I apply everywhere and hear nothing. What am I doing wrong?

On this page

Short version
Who it's for
What to do
Your options
Beliefs checked
Questions to ask

The short version

One application with no reply tells you almost nothing about you: only 8% of applicants got past the first screen (Gem, a company that sells hiring products, 2026 recruiting benchmarks, over 165 million applicants from June 2021 to May 2025).

A long run of silence does say something, though no tested number tells you when. Tailored resumes got about 6 interview chances per 100 applications and generic ones fewer than 3 (Huntr's Q2 2025 report, from outcomes its job-tracker users recorded). So by our reading, if 100 applications have brought no interview at all, it is time to change how you apply.

The famous claim that 75% of resumes are never seen by a human has no published research method behind it, according to The Interview Guys, a career-advice site. Screening is still real, and it catches qualified people too: in the 2021 Harvard-Accenture Hidden Workers study, as The Register reported it, 88% of employers agreed that qualified high-skills candidates get screened out for not matching the exact job-description criteria. So silence on its own is not proof that you are unqualified.

The evidence points to three changes worth trying, in no order it ranks: tailor each application to the posting's stated requirements, get referred where you can, and add a few related occupations to your search. Most numbers here come from vendors (companies that sell hiring or job-search products) and from surveys, not controlled studies, so treat them as rough guides.

Who this is for

You have sent a lot of applications, maybe for months, and almost nothing comes back. You may already have done what you were told, like tailoring your resume and asking around; if so, step 6 below is the place to start. You want to know what to change, and whether software is turning you down before a person reads your resume.

What to do this week (our advice)

The evidence doesn't rank these steps, so our advice is to start with step 6: your own count is the only thing here that tells you about your search rather than everyone's.

For the few roles you want most, look for a referral, ideally from someone who does the same kind of job. Referred candidates convert far better than people who apply on their own in vendor data, though they are an already-selected group. It may not need a close contact: moderately weak ties, people you share some mutual connections with, created the most job moves in LinkedIn experiments reported by Stanford's Digital Economy Lab. That study was about job moves in general, not referrals (option 3).
Answer every yes/no screening question truthfully and completely. All 25 recruiters Enhancv interviewed said they use these "knockout" questions when a posting has them.
Rewrite your resume for each posting, using its own required terms where they are true for you. Write it naturally, with no keyword stuffing and nothing hidden. This means sending fewer applications, and our advice is to make that trade, given the small effects of overall search effort in option 1.
Add a few related occupations to your search, and keep your main target. In the Belot, Kircher and Muller experiment, people who had searched narrowly got twice as many interviews as similar narrow searchers in the control group (option 4).
Make your resume easy to skim: clear sections, headings, short sentences and no images. Both a 2018 eye-tracking study by Ladders, as HR Dive reported it, and Workday's administrator guide point that way.
Keep a count: for each application, note how you sent it (one-click, company website, or referral) and whether anyone replied. Judge the results across many applications, not week by week, because silence is common even for qualified people and the published numbers on each route disagree.
Keep other applications going while you wait for any one answer, and if you doubt a posting is real, ask someone at the company.
If you have a gap in your work history, give the reason. In ResumeGo's 2019 field experiment, fictitious applicants who gave a reason for a gap got more callbacks than those who gave none.

Your options, and what each one costs

1. Send more applications, the same way

What happens. You keep your resume and your targets as they are and send more.

What it gives. A modest link to more interviews. How hard people searched predicted how many interviews, offers and jobs they got: corrected correlations of .23 for interviews, .14 for offers and .19 for being employed, on a scale where 0 means no link and 1 a perfect one (a 2021 review of many studies in the Journal of Applied Psychology). Searching hard here means effort in the whole search, not just the number of applications.

What it gives up. Those are small effects, and the same 2021 review found that searching harder did not predict the quality of the job people found.

What the evidence says. Some searches do take many applications. The most common path to an offer took 10 to 20 applications (20.8% of users), and 14.3% of users sent more than 100 applications before an offer (Huntr's tracker data for April to June 2025, as its users recorded it).

2. Tailor each application to the posting

What happens. You answer every screening question exactly, and rewrite your resume so the posting's required terms appear in plain, natural wording.

What it gives. Possibly a better place in the queue. That is our own inference: no source here tested whether tailoring raises a match score (a ranking of how well a resume fits the posting).

In Huntr's Q2 2025 report, users who tailored their resumes reached an interview or offer on 5.75% of applications, against 2.68% for those who didn't, by the outcomes they recorded themselves.

All 25 US recruiters Enhancv interviewed said they use knockout questions when a posting has them, which Enhancv defines as yes/no eligibility checks such as work authorization, a required license and location (Enhancv, a company that sells job-search products; undated, page read October 2026). That doesn't show how many postings use them or how many applicants they remove.

Enhancv's write-up also says many teams use match scores and filters to decide whom to review first, and that a low score is not an automatic rejection, but in high-volume roles it can push an applicant to the bottom of the pile.

What it gives up. Time on each application, so you send fewer.

Don't hide instructions in your resume: 22% of hiring managers said they had caught applicants hiding prompt injections (hidden instructions to AI screeners) in resumes (Greenhouse, a hiring platform, 2025 survey of 1,236 recruiters and hiring managers in the US, UK, Ireland and Germany).

What the evidence says. In Enhancv's interviews, 76% of the 25 recruiters said natural use of keywords, not keyword stuffing, is among the things they value most. Using AI to write your resume is a different question, with its own evidence: see Should I use AI to write my resume?.

3. Get referred

What happens. For the roles you want most, you find someone at the company who can refer you, instead of applying cold.

What it gives. Referred candidates do much better in vendor data, though part of that gap may be who gets referred, not the referral itself. Gem, which sells hiring products, says in its 2026 recruiting benchmarks that referrals convert at 11 times the rate of people who apply on their own. In Ashby's analysis of over 38 million applications from January 2021 to December 2024, as Huntr, a job-tracker company, reports it, about 40% of referred candidates reached an interview, against 3% of people who applied on their own.

What it gives up. You need contacts, and a referral is no guarantee. Huntr, a job-tracker company, cautions that referred candidates are a different, already-selected group. Gap: we hold no tested way to get a referral when you know no one at the company.

What the evidence says. Referrals are rare: in Ashby's analysis of over 38 million applications, referrals were 1% of applications on average. People applying on their own remained the largest source of hires, at 43% to 52% of all hires from 2021 to 2025 (Ashby, a company that sells hiring software, from nearly 250,000 hires).

Ashby's analysis of over 54 million applications to March 2026 says referrals more often lead to interviews when the person referring works in the same job function as the role. Moderately weak ties, measured by mutual connections, created the most job moves, and weak ties helped more in digital industries while strong ties helped more in less digital ones (randomized LinkedIn experiments on over 20 million people over 5 years, reported by Stanford's Digital Economy Lab).

Outside vendor data, Pallais and Sands (NBER, page read October 2026) start from the finding that referred workers are more likely to be hired, all else equal. In 3 field experiments in an online labor market, Pallais and Sands found that referrals carry positive information about performance and persistence that a worker's observable characteristics do not.

What happens. You keep your main target and add a few neighbouring kinds of job to your search.

What it gives. More interviews, in one experiment. Showing job seekers alternative occupations raised their total job interviews by 44%, and those who had searched narrowly got twice as many interviews as similar narrow searchers who were not shown them (randomized experiment in Edinburgh by Belot, Kircher and Muller, page read October 2026).

What it gives up. Some focus. An exploratory search led to lower-quality jobs 8 months later, while a search that was both focused and exploratory added to the number of job offers (a study that surveyed 248 unemployed people twice, "Job-search strategies and reemployment quality: the impact of career adaptability", page read October 2026). By our reading, keeping your main target is closer to the second.

What the evidence says. The Edinburgh experiment was small, and it did not clearly help people find a job. It had about 300 job seekers, in Scotland, not the US. The 44% rise started from a low base of about 0.09 interviews a week, as job seekers reported them. Its effect on finding a job was negative but too small to tell from chance, in samples What Works Growth calls quite small because many people dropped out.

What Works Growth also reports that the average wages of the jobs people interviewed for did not change significantly, and that job offers from other search activities rose significantly.

Two popular beliefs, checked

"Software throws out most resumes before a person sees them"

Not shown by the evidence we hold. The Interview Guys, a career-advice site, say the 75% figure came from a 2012 sales pitch by Preptel, a resume-optimization seller that closed in 2013, and that no research method was ever published to back it.

Of the 25 US recruiters Enhancv interviewed (undated; page read October 2026), 23 (92%) said their applicant tracking systems (ATS) do not auto-reject resumes for formatting, content or design, and 8% had AI or content-based auto-rejection set up.

A 2025 survey by Resume.org, which sells job-search products, of 1,399 full-time US workers who knew their company's hiring practices, reported by CPA Practice Advisor, points the other way. In that 2025 Resume.org survey, 57% of companies were said to use AI in hiring, and among those, three in four were said to let it reject candidates with no person reviewing them, 39% only at the first resume screen. Both come from companies that sell job-search products, asking different people (recruiters in one, workers in the other), so neither settles how much rejecting software does on its own.

63% of employers used a recruitment management system, and over 90% of those used it to make a first cut or rank applicants (the 2021 Hidden Workers study by Harvard Business School and Accenture, a survey of 2,275 executives in the US, UK and Germany, as Harvard Business School reports it).

So the likelier picture in this evidence is that software sorts and ranks, and knockout questions screen on eligibility, while how often software alone rejects a resume is unsettled.

"If nobody replies, I'm not qualified"

Not on its own. Two things can cause silence that has nothing to do with your qualifications: screening on exact criteria, and sheer volume. 94% of employers agreed that qualified middle-skills candidates get screened out for not matching exact job-description criteria, and 88% said so for high-skills candidates (the 2021 Hidden Workers study by Harvard Business School and Accenture, as The Register reported it). And volume has grown: Ashby, which sells hiring software, says in its 2026 recruiter productivity report that applications per hire tripled from 2021 to 2024 and stayed above 300 throughout 2025.

Questions to ask yourself

Are the postings real, or ghost jobs? Some may not be. Our advice is not to spend long trying to tell: there are no official statistics on how many there are, the estimates that exist are hard to assess, and we found no tested way to spot one. In any given quarter, 18% to 22% of jobs posted on Greenhouse's own platform are classified as ghost jobs (advertised with no intent to hire), by its own classification (Greenhouse, a hiring platform, 2024 State of Job Hunting report). An August 2026 Allwork.Space article reports that Greenhouse's figure of about 19% of jobs advertised on digital job boards in the second quarter counted postings that got at least one application but no hiring activity, such as interviews or job tests. No hiring activity is not proof that nobody meant to hire. Ashby, which sells hiring software, reports that among its customers, 82% of jobs open in 2024 led to a real hire and 18% were not filled. A Congressional Research Service report (April 2025) says there are no official statistics on the size of the ghost-job problem, and that the firms behind the estimates sell hiring-related services and have released varying information about their methods, which makes the estimates hard to assess. Gap: we hold no tested way for a job seeker to tell a ghost posting from a real one.
How long have I waited, and over how many applications? A month or two of silence is common, and by our reading many applications with no interviews at all is a reason to change something. 33% of professional candidates in North America said in 2023 that they were still waiting on next steps one to two months or more after applying, and 36% said so in 2024 (CandE Benchmark Research reported by ERE, from 170,000 responses overall in 2023 and 3,000 responses and counting in 2024). Gap: we hold no tested threshold for how many unanswered applications should make you change course.
Am I relying on one-click applying? Maybe, but no independent test says so. The two companies here that compare routes both put company websites ahead of one-click applying, but disagree widely on by how much. In a 2026 guide, LoopCV puts response rates at 1% to 2% for one-click applications and 3% to 5% for a company website, while Apt (2025), a seller, puts them at 3% to 13% and 15% to 30%. Apt (2025) says only about 3% of LinkedIn Easy Apply submissions (one-click applications) get a human review. AutoApplyMax reports, from what it calls a dataset on auto-apply outcomes, on 100,317 LinkedIn Easy Apply submissions from 2,412 active users in 40 countries (34% in the US) from January to August 2026. In that 2026 AutoApplyMax data, 52% were counted as reaching a human reviewer, meaning they got a reply or were measurably viewed on LinkedIn's view count, and the average recruiter reply rate was 4.7%. By our reading those two figures may measure different things, since a view count is not the same as the human review Apt describes. Gap: we hold no independent test comparing a one-click application with a company-site application for the same kind of job.
Has my search run long, or could my age be part of it? Possibly: field experiments have found callbacks falling with time out of work or with age, though they do not all agree. The chance of a callback fell the longer someone had been unemployed, with most of the drop in the first 8 months, though the findings speak most directly to younger job seekers with little work experience (a 2012 field experiment by Kroft and colleagues, which sent fictitious resumes to real job postings in 100 US cities). A 2016 audit study of experienced college-educated women applying for administrative support jobs found no link between callbacks and time unemployed, but among these women, those aged 50 and older were significantly less likely to get callbacks. ResumeGo, a resume-writing company, reports from a 2019 field experiment with fictitious applicants that those who gave a reason for a work gap got a 6.8% callback rate, against 4.3% for those who gave none. Gap: we hold no tested step that offsets age in callbacks.
Would my resume hold up on a quick read? Check it: the evidence suggests that both skimming readers and resume-reading software reward a plain layout. HR Dive reported that a 2018 eye-tracking study by Ladders, which sells job-search products, found recruiters skim a resume for an average of 7.4 seconds at the first screen. In that study as reported, simple layouts with clear sections and headings held attention, while cluttered layouts, multiple columns and long sentences did poorly. Workday, which sells hiring software, says in its administrator guide that resume parsing results (how well its software reads a resume) can vary with a resume's format and word order, and that resumes without images or image-based styles work best.

How we checked this: this page is the output of a rigorous research pipeline run with AI tools: every factual sentence rests on a finding checked against the source's own words, and a person approved the page before it went up. How we check our facts. Updated November 12, 2026.