Britain lost 19,000 vacancies in a year. Firms with under ten staff lost 18,000 of them.
A 2.7% fall reads like a market drifting gently downwards, and on its own it barely reads at all: it is smaller than the survey's own margin of error. Underneath it the movement is not gentle and not evenly spread. The smallest employers in the country - the ones with fewer than ten people on the payroll - cut their advertised vacancies by 16.1% in twelve months, while the rest of the market stayed close to flat. Every figure below is from a single Office for National Statistics bulletin published on 18 August 2026, linked at the foot of this piece.
16.1%annual fall in vacancies at businesses with 1 to 9 employees, to 95,000 - against 2.7% across all employers
2.7%annual fall in total UK vacancies, to 707,000 - a change smaller than the survey's own 95% interval of about plus or minus 32,000
95,000vacancies still advertised by businesses with 1 to 9 employees, roughly an eighth of the UK total
The average hides the whole story
There were 707,000 vacancies advertised in the United Kingdom in the three months to July 2026, 19,000 fewer than a year earlier. As a percentage that is 2.7%, which is the kind of number a reader skips. It deserves to be skipped for a second reason too: the survey that produces it carries a 95% confidence interval of roughly plus or minus 32,000, so a fall of 19,000 is not, on its own, distinguishable from no fall at all. The statisticians say as much, and advise readers to look at long-term movements rather than short-term changes.
What is not inside that fog is the composition. Vacancies at businesses with 1 to 9 employees fell by 18,000 over the same year - 16.1% of that band - leaving 95,000. Set the two together and the arithmetic is uncomfortable: the smallest employers in the country account for very nearly the whole of the national decline, and they did it from about an eighth of the national stock.
The band above them moved the same way and less far. Businesses with 10 to 49 employees were down 8,000 on the year, 7.7%. All three of the smallest bands fell over twelve months. The bulletin does not publish an annual figure for the two largest bands, so what happened above 49 employees has to be left as a question rather than answered by subtraction.
UK vacancies, three months to July 2026. The bands describe the EMPLOYER's headcount, not the number of posts it is advertising. The bulletin publishes a level and both changes for all employers and for the 1-to-9 band; for the 10-to-49 band it publishes the annual change only, so the other cells are left empty rather than estimated. The 95% confidence interval on the total is about plus or minus 32,000.
Employer size
Vacancies, May to July 2026
Change on the year
Change on the quarter
All employers
707,000
-19,000 (2.7%)
-6,000 (0.8%)
Businesses with 1 to 9 employees
95,000
-18,000 (16.1%)
-8,000 (7.8%)
Businesses with 10 to 49 employees
not published separately
-8,000 (7.7%)
not published separately
What a size band is, and what it is not
A size band here counts the people a business employs, not the jobs it is advertising. "Businesses with 1 to 9 employees" is a corner shop, a two-person agency, a small care provider - the employer, not the posting. The 95,000 figure is how many roles that entire category of firm was advertising across the country.
That distinction matters because it decides what the number can be used for. It cannot tell you that small firms hold most of the vacancies; they hold roughly one in eight. It tells you where the change happened. A market can shed jobs from its smallest employers while its largest ones hire steadily, and the national total will barely twitch - which appears to be close to what the last twelve months did.
It also cannot tell you why. A vacancy disappears when it is filled, when it is withdrawn, or when the business stops advertising it. The bulletin counts advertised posts; it does not ask what became of them. A 16.1% fall is consistent with small employers hiring successfully and with small employers giving up, and nothing in this release separates the two.
Why this desk is reporting it
The week's earlier editions described a market where few people are fired and few are hired. This is the same market seen from the employer's side, and it locates the contraction somewhere specific: not across the economy evenly, but concentrated in the firms least able to absorb it. Total vacancies are still 81,000 below where they stood in the first quarter of 2020, a gap of 10.3%.
This desk belongs to a company that builds a job app, and half of that app is a hiring tool for small employers, so we have an obvious interest in this number. That is a reason to be careful with it, not a reason to leave it out. What we can say is what the bulletin says: the decline is real in the small-employer band, it is large in proportional terms, and the national headline it sits inside is too small to be read on its own.
For anyone applying: a shrinking count of advertised posts at the smallest firms is not the same as those firms not hiring. It is a count of what is advertised. The roles that never reach an advert are exactly the ones a candidate cannot find by searching, and they do not appear in this series at all.
We publish how we test our matching for bias, including where it still falls short. Read the bias audit.