How Many People Actually Apply for One Job? Tracing the '250 Applicants' Statistic Against Three Modern Datasets

A reader sent me a spreadsheet in June. Four hundred and eleven rows, one per application, with columns for company, date, source, and outcome, colour-coded down to the shade of grey she used for 'no response.' At the bottom she had typed a small piece of arithmetic: 411 applications, times 250 competitors each, equals 102,750 people. Underneath it, in red, she had written four words: 'Is that actually true?'

It is a fair question and I could not answer it. I have spent years working with resume-parsing data, I have quoted the 250 figure myself, and when I sat down to check it I could not have told you where it came from. So I went looking. What I found was not a study. It was two blog posts, one of them more than a decade old, and a citation chain that loops back on itself until it disappears.

The trail ends at two blog posts, and neither one shows its work

The canonical wording is worth quoting exactly, because the precision is part of why it travels: 'On average, each corporate job opening attracts 250 résumés. Of these candidates, four to six will be called for an interview and only one will be offered a job.' That sentence, in that form, appears in a Glassdoor blog post titled 50 HR & Recruiting Stats That Make You Think, dated January 2015. Nearly every version circulating today, on career sites, in university career-centre handouts, and in the answers you get from AI assistants, is a descendant of that post.

The number itself is older. Its earliest printed appearance I can locate is a May 2013 column by Dr. John Sullivan in the recruiting trade publication ERE, Why You Can't Get A Job … Recruiting Explained By the Numbers, which states flatly that 'on average 250 resumes are received for each corporate job opening.' It sits in a list of figures assembled to make a rhetorical point about how brutal the funnel had become. It was persuasive writing, and it was never presented as original research.

Here is what is missing from both. There is no sample size. There is no year the data was collected. There is no definition of 'corporate job opening,' which matters enormously, because a requisition at a 200,000-person retailer and a requisition at a 40-person software company are not the same object. There is no industry breakdown, no geography, and no statement of whether 250 is a mean or a median. There is no link to a dataset. Ten years of career advice rests on a sentence with none of the apparatus that would let you check it.

Three of the largest hiring datasets in the world disagree by a factor of thirty

The good news is that the question is answerable now in a way it was not in 2015. Applicant tracking system vendors sit on enormous, unglamorous datasets and several of them publish aggregates. Three of those datasets are large enough to take seriously, and the interesting thing is not what any of them says. It is how violently they disagree.

Greenhouse's benchmark work, drawn from more than 640 million applications across over 6,000 companies between 2022 and 2025, reports that applications per job rose from roughly 115 in 2022 to 244 in 2025, an increase of 111%. (Greenhouse recruiting benchmarks) Read that sequence again with the 250 claim in mind. The earliest large-sample benchmark I can find puts the real figure at 115, less than half the number that had already been circulating as settled fact for seven years. The 250 did not describe 2015. It arrived at approximately the right answer roughly a decade after it was written, and by accident.

Ashby, whose customer base skews toward technology companies and high-growth startups, reports a higher figure: more than 300 applications for the average open role, up from roughly 100 in 2021, based on an analysis of more than 100 million applications across 200,000 jobs. (Ashby Talent Trends, reported by HR Dive) Two credible platforms, overlapping periods, a 25% gap between them. That gap alone should end the practice of quoting one universal number.

Then there is Workday, which is not a startup tool but the HR backbone of a large share of the Fortune 500. Workday published its own totals for calendar 2024: customers created 38 million jobs, candidates submitted 356 million applications, and customers created 28 million offers. (Workday, 'How HR Leaders Can Thrive in a Complicated Job Market') Nobody seems to have done the division in public, so here it is: 356 million divided by 38 million is about 9.4 applications per requisition. Not 250. Not 300. Nine.

Before anyone concludes that Workday has debunked everyone else, it has not. The gap is almost entirely definitional, and unpicking it is the most useful thing in this article. Workday counts every requisition its customers open, which includes hourly retail and hospitality roles, healthcare shift positions, internal-only postings, evergreen requisitions that stay open for a year, and single requisitions that hire fifteen people. Greenhouse and Ashby serve a narrower slice: mostly salaried, mostly professional, mostly posted publicly on job boards where any interested person can see them. Those are the postings that get flooded.

So all three numbers are probably correct about the thing they measure. They just measure different things, and none of them measures the thing you actually want to know, which is how many people applied for the specific job in the specific tab you have open right now.

An average is the wrong statistic for a question about your odds

Application volume per posting is a textbook heavy-tailed distribution. A remote, entry-level marketing coordinator role at a company whose logo you would recognise can pull four thousand applications in a week. A night-shift maintenance technician position in a town of 12,000 people can sit open for two months with eleven applicants, nine of whom are unqualified. Both are one job opening. An average computed across them describes neither.

This is the same failure mode that makes 'the average American household has 1.9 cars' useless for predicting what is in any particular driveway, and it has a specific consequence here. When a distribution is heavy-tailed, the mean sits far above the median, because a small number of enormous values drag it upward. If postings behave the way every application dataset I have worked with behaves, the median job posting receives dramatically fewer applications than the average job posting. Not one of the sources quoting '250' reports a median, which is a small tell about how much thought went into the number.

The practical translation is that the competition you face is not a property of the labour market. It is a property of the posting you chose. Visibility, remoteness, seniority, brand recognition, and how long the listing has been live explain far more of the variance than the year does. Two applications submitted on the same afternoon can face wildly different fields.

What actually changed between 2015 and now

Something real did change, and it is worth separating from the folklore. The number of people looking for work did not triple. The cost of submitting an application collapsed. LinkedIn reported that applications on its platform were running at roughly 11,000 per minute in 2025, up more than 45% year over year, a surge the company and outside observers attribute substantially to AI tools that fill and fire off applications in bulk. (CNBC, October 2025) The same job seekers, armed with better automation, now generate several times the volume they did two years ago.

On the other side of the pipe, capacity moved the opposite way. Greenhouse's benchmark data shows recruiters per organisation falling from 10.43 to 4.62 between 2022 and 2025, a 56% cut, while the applications each of those remaining recruiters had to handle rose by more than 400%. That is the actual structural story of the last three years, and it is far more useful than any single volume figure.

What the volume actually implies about your chances

The tempting move is to take 250 applicants and one hire and conclude you have a 0.4% chance. Several sites now publish exactly that figure. It is wrong, and not in a comforting-lie way, in a genuinely-bad-arithmetic way. That calculation assumes every applicant is an equivalent draw from the same urn, which would only be true if applications were assigned at random rather than submitted by people with wildly different backgrounds and wildly different degrees of fit.

They are not. LinkedIn's own research found that around 70% of hirers say fewer than half the applications they receive meet all the criteria stated in the posting. (Forbes, January 2025) In a field of 250, that means the population you are genuinely competing with is closer to a hundred, and the population that is competing with you on the specific requirements the hiring manager actually cares about is smaller still. Volume statistics count submissions. Hiring decisions are made over a much smaller set.

There is a second reason the naive odds calculation misleads, and I say this as someone who once ran the numbers the wrong way myself. Applications are not independent trials. The same resume submitted to forty roles fails or succeeds for correlated reasons: the same missing keyword, the same unexplained gap, the same summary written for a different job. Sending more applications with an unchanged document does not roll the dice more times. It rolls a loaded die more times. That is why our analysis of how long a job search actually takes found application counts to be such a poor predictor of outcome.

What to do with this, concretely

None of the above is a reason for despair, and it is definitely not a reason to stop applying. It is a reason to change what you optimise. If volume on the employer's side has quadrupled while reading capacity has halved, the scarce resource in the entire system is attention, and everything that follows is about buying attention cheaply.

1. 1Estimate the field before you apply, not after. LinkedIn shows an applicant count on most postings. Treat it as a rough signal: a role at 400 applicants after two days is a different proposition from the same role at 12 applicants after a week, and it should change how much time you spend on it.
2. 2Apply while the posting is young. Recruiters facing several hundred applications rarely read them in a fair, exhaustive order, and shortlists in practice get built before the queue empties. Being in the early portion of the pile is one of the few advantages available at zero cost.
3. 3Meet the stated criteria explicitly, in the posting's own words. If 70% of hirers say most applicants miss the listed requirements, then plainly satisfying those requirements on the page is not a minimum standard, it is a differentiator. Mirror the exact phrasing of the must-haves rather than a synonym you prefer.
4. 4Trade breadth for depth once you cross about 30 applications with no interviews. At that point the evidence is about your document, not about the market. Two hours spent rewriting the top third of your resume will outperform two hours spent submitting fifteen more copies of it.
5. 5Prefer postings where the field is structurally smaller. Roles that are onsite, in a specific city, at an unglamorous employer, in a specialised function, or advertised somewhere other than the largest job boards all attract a fraction of the volume, with no corresponding reduction in salary.
6. 6Attach a human to the application wherever you can. A referral or a note from someone who has actually seen you work moves you out of the volume pile entirely, which is the one intervention that survives contact with the research (see our review of what the evidence on referrals actually shows).

How to check a job-search statistic yourself in about five minutes

This is the fourth number we have traced back to its origin on this site, after the 75% ATS rejection claim, the six-second recruiter scan, and the hidden job market. The pattern repeats so reliably that it is worth handing you the method rather than only the results.

1. 1Search the exact sentence in quotation marks rather than the topic. Statistics travel as verbatim strings, so quoting the phrasing surfaces the chain of copies and, usually, the oldest one.
2. 2Follow every link until it stops. Career-site citations tend to point at other career sites. Keep clicking. If you reach a page with no outbound link for the claim, you have found the origin, and if that page is a blog post rather than a report, you have your answer.
3. 3Look for the four things a real finding always has: a sample size, a collection period, a definition of what was counted, and a statement of whether the figure is a mean or a median. A number missing all four is an assertion.
4. 4Ask whether anyone was in a position to know. A claim about all corporate job openings would require access to hiring data across the entire economy. Ask who plausibly had that access, and what they were selling at the time.

I sent the reader with the spreadsheet a shorter version of this. She wrote back that she was less worried about the 102,750 and more annoyed at herself for having believed it, which I thought was the wrong way round. Believing a specific-sounding number that everyone repeats is not a personal failing. Continuing to plan a job search around one after you know it has no source would be.