Full Research: Latest HR & Talent Statistics (August 2026)

Last reviewed August 2026

This page collects the HR, talent acquisition, engagement, culture and learning statistics worth citing in 2026, with every figure traced back to the organisation that produced it. Where a widely repeated number turned out to have no source, we have said so rather than passing it on.

One finding shapes most of the others. Across Australia, people have stopped moving. Job mobility has fallen four years running, recruitment difficulty is at a series low, and average turnover is down. At the same time, employers still cannot fill roughly a third of the vacancies they advertise. Fewer people are moving, and the ones who do move are harder to place well. That combination raises the cost of every individual hiring and development decision you make.

How to use this page

Every statistic below names its source, its sample and the period it covers. Quote it with that attribution attached, because a number without a date and a population behind it is not evidence.

• Cite the original, not us. We link to the publisher so you can check the figure yourself.
• Watch for different instruments. Gallup reports 20% engagement and Qualtrics reports 68%. Both are correct on their own scales, and putting them side by side without explaining that is how HR content loses credibility. If you are building your own measure rather than buying a benchmark, our employee survey question banks cover engagement, pulse, onboarding, exit, manager effectiveness and culture.
• Market sizing is a forecast, not a measurement. Where we quote market values we name the firm and the base year.
• Check the last-reviewed date. Labour market data goes stale within a quarter.

Hiring and shortlisting in 2026

Australian employers are receiving fewer applicants per vacancy than two years ago, and still filling fewer roles. The Jobs and Skills Australia survey asks recruiting employers how many people applied, how many were qualified, and how many were actually suitable. The gap between those three numbers is the whole problem.

Suitable applicants per vacancy have actually improved, from 3.2 in the March quarter of 2024 to 5.2 two years later. The fill rate has not followed. It peaked at 72.1% and has fallen back to 68.2%, so close to a third of advertised Australian vacancies still go unfilled.

Application volume has doubled globally

Outside Australia the picture looks different, and the reason is that most global benchmarks measure applications through an applicant tracking system rather than surveying employers. Greenhouse, working from 640 million applications across more than 6,000 companies, recorded applications per job rising from 116 in 2022 to 244 in 2025 while the number of recruiters per organisation more than halved.

In Australia, SEEK reported applications per job ad at their highest level on record in July 2026, after seven consecutive monthly increases, while job ads themselves fell 6.0% year on year. More people are applying for fewer roles, and employers still report that only a quarter of them are suitable. That is a shortlisting and matching problem, not a sourcing one.

Employee engagement statistics

Global engagement fell to 20% in 2025, the lowest Gallup has recorded since 2020 and the first back-to-back annual decline in the series. The interesting part is who accounts for the fall.

That matters because managers are the mechanism organisations rely on to deliver everything else. Retention, culture, performance and development all run through the same layer that is currently degrading fastest.

Engagement by region

Australia and New Zealand sit almost exactly on the global average for engagement. On wellbeing the region looks completely different: 55% of workers are thriving, second highest in the world. People here are living good lives and having ordinary workdays.

Workplace culture benchmarks

Culture stopped being a soft agenda item in 2026 and became an AI readiness question. Deloitte's global human capital research found 65% of organisations believe their culture needs significant change because of AI, and 34% say culture is actively inhibiting their AI goals.

The clearest evidence that reading your people correctly is worth something comes from SHRM. Workers who say their organisation effectively addresses workplace needs report 91% job satisfaction. Where workers say it does not, satisfaction is 44%.

New hires are having the worst of it

Qualtrics, surveying 33,831 employees across 24 countries, found engagement among new hires down seven points to 65%, the sharpest fall of any group. New hires' sense that they could challenge the status quo dropped from 64% to 50% in a year.

The onboarding window is where culture either takes or fails, and it is deteriorating faster than anywhere else in the employee lifecycle. Measuring it is the job of a culture and engagement platform, and definitions for the terms used on this page are in the HR glossary.

Talent management statistics: mobility, turnover and retention

Australians have stopped changing jobs. Job mobility has fallen four years running, to 7.2% in the year to February 2026, the lowest in the current series.

Employer-reported data agrees. Average Australian turnover was 14% for the twelve months to December 2025, and the share of organisations running turnover above 20% fell to 23%, the lowest AHRI has recorded since it began measuring in May 2023. Recruitment difficulty fell to 25% of recruiting employers, also a series low.

Who employers screen out before they start

The most uncomfortable Australian finding in this year's data is that deliberate exclusion is rising even as hiring gets easier. Almost seven in ten Australian employers report actively excluding people with certain characteristics when recruiting.

AI in hiring: adoption, evidence and candidate reaction

This is where the gap between what gets said about AI and what gets measured is widest. Adoption is lower than the discourse implies, the benefits being reported are almost entirely about speed, and nobody has published credible evidence that AI improves who gets hired.

What organisations say they get from it

That split is worth sitting with. Speed is a process problem, and process problems are what most HR technology already solves. Working out who will actually do well in the role is a different problem, and it is the one the tooling is not yet touching.

Adoption is lower than you would think

• 39% of organisations use AI anywhere in HR, and only 18% use it broadly in hiring.
• 49% of Australian employers use no AI at any point in recruitment or selection.
• 17% of Australian employers use AI for screening and shortlisting.
• Inside HR software specifically, only 9% of organisations report using an HR application with embedded AI. Most reported AI use in HR is people opening ChatGPT.

The two sides of the desk do not agree

The candidate objection is more specific than a general dislike of the technology. Only about one in five candidates want less AI in hiring overall. What they object to is being assessed without being told, and being left with no reply.

Fake candidates, and what is actually measured

Regulation is moving, and not only forwards

In Australia, the Australian Public Service Commission's principles for agency use of AI in recruitment took effect on 1 June 2026. Two provisions matter beyond the public service, because they are a reasonable statement of where the standard is heading: AI tools must not make the final decision in recruitment outcomes, and all recruitment and employment decisions must be explainable and defensible.

Internationally the picture is less tidy. Several jurisdictions have delayed rather than advanced their obligations, with the EU's high-risk employment provisions and Colorado's AI Act both pushed back during 2026. Anyone planning compliance work on the original dates should recheck them.

AI, skills and the shape of work

The headline claim of the last three years has been that skills are churning faster every year. The World Economic Forum's own measure says the opposite.

Is AI eating entry-level work?

In the United States, Stanford's Digital Economy Lab found a widening employment gap for 22 to 25 year olds in AI-exposed occupations, rising from 15% to 19%. In Australia, the federal department responsible for employment looked for the same effect and did not find it.

Australian employers report the same thing from the other direction. In AHRI's December quarter survey, 41% of organisations said AI had led to an increase in entry-level roles, against 19% reporting a decrease.

What AI is measurably doing to productivity

Self-reported productivity gains from AI are large and consistent. Measured gains are neither. In a randomised trial of experienced open-source developers, METR found that using AI tools made them about 19% slower, while the same developers believed they had been roughly 20% faster. They misjudged their own productivity by around 40 percentage points.

That does not mean AI has no value. It means self-report is not evidence of it, and any business case built on people saying they feel faster is built on sand.

What actually predicts job performance

The most quoted table in hiring is out of date. For twenty-five years, selection practice leaned on Schmidt and Hunter’s 1998 meta-analysis. In 2023, Sackett, Zhang, Berry and Lievens rebuilt the estimates and most of them came down, some sharply, and the order changed.

The correction was not a mistake in the maths. It was a mistake in what the maths was applied to. Around four in five validity studies are concurrent (existing employees tested at a point in time), but the range-restriction corrections being used had been derived from predictive studies (applicants tested, then followed). Applying one to the other inflated the results.

Read the table below as evidence about how to assess, not as a leaderboard to pick a winner from. The highest figure in it is 0.42, which still leaves most of job performance unexplained by that method on its own. The useful findings are about the gap between structured and unstructured judgement, and about how little any single signal carries.

What this changes in practice

Structure is the whole game. A structured interview scores 0.42. The same conversation held without structure scores 0.19, and the bottom of its credibility interval is minus 0.01. An unstructured interview can predict nothing at all, and it is still the most common way people get hired.

No single method carries a hire. Even the strongest predictor in the table explains a minority of what happens after someone starts. That is an argument for combining evidence and knowing what each signal is worth, rather than searching for one method that settles it.

Years of experience barely predicts anything. At 0.07 it is close to useless as a screen, which is awkward given how many job ads still lead with it.

The old rankings are unreliable. Cognitive ability fell from 0.51 to 0.31, and to 0.23 when Griebe and colleagues restricted the analysis to 21st-century studies. Empirically keyed biodata and candidate interests went up. Anyone still designing a process around the 1998 order is working from numbers that have since been revised.

Learning and development statistics

Australian employers spent an average of A$1,122 per employee on training in 2025 and trained 38% of their workforce. Fifteen per cent of Australian organisations have no training budget at all. Spending intent has climbed for three years, from 37% of employers planning an increase in 2024 to 60% in 2026.

The training spend gap by employer size is wide: A$549 per employee at small employers against A$1,547 at large ones. And while 79% of Australian organisations have a training plan and 65% have a strategic workforce plan, only 20% say the two are integrated to any great extent. Joining capability data to workforce planning is what a learning platform should do, and mostly does not.

More money, less time

In the United States, spend per learner rose to US$874 from US$774, while training hours per employee fell from 47 to 40. Employers are paying more for less of people's time.

The largest capability gap in the region is AI itself. Across more than 7,000 Australian and New Zealand respondents, 60% use AI regularly at work and 78% have had no formal training in how to use it. Lack of training is the second largest barrier to adoption.

Australian apprenticeships

Non-trade apprenticeship and traineeship contracts in training have fallen 43.8% since 2021, from 130,730 to 73,475, while trade contracts fell only 3.6%. The trade shortage story gets reported constantly. The collapse of the non-trade half of the system does not.

HR technology and applicant tracking systems

HR technology budgets are tightening. The share of organisations planning to increase HR tech spend has fallen from a five-year high of 47% to fewer than one in three.

How many systems HR is actually running

On AI inside HR software specifically, the gap between what vendors ship and what customers use is stark. Awareness of embedded AI in the core HR system moved from 14% to 15% in a year, roughly a quarter of buyers do not know whether their system has AI at all, and only 4% of organisations report using AI agents in HR.

One more counterweight to the forecasts that AI is about to shrink HR. The HR-to-employee ratio has risen from 1.58 per 100 employees in 2017 to 1.98 in 2025, and HR's share of operating expense has doubled from 1.2% to 2.4%. HR has been growing, not shrinking. If you are weighing systems on the back of these numbers, our platform comparisons set out where each of the major vendors fits.

United Kingdom, United States and Asia Pacific

United Kingdom

• Vacancies fell to 707,000 in the May to July 2026 period, 10.3% below the pre-pandemic level and back to where they sat in late 2014.
• Private sector recruitment plans dropped to 57%, the lowest since the measure began in 2016.
• Despite that, 31% of employers still report hard-to-fill vacancies.
• Only 10% of UK employees are engaged at work, half the global average.
• 34% of UK employees use generative AI monthly, two-thirds of them untrained, and 45% distrust the output.

The Employment Rights Act 2025 is rolling out in phases, with the unfair dismissal qualifying period dropping to six months from 1 January 2027. Two credible sources disagree sharply about its effect: the government's own impact assessment puts the cost at around £1 billion a year and no more than a 0.1% increase in total employment costs, while CIPD's survey of 2,082 employers found 74% expecting costs to rise and 37% planning to cut permanent recruitment. Both are worth knowing.

United States

• Quits rate 2.0% and layoffs 1.1% in June 2026: a low-hire, low-fire market.
• Graduate hiring for the class of 2026 swung from -2.4% to +5.6% on a like-for-like median basis.
• 51% of workers who see their employer as ineffective say they are likely to leave within a year.

Asia Pacific

• Japan's job openings to applicants ratio sits at 1.18 with unemployment at 2.5%.
• Korea's graduate unemployment rose 18.8% year on year while unemployment among the least qualified fell 17.6%.
• Singapore recorded a 19th consecutive quarter of employment growth alongside retrenchments rising 18% on the quarter.
• New Zealand unemployment is 5.6% with underutilisation at 13.8%, and Māori underutilisation at 22.8%.

Quick reference: more verified numbers

Additional figures from the same research, grouped by topic. Each one carries the organisation that produced it so you can attribute it correctly.

Employee engagement and wellbeing

• Global employee engagement was 20% in 2025, down from a 2022-23 peak of 23% and the first back-to-back annual decline Gallup has recorded. (Gallup)
• Employee thriving globally sits at 34%, broadly flat year on year. (Gallup)
• Mercer's separate thriving measure fell from 66% in 2024 to 44% in 2026, its lowest since the measure began. The two are different instruments and should not be charted together. (Mercer)
• 50% of Australian workers experienced a lot of stress the previous day, against a global average of 40%. (Gallup)
• Australian mental stress workers' compensation claims take a median 36.8 weeks off work at a median cost of $69,900, against 9.2 weeks and $19,400 for body stressing claims. (Safe Work Australia)
• Employee concern about losing their job to AI rose from 28% in 2024 to 40% in 2026. (Mercer)
• 62% of employees say leaders underestimate AI's emotional impact, while only 19% of HR leaders factor those impacts into digital implementation. (Mercer)

Hiring, recruitment and candidate experience

• Applications handled per recruiter rose from 146 in 2022 to 746 in 2025. (Greenhouse)
• Offer acceptance rates fell from 74% in the second quarter of 2023 to 51% in the second quarter of 2025. (Gartner)
• Close to half of employers automatically screen out any CV showing an employment gap of more than six months. (Harvard Business School and Accenture)
• UK students and graduates using AI in their applications rose from 55% to 73% in a year, climbing to 66% among those applying for more than 100 roles. (Prospects at Jisc, n=4,990)
• The share of UK job postings disclosing pay fell from 64.9% to 55.3% in eight months, the lowest since December 2021.
• 91% of Australian organisations making redundancies were recruiting at the same time. (AHRI)
• Australian manager roles got easier to fill over the past year, with the fill rate up 2.7 points to 79.1%, while professional roles got harder, down 5.3 points to 66.5%. (Jobs and Skills Australia)

AI in HR and hiring

• 39% of organisations use AI anywhere in HR, and 18% use it broadly in hiring. (SHRM)
• 4% of organisations report using AI agents in HR. (Sapient Insights Group)
• 38% of candidates have withdrawn from a hiring process because it included an AI interview, and a further 12% say they would. (Greenhouse)
• 6% of 3,000 candidates surveyed admitted to interview fraud in the second quarter of 2025. (Gartner)
• The advertised wage premium for AI skills is 62%, up from 57%. (PwC AI Jobs Barometer)
• Australian job ads mentioning AI skills roughly doubled, from about 20,000 to 41,000. (SEEK)
• The employment gap for 22 to 25 year olds in AI-exposed US occupations widened from 15% to 19%. (Stanford Digital Economy Lab)
• Around 1 in 10 employees at AI-adopting organisations strongly agree AI has changed how their work gets done. (Gallup, n=23,717)

Learning and development

• US training spend per learner rose to US$874 from US$774, while hours per employee fell from 47 to 40. (Training magazine)
• Australian employers spent an average of A$1,122 per employee on training and trained 38% of their workforce. (AHRI)
• 15% of Australian organisations have no training budget at all. (AHRI)
• Training spend per employee runs A$549 at small Australian employers against A$1,547 at large ones. (AHRI)
• New Zealand workplace-based learners fell from a peak of 158,570 in 2022 to 108,315 in 2025.
• Asia Pacific leads the world on skills-based workforce planning at 41%, against North America's 22%, with a global average of 30%. (LinkedIn Talent Report, 2026)
• Only 12% of organisations are mature in skills-based workforce models. (i4cp)

HR technology and applicant tracking systems

• 97.8% of Fortune 500 companies had a detectable applicant tracking system in 2025, 489 of 500. (Jobscan)
• Workday holds 39% of Fortune 500 talent acquisition and SAP SuccessFactors 13.2%, so two vendors cover just over half the top of the market. (Jobscan)
• The global applicant tracking software market was US$2.5 billion in 2024, with the top ten vendors holding only 51.1% between them. (Apps Run The World)
• 43% of organisations now have a defined HR technology function, but 58% of the people in those roles have less than three years of experience in them. (Sapient Insights Group)
• People analytics teams average one practitioner per 2,500 employees, and 52% of organisations report measurable business improvement from the function. (Insight222, 372 organisations)
• Awareness of embedded AI in the core HR system moved from 14% to 15% in a year. (Sapient Insights Group)

Australia and New Zealand labour market

• Australian unemployment was 4.5% in July 2026. (ABS)
• Australian job vacancies stood at 329,500 in May 2026, well down from the May 2022 peak of 473,100. (ABS)
• The Wage Price Index rose 3.2% over the year to the June quarter of 2026. (ABS)
• 36% of employed Australians usually worked from home in August 2025, effectively unchanged on the year, rising to 59% among managers and professionals. (ABS)
• Australia's average total remuneration gender pay gap narrowed to 21.1%, while the CEO gap widened 1.2 points to 26.2%. (WGEA)
• Employer-sponsored skilled migration places rose from 44,000 to 58,040 for 2026-27, funded by cutting regional places from 33,000 to 14,110. (Department of Home Affairs)
• Australian employers expect to raise pay by 3.3% over the coming year. (AHRI)
• New Zealand unemployment was 5.6% in the June 2026 quarter, with underutilisation at 13.8% and Māori underutilisation at 22.8%. (Stats NZ)

Statistics that get repeated but do not hold up

Checking every figure on this page meant tracing a lot of famous numbers back to their source. Some of them do not have one. These are the ones we found circulating most widely without anything underneath them, and we have left them off the page above.

This matters more than it used to. Unsourced figures are now being ingested and repeated by AI assistants, which launders them one more time and puts them in front of people making real decisions.

How we checked these numbers

Every figure on this page was traced to the organisation that produced it, and read on that organisation's own page or in its own PDF. Where a number is quoted from secondary coverage because the original sits behind a paywall, we have said so at the point of use.

• Primary sources only. Statistical agencies, named research reports with disclosed samples, peer-reviewed literature and regulators. No statistics roundups.
• Samples and dates travel with the number. A figure without a population and a period is not evidence.
• Forecasts are labelled as forecasts. Market sizing and multi-year predictions are not measurements, and we do not present them as such.
• Figures we could not verify were left out, and the most widely repeated of them are listed below so you can avoid them too.

One example of why this matters. For the same learning management system market in the same base year, published estimates run from US$24.09 billion to US$31.42 billion depending on which firm you ask, with ten-year forecasts diverging by roughly half. Rather than pick one, we would suggest treating any single market-size number with care.

Sources

Every organisation whose data appears above, with a link to where it publishes it.

• Australian Bureau of Statistics. Job Mobility, February 2026; Working Arrangements; Labour Force. View source
• Jobs and Skills Australia. Survey of Employers who have Recently Advertised, March quarter 2026. View source
• Australian HR Institute. Quarterly Australian Work Outlook, March quarter 2026. View source
• Gallup. State of the Global Workplace 2026. View source
• Greenhouse. Recruiting benchmarks and AI in hiring research, 2025 and 2026. View source
• SHRM. State of the Workplace 2026, Talent Trends 2026, State of AI in HR 2026. View source
• Deloitte. 2026 Global Human Capital Trends. View source
• Mercer. Global Talent Trends 2026. View source
• Qualtrics. 2026 Employee Experience Trends. View source
• Sackett, Zhang, Berry and Lievens. Revisiting the design of selection systems, Industrial and Organizational Psychology 16(3), 2023. View source
• Harvard Business School and Accenture. Hidden Workers: Untapped Talent, 2021. View source
• Sapient Insights Group. Annual HR Systems Survey, 27th and 28th editions. View source
• World Economic Forum. Future of Jobs Report 2025. View source
• LinkedIn. 2026 LinkedIn Talent Report (the report that replaced the Workplace Learning Report). View source
• SEEK. Employment Report and AI job ad data, July 2026. View source
• Fosway Group. Digital Learning Realities 2026. View source
• NCVER. Apprentices and trainees, June 2026. View source
• Hays. Salary Guide FY26/27. View source
• Office for National Statistics. Labour market overview, August 2026. View source
• CIPD. Labour Market Outlook, Summer 2026. View source
• The Conference Board. Job Satisfaction 2026. View source
• Australian Public Service Commission. Principles for agency use of AI in recruitment, effective 1 June 2026. View source
• WGEA. Employer gender pay gaps, 2024-25. View source

Frequently asked questions

What are the most important HR statistics for 2026?

The four that change the most decisions: Australian job mobility has fallen four years running to 7.2%, global employee engagement is 20% with manager engagement down to 22%, Australian employers get 21.5 applicants per vacancy but judge only 5.2 suitable, and 89% of organisations using AI in HR report efficiency gains while only 24% report better identification of top candidates.

How many applicants does a job vacancy get in Australia?

Australian employers who recently advertised reported 21.5 applicants per vacancy in the March quarter of 2026. Of those, 8.3 were considered qualified and 5.2 suitable. The national fill rate was 68.2%, so close to a third of advertised vacancies went unfilled. Source: Jobs and Skills Australia.

What percentage of employees are engaged at work?

Gallup puts global engagement at 20% for 2025, the lowest since 2020. Australia and New Zealand sit at 21%, the United States at 31% and the United Kingdom at 10%. Other providers report much higher figures because they measure engagement differently, so always check which instrument a number comes from before comparing.

Is it true that 75% of CVs are rejected by an applicant tracking system?

No credible source supports it. Every version of the claim traces back to a resume optimisation vendor that promoted it around 2012 and closed in 2013 without publishing any study. The real 75% in this area is that around 75% of United States employers use an applicant tracking system, from Harvard Business School and Accenture research. It was never a rejection rate.

How often is this page updated?

It is reviewed and refreshed as major research releases land, and the last review date is shown at the top. Labour market figures change quarterly, so check the reference period on any statistic before using it.