Seven in ten job adverts never reach ten applicants
Every conversation about hiring starts from the same assumption: that the problem is volume. Too many applications, too little time, a shortlist buried under unsuitable CVs. Candidates are told to apply within the hour because the
queue forms fast.
We were able to check, because Reed publishes a live application count on every advert. A collector running once an hour turns that number into a growth curve. From 18 to 31 July 2026 we took 168 hourly snapshots across ten UK occupations
and kept the 1,691 adverts that appeared during the window, so every one is watched from the hour it went live rather than joined partway through.
Of those 1,691 adverts, 524 ever reached ten applications. The other 1,167 did not. Seven in ten of the vacancies we watched never got ten people to apply, at any point in a fortnight.
The queue builds far slower than anyone plans for
Where a queue did build, the tenth application arrived after a median of 28 hours. A quarter of those adverts took more than three days to get there.
The median advert had gained one application by the time it was a day old, and three by the time it was three days old.
The shape matters more than the total. For adverts that gained a real queue and were tracked for at least four days, a median of 24 per cent of their applications arrived in the first 24 hours. In every one of the nine occupations where we had
enough data to measure it, that first-day share sat between 17 and 29 per cent. Not one of them front-loaded a majority of its applicants into day one.
Three quarters of the eventual applicants had not arrived by the end of the first day.
A vacancy sitting at two applications on the morning of day two is not a failing advert. It is a normal one. That has a direct operational consequence, because the decision to refresh, repost or rewrite is routinely taken on day two, before the data that would justify it exists.
The occupation decides almost everything
The averages hide a spread that dwarfs them.
Receptionist adverts gained a median of 15 applications over the window, and 144 of 247 reached ten, typically inside a day. Staff nurse adverts gained a median of nothing at all. Three of 187 reached ten applications. Same country, same
fortnight, same job board. Payroll administrator sat between the two, with 196 of 448 adverts reaching ten and a median of 26 hours to get there. Occupational therapist behaved like nursing: three of 80. We measured the same gap in a much larger cross-section last month, where the median receptionist advert carried 45 applications and the median nurse advert carried 2. The hourly panel adds the half of that picture nobody usually has, which is that the gap is already visible on day one rather than accumulating slowly over the life of the advert. A nursing vacancy is not quiet because it is new. It is quiet because it is a nursing vacancy.
Why this costs HR teams money
Three habits follow from the volume assumption, and the data undermines all of them.
The first is judging an advert on its first day. In these occupations the first day carries a fifth to a quarter of the eventual applicants, so a quiet Monday tells you very little about the vacancy and quite a lot about the calendar. Teams that rewrite on day two are rewriting adverts that were working.
The second is reposting. Reposting exists to catch a burst of attention, and in the slow occupations there is no burst to catch. When 184 of 187 nursing adverts never see ten applications, the constraint is candidate supply, not advert visibility, and the spend buys a fresh timestamp and little else.
The third is the shortlist target itself. A hiring manager expecting ten applicants before interviewing has set a threshold that seven in ten adverts in this panel never crossed. That threshold does not filter for quality. It just delays the process until the advert has aged enough to look like a failure, at which point the vacancy gets escalated, rewritten, or handed to an agency, all of which cost more than interviewing the four people who did apply.
There is a fairness dimension too. The occupations that fill fast in this data are the administrative and front-of-house roles. The ones that do not are nursing and occupational therapy, where the shortage is already well documented. A process tuned to the first kind of vacancy will consistently misread the second.
Three things to change
Set the review point by occupation, not by calendar. If a receptionist vacancy has two applicants after 48 hours, something is wrong with it. If a nursing vacancy has two applicants after 48 hours, that is the market, and it will still be the market after a rewrite. One review rule applied across a mixed vacancy list will be wrong for most of it.
Stop treating ten applicants as the point where hiring begins. Decide in advance how few candidates you are willing to interview, and hold to it. In seven of these ten occupations the realistic answer is a handful.
Spend the reposting budget where there is attention to catch. In fast-filling roles the first day is genuinely worth watching and a repost can work. In the slow ones the same money goes further on reaching people who are not currently
looking at job boards at all.
What this does not say
The ten occupations were chosen because they are small enough to enumerate completely, so this is a panel rather than a cross-section of the UK market. The window is 13 days, so adverts appearing late were watched for less time,
which is why the age-anchored figures rather than the totals are the ones to quote. The cadence was hourly by design and closer to one snapshot every 90 minutes in practice, so times are accurate to a couple of hours rather than to the minute. And Reed can only count applications made through Reed, so any application made on an employer’s own site is invisible to this method.
The full analysis, with the per-occupation timings, is at https://sausagedog.io/blog/how-fast-job-adverts-fill-up
Anthony Clasper is the founder of Sausage Dog, a UK CV tool built on original analysis of 16,000+ UK job adverts. https://sausagedog.io