talenbrium Hiring Difficulty Analysis, Q1 2026Mismatch 3 — The Unicorn Specification
Talenbrium's co-hiring cluster analysis reveals a consistent pattern: organisations post for a single role that implicitly requires the output of three distinct talent profiles. A common example in 2025–2026 is the AI/ML engineer posting that also requires cloud architecture experience, model governance knowledge, and regulatory compliance familiarity. The candidate who holds all four at senior level does not exist in sufficient numbers in any market.
When Talenbrium analyses which roles are typically hired alongside AI/ML engineers in practice — the co-hiring cluster — the pattern is clear. Cloud/DevOps engineers are co-hired in 71% of cases. Data engineers in 64%. AI governance analysts in 31%. Organisations that define these as separate, sequential hires instead of simultaneous cluster hires close roles 40% faster than those trying to find the unified specification.
MID-PAGE CTA
Need this analysis for your specific roles and markets?
Talenbrium produces bespoke hiring feasibility assessments — Hiring Difficulty Scores, talent pool sizing, and salary benchmarks — for the exact role families and geographies your organisation is planning to hire into.
Discuss a Custom Requirement Download Sample ReportWhat the Data Actually Says About Your Market
The structural mismatches above are not abstractions. They translate into specific, measurable consequences that appear in Talenbrium's employer database tracking. Three patterns repeat consistently across markets and sectors.
First, organisations with above-specification hiring — roles requiring experience profiles that exist in less than 15% of the available talent pool — experience time-to-fill rates 2.6 times higher than sector peers. This is not because they are less attractive as employers. It is because the candidate they are looking for does not exist in sufficient numbers to be found by any recruiting process, however sophisticated.
Second, geography-agnostic hiring plans — organisations that apply national salary benchmarks without city-level talent supply analysis — consistently under-offer relative to the local competitive market. Talenbrium's compensation benchmarking shows the city-level variance for equivalent senior engineering roles running as wide as 35% between the highest-cost market (San Francisco Bay Area) and mid-tier growth markets (Denver, Atlanta). An organisation benchmarking against national medians in San Francisco is offering 20–25% below what the local market requires.
Third, organisations that separate co-hiring cluster roles — hiring sequentially rather than simultaneously — experience compounding delays. When the AI engineer is hired before the MLOps engineer and the data engineer, the AI engineer spends the first 60–90 days of their tenure blocked on infrastructure that the missing cluster roles would have provided. The effective productivity cost of sequential cluster hiring exceeds the cost of carrying the open requisition.
Organisations that have been running above-specification roles open for 90+ days face a compounding problem: high-intent candidates who applied early have accepted offers elsewhere, leaving only passive or below-specification candidates in the active pipeline. Every 30 days above 60 total days-to-fill reduces the qualified candidate pool by an estimated 18–22% as alternatives are accepted.
A Framework for Making Hiring Decisions with Market Intelligence
The solution to the hiring paradox is not more recruiting. It is information — specifically, the three questions that any hiring decision should be able to answer before the requisition is opened.
| Question | What you need to know | Risk if unknown |
|---|---|---|
| Is this role achievable in this market? | Talent pool size for the role + seniority + location combination; Hiring Difficulty Score; active competitor posting density | High — opening a requisition with no achievable candidate pool |
| Is the compensation competitive enough to close? | P25 / P50 / P75 / P90 for the exact role in the exact market; YoY salary movement; skill-specific premium requirements | Medium-High — sourcing candidates you cannot close |
| Are we hiring the right cluster? | Co-hiring frequency analysis for the core role; whether adjacent roles should be opened simultaneously; expected time-to-fill for each cluster position | Medium — sequential hiring extending total time-to-productivity |
| Is the specification realistic? | What % of the available talent pool holds the exact skill combination specified; which requirements are universal vs. differentiating | High — unicorn specification eliminating 80%+ of viable candidates before interview |
| When is the best time to hire in this market? | Seasonal hiring peaks; competitor hiring cycle analysis; talent supply seasonality by role family | Low-Medium — competing in peak demand windows when cost and time-to-fill are highest |
These five questions are not answered by a recruiter's intuition or a LinkedIn search. They are answered by labour market data — job postings analysis, salary benchmarking, talent pool sizing, and competitive employer hiring intelligence. The organisations that answer them before opening a requisition consistently outperform on time-to-fill and offer acceptance rates.
Among HR respondents whose organisations conducted market intelligence assessments before opening senior technical requisitions, average time-to-fill was 47 days — versus 89 days among those who relied on recruiter judgment and ad-hoc compensation benchmarking. The intelligence gap, not the recruiting gap, explains the majority of the difference.
The Sector Picture: Where the Paradox Is Most Acute
The hiring paradox is not uniform across sectors. Talenbrium's sector-level analysis identifies three industries where the mismatch between employer specification and available talent supply is most structurally embedded — and where market intelligence has the highest decision value.
Technology. AI/ML engineering roles represent the most acute supply-demand imbalance in Talenbrium's US dataset. Postings grew 163% between 2024 and 2025, while the experienced talent pool grew at a fraction of that rate. The result is a market where three qualified candidates exist for every one open role — but 78% of postings ask for the top 12% of the experience distribution. The apparent shortage is partly real and partly specification-driven.
Healthcare. The Bureau of Labor Statistics 2024–34 Employment Projections confirm that registered nursing generates approximately 194,500 openings per year. Total projected nurse graduate supply over the same decade is barely sufficient to cover a single year's openings. This is not a specification problem — it is a pure pipeline problem, compounded by 42 US states projected to face nursing shortages within five years, according to US government health workforce projections. Market intelligence here changes not the hiring decision but the location decision: where to source, where to place roles, and where to build long-term pipeline partnerships.
Financial Services. BFSI hiring difficulty is concentrated at the intersection of technical and regulatory expertise. Roles requiring data science skills combined with regulatory compliance knowledge — risk modelling analysts, regulatory technology engineers, AI governance specialists — are the fastest-growing category in Talenbrium's financial services employer database and the hardest to fill. The specification challenge is real: the combination of quantitative modelling experience with EU regulatory knowledge (DORA, EU AI Act) or US SEC guidance on algorithmic employment decisions represents a talent profile that no graduate pipeline yet produces at scale.
What This Means for 2026 Workforce Planning
The hiring paradox will not resolve itself in 2026. The structural supply constraints in technology, healthcare, and financial services are multi-year problems rooted in education pipeline capacity, demographic trends, and the pace of skill evolution. But the portion of the problem that is an information failure — organisations making hiring decisions without labour market intelligence — is solvable immediately.
SHRM's data shows that over a quarter of organisations required new skills in 2025 that did not exist in their job descriptions two years prior. Cedefop's 2025 Skills Forecast for the European Union confirms that ICT professionals and electrical engineers represent some of the most widespread labour shortages across EU member states simultaneously. The BLS projects that healthcare and social assistance will account for the majority of US employment growth through 2034 — but training capacity prevents supply from meeting that demand at the pace the projections suggest.
These are not trends that more recruiting solves. They are structural realities that inform strategy — where to hire, what to specify, what to pay, when to build pipeline rather than buy it, when to invest in internal reskilling rather than external hiring. The organisations that build workforce intelligence into these decisions before the requisition is opened will outperform those that treat every open role as a recruitment problem to be solved.
The 70% hiring difficulty figure will not move materially until organisations change the question they are asking. The question is not "how do we fill this role?" The question is "can this role be filled in this market at this specification and price — and if not, what has to change?" That question requires labour market intelligence, not more recruiters.
Want this intelligence applied to your hiring decisions?
Talenbrium provides bespoke Hiring Feasibility Assessments, Talent Pool Analysis, and Salary Benchmarking for specific roles, markets, and industries. Reach out to discuss your requirements — there is no template engagement.
