One real week: 5,058 roles found, 773 scored, 249 worth it, 124 applications sent by hand. How many jobs to apply to per week, from the data.
Search "how many jobs should I apply to per week" and you get answers from 5 to 500. Most come with no numbers behind them. One person on Reddit swears by volume. A blog post says ten good ones beat a hundred quick ones.
I submit applications for a small group of people every week, so I have real numbers. Last week, 3 to 9 Nov 2026, I sent 124 applications for seven people. Here is what that week looked like, and what it says about the right number for you.
The 124 went to 87 different employers. That works out to about 18 applications per person for the week, though the split wasn't even. Some people had more open roles in their field than others.
| Day | Applications |
|---|---|
| Tue 3 Nov | 5 |
| Wed 4 Nov | 19 |
| Thu 5 Nov | 47 |
| Fri 6 Nov | 19 |
| Sat 7 Nov | 14 |
| Sun 8 Nov | 10 |
| Mon 9 Nov | 10 |
The big Thursday came mostly from one person's list. They had approved a long batch at once, and I worked through it in a day.
Every one of the 124 was read before it went. A tool filled the form, and I checked it and pressed submit myself. That limit, a person reading each one, is part of why the number is 124 and not 1,200.
88 of the 124 came from searches I ran for one person at a time. Only the rest came from my scanner's daily queue. That surprised me, and it's the most useful thing in this post.
My scanner reads company job boards every night for each person. Last week it found 5,058 roles for six of them. Most never reached the scoring step.
| Stage | Roles | What happened |
|---|---|---|
| Found by the scanner | 5,058 | Every role that matched a title and place |
| Set aside before scoring | 2,697 | Wrong city, wrong level, a repost, a title they ruled out |
| Scored against the CV | 773 | Each with a written reason |
| Scored 3 out of 5 or higher | 249 | Worth a closer look |
| Scored under 2.5 | 469 | Not a real fit |
So of 773 roles that looked right by title and place, about a third cleared the bar. Two thirds did not, once someone read them against the CV.
That ratio matters for your own search. If you apply to everything that matches your title, roughly two of every three may be weak fits. Those take the same evening to fill in.
A note on the 124: most of them didn't come from this pool. They came from the direct searches I describe next. I mention that so nobody reads this as "5,058 became 124".
Some search rules don't fit a filter. One person needs employers with a record of H-1B approvals. Another only takes remote roles whose posted pay starts above his floor, at companies he hasn't applied to yet.
A job board can't filter for either. So for those people I search the boards directly and check each role against their rule by hand. That's slower, but it's where most of last week's applications came from.
The lesson for you is simple. If your search has a rule like this, a visa, a pay floor, remote only, a filter will show you too much. You'll need to check each posting yourself, which caps how many you can do well in a week.
Before I fill anything, I open the posting and check three things:
If a posting fails any of these, it doesn't go, however good the title looks.
There isn't one right number, and anyone who gives you one without asking about your field is guessing. What I can share is what we see.
For the people I search for, about 18 a week per person is where careful applications land. That's with a strict fit bar, a tailored CV for each, and someone reading every form. In thin fields it's fewer, because there simply aren't more good roles.
| Your situation | What tends to set the number |
|---|---|
| Large field, few rules | How many strong fits appear each week, often more than you can do well |
| Narrow field or a visa rule | How many roles pass the rule, checked one by one |
| Very senior roles | Fewer roles, each with a note to the hiring manager |
These patterns come from our own work, not a study. The point is the shape. Your number should come from how many good-fit roles exist for you, not from a target you picked first.
If you want to know how many roles exist for your title right now, the free counter at How many jobs shows total, remote and senior counts.
Autofill saves typing, and that's fine. The risk is what autofill can't do. It doesn't read the posting, check the location or notice that the form asks about office days.
That's the split I use. A tool fills the fields, then a person reads every answer and the posting before submitting. Fully automatic tools skip the reading part, and that's where weak and wrong applications come from.
There's also the content. Some tools write cover letters and CV lines that claim things the person never did. When we measured our own tailored CVs in October, 16 of 18 drafts invented a tool, number, title or degree, or lifted words from the job description that weren't in the real CV. We now check every tailored CV against the original in code, and send the original when the check fails.
So for the question of autofill or manual, the honest answer is both. Let software type, and let a person decide.
If you apply to fewer, better-fit roles, you get evenings back. Spend some of them where a single application can't reach.
A hiring manager note works best when it's short. Two or three sentences: the role, one piece of your work that matches what they wrote, and a plain ask for a conversation. Send it after you apply, not instead of applying.
These take less time than ten rushed applications, and they compound. For more on why quick applications often go quiet, see why you're getting ghosted.
This post describes my week, and it's also what Zipply does. Each night it reads company job boards for you, scores each role against your CV with a written reason, and skips weak fits with a note on why. For the strong ones, it tailors your CV and fills the application, and I review and submit each one myself. You get one email each morning showing what went out.
It costs $50 a month, or $550 a year. It only works well when your field has enough open roles, so it checks that first and tells you plainly if the number is thin. Nobody can promise you a job, and we don't. If you'd like this run for your search, you can start here.
How many jobs should I apply to per week has a real answer, but it's yours, not a universal number. In one week we sent 124 careful applications for seven people, about 18 each, and only a third of the roles we scored were worth sending.
Count how many good-fit roles exist for you. Apply to those with care, and spend the time you save on notes and introductions.
It depends on how many good-fit roles exist in your field. For the people I search for, careful applications averaged about 18 a week per person in one recent week, fewer in narrow fields. Count the real openings for your title first, then apply to the strong ones with care.
It's best to do both. Let autofill handle the typing, then read every answer and the posting yourself before you submit. Autofill can't tell when a remote job is hybrid, when a question needs your own answer, or when the role is the wrong level. A person has to check those.
To apply manually means you fill in and submit the employer's application form yourself, rather than letting a tool submit it for you. You can still use autofill for typing. The key part is that you read the form and the posting, and you press submit.
Put this into practice
Your personal job search concierge. Zipply watches the market, scores every role against your CV, and applies on your behalf - only when the fit is right.
Get started →