Cost per Hire in Japan: About 1 Million Yen, Seven Years Old

A hire in Japan costs about a million yen before the person does a day of work. Many foreign managers assume that money is the recruiter's fee, and that hiring without a recruiter is close to free. The best-known Japanese figures say otherwise: 1.03 million yen per mid-career hire and 0.94 million per new graduate, and as far as I can tell graduate hiring rarely runs through a recruiter. The catch is that these figures describe fiscal 2019, and I have not found a later edition of the same white paper that reports them.

Who is counted in the million-yen figure

Companies that recruit new graduates, reporting what they spent per hire on the new-graduate and mid-career hiring they did in fiscal 2019. The source is the Shushoku Hakusho 2020 (就職白書2020), a white paper from Recruit's Shushoku Mirai Research Institute. Average cost per hire was 936,000 yen for new graduates and 1,033,000 yen for mid-career hires. A year earlier the figures were 715,000 and 830,000. Both rose by about 200,000 yen in one year.

Three caveats come with it.

It is seven years old. Fiscal 2019 ended before the pandemic. It is the most recent edition of the white paper in which I could find a per-hire cost.

The sample is graduate recruiters. The white paper's company survey went by mail to 4,853 companies with five or more employees that run new-graduate hiring, and 1,256 answered, a 25.9% return, between December 2019 and January 2020. A small company that only hires mid-career is not in it.

It does not say what the money includes. The chart note reads only "companies that conducted new-graduate and mid-career hiring in fiscal 2019, numeric answers". Advertising and agency fees are surely in. Whether anyone counted the recruiting team's salaries or the interviewers' hours, the white paper does not say. So the figure cannot be added to numbers from other surveys.

What does a company spend on mid-career hiring in a year now?

An average of 6.099 million yen in 2025. That is a company's total for the year, from Mynavi's Mid-Career Hiring Survey, 2026 edition (中途採用状況調査2026年版). Mynavi surveyed 1,500 HR staff online in December 2025. All worked at companies with three or more employees, handled mid-career hiring in 2025 and were involved in managing the hiring budget.

The total was 4.843 million in 2021 and 6.506 million in 2024, then fell to 6.099 million in 2025. The average budget was 6.960 million, so companies spent about 860,000 yen less than they had planned.

The same survey reports 12.3 hires per company in 2025 against a target of 16.7. Divide 6.099 million by 12.3 and you get roughly 500,000 yen per hire. That is my division of one average by another. It is a rough order of magnitude, and an average of each company's own cost per hire would come out differently. The total also covers only money paid to outside parties: agencies, job ads, direct-sourcing services, job search engines, job fairs, employer branding and miscellaneous costs. In-house recruiters' pay is not in it.

By channel, agencies are the largest line at an average 3.755 million yen actually spent, then direct sourcing at 1.739 million and job ads at 1.251 million. Each channel average covers only the companies that used that channel, so they add up to more than the total.

One figure comes closer to a true unit cost. For hires made through ads on job-change sites, the ad spend per hire averaged 367,000 yen across all job types, against 370,000 the year before. For IT engineers it was 571,000, for AI engineers 599,000, for sales 404,000.

One hire cost about 1 million yen on average in fiscal 2019; job-site ad spend alone was 370,000 per hire in 2025

Cost per hire, in units of 10,000 yen; a full bar is 1.1 million yen. Top: Shushoku Hakusho 2020, companies with 5+ employees that hire new graduates, numeric answers. Bottom: Mynavi Mid-Career Hiring Survey 2026 edition, 2025 results, job-change site ad spend only

Average cost per hire　Shushoku Hakusho 2020

• Mid-careerFY2019
  ¥1.033M
• New graduateFY2019
  ¥0.936M
• Mid-careerFY2018
  ¥0.830M
• New graduateFY2018
  ¥0.715M

Job-site ad spend per hire　Mynavi, 2025

• AI engineers
  ¥0.599M
• IT engineers
  ¥0.571M
• Sales
  ¥0.404M
• All job types
  ¥0.367M

A million yen is a blend of channels. And it is already spent on the day the person joins.

The average hides a spread of several times between channels

A million yen is what you get when you blend every channel. No single channel costs a million.

For agencies, Mynavi's survey describes the service as one where a fee of around 30% of annual salary is charged when the hire is made. On a 5 million yen salary that is 1.5 million. Through job-site ads, the average above is 367,000. For hires through employees and acquaintances, neither survey gives a separate figure.

One number is easy to confuse with this one. The labor ministry's statistics on fee-charging placement agencies also show an average of about 1.03 million yen per placement in fiscal 2024, which I broke down in why Japan's recruiter fees run 30 to 35% of salary. That is the average fee per placement inside one channel. The white paper's 1.033 million is the average per hire across all channels. The two numbers look alike by coincidence.

Agencies handle a minority of job changers in Japan to begin with. The breakdown is in how people in Japan actually find jobs. Whether a company's cost per hire is high or low depends mostly on its channel mix.

What is sunk when a hire goes wrong?

The first thing sunk is that million yen. It was spent by the day the person joined, and it does not come back whether the hire works out or fails. Agency contracts usually carry a refund clause for early departures. Once the period passes, the fee is gone too.

Two more amounts follow. One is salary plus the employer's share of social insurance while the person is on payroll, which comes to roughly 1.2 times salary from the second year by the calculation in what one hire in Japan really costs. The other is hiring again, at the same unit cost.

A worked example. Someone on 5 million yen turns out to be the wrong fit after six months. Take the white paper's average of 1.03 million for the hiring cost. Six months of salary and social insurance is 5 million times 1.2, divided by two: 3 million. That is about 4 million so far. Hiring a replacement adds another million. This is an illustration and every company's numbers differ, but the order of magnitude holds. In this example, one wrong hire costs around 5 million yen.

From the employer's side, this explains a lot of behavior that outsiders call conservative. Japanese companies run many interview rounds, decide slowly and prefer people with experience. What I see is a company protecting money on the scale of that example. An extra interview round costs internal hours. A wrong call costs millions.

The cheapest protection comes before the start date. Say in the interview what the role has to deliver, what you will look at in the first three months, and what the company cannot offer yet. A candidate who walks away after hearing that has left at the cheapest possible moment. The other side of that, good calls where the person still does not come, is in reading Japan's offer decline rate.

Budgets were underspent, and average hires fell short of the average target

Two gaps in the Mynavi survey caught my eye. The average budget was 6.960 million yen and the average spend 6.099 million. The average target was 16.7 hires and the average result 12.3. More than a tenth of the money went unspent. On headcount, dividing one average by the other gives 73%. Mynavi does the same division and notes that it is not the average of each company's own achievement rate.

The same report carries a number that points the other way. In 2025, 92.9% of companies met their hiring target, the highest on record, up from 65.2% in 2024. The report suggests targets may have been lowered because hiring is hard. Putting the two together, my reading is that most companies set a smaller target and met it, and the shortfall sits with the few whose targets were large. That is my inference. The report does not say it.

So I will only go this far. On average the budget was not used up, and leftover money did not turn into more hires.

If I were setting a hiring budget for a Japan subsidiary, I would keep two lines. The first is money paid out, split by channel, with hires and cost per hire for each. The second is internal time: who interviewed, how many rounds, how long. Neither the white paper nor Mynavi reports that second line, and I know of no public survey that does. You have to keep it yourself.

My guess is that after six months of keeping both lines, you find the most expensive channel is not always the one whose hires stay longest. I have no public data for that. By then it will not matter much, because you will have stopped needing the national average.

---