How Many Jobs Should You Apply To?

Ten to fifteen well-targeted applications a week, not fifty, and the reason is arithmetic rather than principle. Once you subtract postings that are not real openings and apply a realistic interview rate, mass applying produces a worse expected outcome than a smaller, better-aimed effort — while consuming far more of your time and morale.

This post is the math, the evidence behind each input, and where the effort actually pays.

Start with how many postings are real

The first number most people never account for is that a meaningful share of listings are not attached to an opening.

Greenhouse's 2024 State of Job Hunting report, based on a survey of 2,500 workers plus data from its own applicant tracking platform, classified 18–22% of jobs posted through Greenhouse each quarter as ghost jobs. Close to 70% of its customers had posted at least one during Q2 2024.

A 2026 Clarify Capital analysis took a different approach: scraping 176,268 unique Indeed listings across 49 industries and every US state in February 2026, then flagging anything still active after 30 days. That method produced roughly 1 in 7 overall, with senior-level roles closer to 1 in 5 and wholesale above 50%.

The two studies disagree, and the disagreement is methodological rather than contradictory. Greenhouse can see whether a requisition ever produced a hire. Clarify Capital can only see how long a post stayed up, which catches slow hiring alongside fake hiring. Treating the honest range as 15% to 25% of listings, higher for senior roles is defensible; picking a single number is not.

Neither figure supports the "1 in 3 listings is fake" claim that circulates widely. That number comes from an employer survey in which roughly a third of companies admitted to having posted a job with no immediate intention to fill it, which is a statement about companies, not about listings. A single company posting one ghost job among forty real ones counts as a yes.

Then the response rate

Of the postings that are real, the conversion rate from cold application to interview sits at roughly 2–3%. That figure is consistent across enough independent sources to be treated as a working assumption, though almost none of them publish per-role methodology, so treat it as an order of magnitude rather than a precision estimate.

Run it forward. One hundred cold applications, with 20% of them attached to no real opening:

That is the mass-apply model at its own best case, and it is why people describe hundreds of applications before an offer. The model is not broken. It is just extremely inefficient, and every input in it is something you can move.

What moving each input costs

This is the part the volume framing obscures. Three levers, ranked by return on the hour.

Match quality. A 3% rate is the average across applications including ones where the candidate met half the requirements. Applying only where you match most of the requirements listed first raises the rate materially, and costs nothing but restraint. The single cheapest improvement available is not applying to roles you will not get.

Referrals. Referred candidates convert to hire at multiples of the cold-application rate. The published figures vary wildly by source and company size, which is a reason to distrust any specific multiplier, but every dataset points the same direction. One conversation with someone inside a company is worth more than ten applications to it, and it is the highest-leverage hour in a job search by a wide margin.

Tailoring. Twenty minutes of tailoring makes a resume visibly written for the role in front of the reader. At scale this is impossible, which is precisely why it works. When a large share of a pile is machine-broadcast and approximately relevant, a genuinely specific application is unusual.

Notice that all three levers require fewer applications to be viable. That is the actual argument against volume. Not that volume is lazy, but that it is the only strategy that forecloses the other three.

The realistic weekly number

For someone searching full time:

Fifteen tailored applications is around five hours of work at twenty minutes each, plus the time to find and read the postings. Add outreach and follow-up and the week is full. Anyone claiming fifty tailored applications a week is either not tailoring or not sleeping.

For someone searching while employed, halve it. Five to eight applications a week sustained over three months beats forty in one desperate weekend followed by six weeks of nothing, and the sustainability matters because searches now run long.

Spotting a ghost job before you spend the time

Not perfectly detectable, but the odds are readable.

Posting age. The strongest single signal, and the basis of the Clarify Capital methodology. A listing that has been up for 45 days on a role that would normally fill in three weeks is either stalled, ghost, or extremely picky. All three are bad odds.

Reposting cycles. The same requisition reappearing every 30 days with a fresh date is usually a listing being refreshed rather than a role being filled.

Evergreen language. "We are always looking for talented people" is a pipeline post, not an opening.

No named team or manager. Real requisitions usually have an owner who wants candidates, and that shows up as specificity about the team, the stack, and what the first ninety days look like.

Salary bands wider than about 40%. A range of $90k–$180k often means the level has not been decided, which means the requisition may not be approved.

Responsibilities that describe a department. When a posting lists the work of three roles, it is frequently aspirational headcount rather than budgeted headcount.

None of these is conclusive. Together they are enough to sort a list of twenty postings into eight worth twenty minutes each and twelve worth skipping.

Where to spend the time you save

Applying is the most visible job search activity and one of the least productive per hour. Reallocating is uncomfortable because outreach feels like it does not count.

Find the person, not the portal. For any role you genuinely want, identify someone on that team. A short, specific message referencing the work rather than asking for a referral outright converts at a rate no application form matches.

Make yourself findable. A portion of hiring starts with a recruiter searching rather than a candidate applying, which is a different funnel with different mechanics. That is a case for a LinkedIn profile built for recruiter search and, for technical and design roles, work that can be looked at.

Fix the base resume once. Every hour spent making the base resume genuinely strong is an hour that pays back on every application afterward. Most people tailor a weak base repeatedly instead of fixing it once.

Track what happens. Fifty applications with no record is fifty data points discarded. Which postings replied, how fast, from which channel — after a month that tells you where your time is converting, which no generic advice can.

On following up

Follow up with a person, once, about a week after applying. Two or three sentences: you applied, here is the single most relevant thing about your background, you are happy to answer questions. Send it to the hiring manager or someone on the team.

Status requests through an application portal go into a queue that is rarely read and never produces a decision. Repeated follow-ups produce a reputation rather than an interview.

If there is no reply after one follow-up, treat it as closed and move on. Employer ghosting is now the norm rather than a discourtesy — Greenhouse found 61% of job seekers reported being ghosted after interviews, up from 52% earlier in the same year. It says nothing about your application.

Keeping it sustainable

The failure mode of a long search is not applying too little. It is a burst of forty applications, three weeks of silence, and a conclusion that the market has rejected you personally, when the arithmetic above predicts exactly that silence.

Two things help. Set the target in applications sent rather than responses received, because the second is not under your control. And keep the per-application cost low enough that fifteen a week is not heroic, which mostly means having a strong base resume and a canonical record of your career to tailor from rather than rebuilding each time.

Plenty of tools support this. Teal and Huntr are built around tracking applications with per-role documents. VeriWorkly keeps a master profile that tailored resumes and cover letters derive from, free and open source, though with a smaller template library and less writing guidance than either. A spreadsheet and a folder of files also works. The structure is what saves the time.

The short version

Between 15% and 25% of postings are not real openings, and cold applications convert to interviews at around 2–3%. Those two numbers make the volume strategy mathematically weak and every alternative strategy mathematically strong. Ten to fifteen targeted applications a week, plus real outreach, plus a base resume worth tailoring, beats fifty untargeted applications on every measure including how long you can keep doing it.