Organizations & Hiring
In-House Recruiters or Agencies in Japan? Split by Role
MHLW Survey on Job Changers 2025
Nobu Sato7 min read
Contents6
"Should we build an in-house recruiting team in Japan or keep using agencies?" is the wrong question, because the answer is different for each role. Sort roles by how scarce the talent is, how many you hire a year and whether the search can be public. Most roles then fall clearly to one side, and the hard cases are easy to spot.
How often does your Japan entity actually hire?
Count this before you budget a team. At small sites, mid-career hiring tends to come in bursts, with long quiet stretches in between.
Japan's Ministry of Health, Labour and Welfare (MHLW) published its 2025 Survey on Job Changers on 25 September 2026. It counts full-time workers on open-ended or one-year-plus contracts who joined from another employer between 1 October 2024 and 30 September 2025. Of establishments with full-time staff, 38.2% hired at least one such person in that year.
Size changes everything. At establishments with 1,000 or more employees the share was 94.6%. At 300 to 999 it was 86.1%, at 100 to 299 it was 77.7%, at 30 to 99 it was 60.3%, and at 5 to 29 it was 32.7%.
These are establishments, not companies, but for a Japan entity with one office the two are the same thing. Even at 30 to 99 people, about four in ten sites hired no one mid-career all year. A full-time recruiter costs the same every month. Hiring at that size doesn't arrive every month.
Demand is rising, though. In the same survey, 63.6% of establishments plan to hire job changers in the next three years, including 90.5% of the largest and 61.2% of the smallest. Of those with plans, 40.0% said they would prioritize experienced hires, against 9.8% for new graduates. The mid-career market in Japan is going to get more crowded.
Two kinds of role
The survey also asked why employers hired job changers, by occupation. The answers split roles into two groups.
The first is backfill. For clerical roles, "to replace someone who left" was the reason in 51.8% of cases. Sales 54.6%, service 60.8%, production, transport, construction and similar work 59.4%. Someone leaves, you refill the seat, and the same opening comes back several times a year.
The second is ready-made expertise. For professional and technical roles, "can contribute right away using their experience" was cited 64.1% of the time and "has specialist knowledge and skills" 51.2%. For managerial roles the figures were 60.7% and 50.2%. Each search is a little different, and the pool is thin.
Backfill roles suit an in-house team with job ads and employee referrals. The more often you run the same search, the better the process gets and the lower the cost per hire. Expertise roles, especially ones where you have to approach people who aren't looking, usually go faster with outside help.
Confidentiality cuts across both. Replacing a sitting leader, or hiring the head of a business you haven't announced, can't be posted, and an internal recruiter making calls will be noticed inside the company. Those go outside.
Hire high-volume standard roles in-house and send scarce ones out; the quadrant people get wrong is high volume and scarce
Who should run the search, by talent scarcity and hires per year. Hiring-reason shares from MHLW's 2025 Survey on Job Changers (establishment survey, by occupation, multiple answers); the allocation is the author's view
Across: scarcity, low to high | Down: hires per year, many at top, few at bottom | Confidential searches go bottom right wherever they would otherwise sit
In-house team plus ads and referrals
Clerical, sales, service. Backfill is 50 to 60% of the hiring reasons, and the process can be standardized
RPO or an exclusive agency
Several engineers of the same type a year. Many agencies at once means duplicate submissions; in-house alone can't reach them
Handled in-house part time
An occasional standard backfill. A job board and personal networks are enough; no team needed
Outside search firm
Managers, specialists, confidential roles. Ready expertise is the main reason, and the right people usually aren't looking
Don't decide in-house or outside for the whole company. Place each role in a quadrant, then pick who runs it.
What about roles that are scarce and high volume?
Give them to one outside partner with a reason to commit. That means an RPO or an exclusive agency, not a crowd of contingency firms.
This is the quadrant foreign companies in Japan hit most often and place worst. Five or six backend engineers a year. A sales team of ten in one go.
The usual move is to brief five or six agencies and pay whoever lands someone first. The same candidate arrives from three of them, the candidate decides you are disorganized, and each agency, seeing low odds, puts in little effort. The opposite move, doing it all in-house, leaves one recruiter juggling dozens of openings with no time to build a talent map.
RPO (recruitment process outsourcing) means paying monthly or per project for people who work inside your process, sourcing, contacting and screening. An exclusive mandate means picking one or two agencies, giving them a fixed window to themselves and getting a dedicated consultant in return. Either way the outside party has a reason to spend time on you.
What the in-house team has to be good at
Posting jobs and booking interviews is admin. The job is more than that.
The core skill is writing the profile. What the role actually screens for, which requirements can flex, how long each takes to learn on the job. A team that can't write that becomes a forwarding address for agency briefs.
Then mapping: knowing which companies employ the people you want and roughly how many. Even if an agency does the approaching, a map in your own hands tells you whether the person they sent is near the top of that pool or just available.
Then managing interviewers. My impression is that candidates in Japan read silence badly, and a week without a decision often means they accept somewhere else. Getting hiring managers to decide on time is a job no agency can do for you.
What to agree with an agency up front
Write down how long the exclusive window runs and what happens if it ends without a hire. Cap submissions per week and ask for accuracy over volume. Commit to replying within a set number of working days after a submission, and to a decision within a set number of days after each interview. That last one binds you more than them. When these partnerships stall, it is often the hiring side that has gone quiet.
On cost, agency fees in Japan are usually quoted around a third of theoretical annual income, and the first-year cost of a hire comes to about 1.5 times salary once social insurance is included, as I worked out in What one hire in Japan really costs: 1.2x to 1.5x salary. Agencies also place fewer than one in ten of Japan's job changers, which is in How People in Japan Actually Find Jobs: Recruiters Are 9%. Put those two facts side by side and it's clear the agency line should be kept for a few seats.
Where a new Japan entity should start
In the first year or two, I don't think a dedicated recruiter is needed. One HR generalist who knows hiring, plus one or two outside partners you use consistently, will cover it. Once you are hiring several people of the same type each year, bring that role family in-house or hand it to an RPO as a block. The order of the first few hires is in Building a Japan entity from zero, and a recruiting team rarely comes first.
The quickest test I know: take last year's hires and drop each one into the figure above. Some of the agency fees will turn out to be money well spent. Others will turn out to be the cost of nobody deciding.
Sources (as of November 2026)