The number: about one hire in three
The authoritative source is the Ministry of Health, Labour and Welfare's Employment Trends Survey (雇用動向調査), which asks every new hire at establishments with five or more employees how they got the job. For non-new-graduate entrants, the entry route (入職経路) breaks down like this.
- Advertising and job media (広告) — 33.2%
- Personal connections and referrals (縁故) — 23.5%
- Hello Work, the public employment service (ハローワーク) — 19.2%
- Other — 16.7%
- Private placement agencies (民営職業紹介所) — 6.3%
- Schools (学校) — 1.0%
So: about a third. Say that in a room full of Japanese HR directors and you will not be wrong.
Two things about the number are more interesting than the number itself.
First, Japan is more board-dependent than the West, not less. US source-of-hire surveys put job boards at roughly a quarter of hires, with referrals doing the heavy lifting at well-run companies. Japan's 33.2% is higher. The cliché that Japan hires through relationships is half right — referral hiring (縁故) at 23.5% is enormous by Western standards — but advertising still out-fills it.
Second, the average misleads. That 33.2% is every hire at every employer with five or more staff: the convenience-store shift leader, the factory line hire, the dispatch conversion, the regional sales rep. Job boards dominate high-volume, Japanese-language, locally mobile hiring, and they are extraordinarily good at it. The distribution inverts as you climb.
Where the cliff is
The tell is the private placement agency (民営職業紹介所) line. Agencies fill 6.3% of non-new-graduate hires nationally — a rounding error. But that 6.3% is not spread evenly. It was 2.1% in 2009 and has tripled since, and it concentrates in large employers, specialist functions and senior roles.
Meanwhile the white-collar placement market grew 12.0% in FY2024 alone, to roughly ¥449 billion, with Yano Research attributing the growth specifically to demand for management-level and high-specialist talent. There are now over 30,500 licensed placement agencies in Japan.
An industry does not grow 12% a year charging 30–35% of theoretical annual salary (理論年収) if job boards are filling those seats.
The same story shows up inside the boards' own products. Japan's senior-hiring platforms — BizReach, Recruit Direct Scout, doda X — do not sell job postings to be applied to. They sell access to a candidate database so that headhunters and in-house recruiters can go out and approach people. BizReach alone has around 8,400 registered headhunters working its 2.81 million members. The platform's own economics concede the point: at the top of the market, nobody applies. Somebody gets approached.
So the practical translation runs in two directions. If you are hiring a bilingual section manager (課長) or above, or any role where the intersection of language, function and industry produces fewer than a few hundred qualified people nationally, budget for a job board as brand spend and inbound optionality — not as your fill channel. If you are hiring ten Japanese-language inside sales reps in Osaka, the board is your fill channel and an agency is an expensive mistake.
The ten that matter: the volume tier
Prices below are standard four-week posting rates before agency discounts. Treat them as order-of-magnitude.
1. Indeed Japan, and Indeed PLUS
What it is. The aggregator that became the distribution layer. Over 35 million monthly users — the single largest audience in Japanese recruiting. Since January 2024, Recruit has routed its own media through Indeed PLUS, a distribution platform that automatically places one posting across up to ten job sites including Rikunabi NEXT, Townwork, Torabayu and Hatarakuiku.
Strengths. Unmatched reach. Pay-per-click, so spend tracks interest. Free organic indexing if your careers page is structured properly.
Weaknesses. Volume without discrimination — you will screen a lot. Weak at senior and bilingual. Indeed PLUS moved pricing and placement control into an algorithm, which several employers experience as a loss of levers.
Use it for. Volume, hourly, regional, and as the free backstop under everything else.
2. Rikunabi NEXT (リクナビNEXT — Recruit)
What it is. Recruit's mid-career flagship and the largest single database in the market: roughly 13 million registered members, about a fifth of the labour force. Now distributed through Indeed PLUS.
Strengths. Brand recognition that means an unknown foreign employer still gets seen. Broadest demographic coverage of any single site.
Weaknesses. Breadth is the problem — heavy applicant volume, heavy screening load, low signal density. ¥200,000–¥1.2M per four weeks.
Use it for. Generalist corporate roles where you need a big funnel and have the headcount to sift it.
3. doda (Persol Career)
What it is. The most useful hybrid in the market: a job board, a placement agency and a scout platform (doda X) running off one candidate pool. Around 9.9 million registered members as of mid-2025, passing 10 million by late 2025.
Strengths. Genuinely strong for experienced mid-career professionals. You can start on the board and escalate the same requisition into the agency arm.
Weaknesses. That escalation path is also a conflict of interest — the same company decides whether your ¥600,000 posting or their ¥3M placement fee gets the candidate. Posting rates ¥250,000–¥850,000 per four weeks, at the pricier end.
Use it for. Experienced professional hiring in Japanese: sales, engineering, corporate functions.
4. Mynavi Tenshoku (マイナビ転職)
What it is. Mynavi's mid-career site, roughly 9.0 million members as of September 2025. Strongest regional footprint of the big four, plus a large offline job-fair operation.
Strengths. Reaches outside Tokyo and Osaka better than anyone else. Young-skewing. Regional SMEs trust it.
Weaknesses. Thin above mid-level. Weak for bilingual. Premium tiers run to ¥1.5M per four weeks.
Use it for. Plant, branch and regional-office hiring; anything outside the Kanto–Kansai corridor.
5. en Tenshoku (エン転職 — en-japan)
What it is. The largest member base of the mid-career boards at roughly 12.2 million, and the most editorially distinctive. en sends its own reporters to interview the employer and writes the ad — including a mandatory section on what is difficult about the job.
Strengths. That honesty section is not a gimmick; it reduces early attrition, which is a real cost in Japan. Over 60% of members are 34 or under.
Weaknesses. Editorial control sits with en, not you. Long lead time to publish. The young skew makes it the wrong tool above manager level. ¥280,000–¥1.2M per four weeks.
Use it for. Junior-to-mid hiring where retention (定着) matters more than speed.
The ten that matter: the senior and specialist tier
6. BizReach (ビズリーチ — Visional)
What it is. The platform that created Japan's senior-hire scout market. About 2.81 million members, around 8,400 registered headhunters, and more than half its postings above ¥10M. Candidates pay a subscription for full access — which is itself the quality filter.
Strengths. The only database where you can reliably find employed, senior, non-actively-looking Japanese professionals. Direct-recruiting (ダイレクトリクルーティング) functionality is mature.
Weaknesses. It is a hunting licence, not a hiring service — it consumes recruiter hours, and most employers underestimate how many. Response rates to generic scouts are poor. Expensive on an annual-licence basis.
Use it for. Director and above, specialist functions, and any search where you will do the outreach yourself.
7. Recruit Direct Scout and doda X
What they are. Recruit's and Persol's answers to BizReach. Recruit Direct Scout carries roughly 10,000 registered headhunters; doda X around 7,300 headhunters and over 400,000 members.
Strengths. Free to candidates, so the pool skews slightly broader and less self-selected than BizReach. Good for reaching people who would never pay for a job site.
Weaknesses. Lower engagement per member. Heavy headhunter traffic means your scout competes with dozens of others in the same inbox.
Use it for. A second net under a BizReach-led direct-recruiting programme, not a first choice.
8. Green, Findy and Wantedly — the tech and startup cluster
What they are. Three different bets on engineering hiring. Green (Atrae) is success-fee only — roughly ¥300,000–¥900,000 per hire depending on location, with no posting cost. Findy scores candidates from their actual GitHub activity: around 267,000 cumulative registrations and over 4,000 client companies. Wantedly is a mission-and-culture platform with over a million users and 30,000 companies, where salary is deliberately not displayed and the call to action is a casual chat, not an application.
Strengths. Green's pricing is the best risk transfer in Japanese recruiting. Findy's signal on engineering ability is real. Wantedly reaches people who are not looking.
Weaknesses. All three skew young, Tokyo, startup. Wantedly's no-salary convention wastes time if your band is fixed and modest. None works above engineering-manager level.
Use it for. Engineers, product and early-stage teams. Green first if you want to pay only for outcomes.
9. type and Onna no Tenshoku type (女の転職type — Career Design Center)
What they are. A Tokyo-metropolitan IT and engineering board, and its sister site dedicated to women's mid-career hiring — a category with no real Western equivalent.
Strengths. Tight, high-intent Kanto audience. Onna no Tenshoku type is the most efficient single channel in Japan for hiring experienced women into roles with defined hours, and it reaches candidates the generalist boards do not.
Weaknesses. Small outside Kanto. Limited above mid-level.
Use it for. Tokyo technical hiring, and gender-balance objectives where you need a channel, not a policy statement.
10. The bilingual layer: CareerCross, Daijob and LinkedIn
What they are. The English-first tier. CareerCross and Daijob (around 710,000 bilingual users) are the two established bilingual boards. LinkedIn had about 5.08 million Japanese members as of May 2025 — 4.1% of the population.
Strengths. The only inventory where the job ad can be in English and the candidate expects it to be. LinkedIn is genuinely useful for employer brand and for outbound approaches.
Weaknesses. Read that LinkedIn number again: 5.08 million, against 13 million on Rikunabi NEXT alone. LinkedIn in Japan is an outbound tool and a shop window, not an inbound channel. The bilingual boards are small and their pools are worked hard by every agency in Tokyo — you will see the same 200 CVs everyone else sees.
Use it for. Outbound and brand. Never as your only bilingual channel.
Also on the board, but not boards
Hello Work (ハローワーク) is free, government-run, and fills 19.2% of non-new-graduate hires — more than double what agencies do. It is close to useless for bilingual professional hiring and close to essential for regional and hourly hiring.
Kyujin Box (求人ボックス), run by Kakaku.com, has around 15.6 million monthly users; Stanby (スタンバイ) has roughly 10 million. Both are domestic search engines that sit alongside Indeed. Cheap, and worth feeding.
The new-graduate market runs on entirely separate infrastructure — Mynavi, with 594,068 registered students as of 1 March 2025, and Rikunabi with 430,398, on a government-blessed hiring calendar. If you want new graduates, none of the ten boards above is relevant.
How Japanese boards actually differ from Western ones
Not in degree. In structure.
- The database is the product; the posting is the by-product. In the West you buy an advertisement and wait. In Japan the premium platforms sell you the right to go hunting — scout functionality is standard, not an upsell. Budget recruiter hours accordingly, because a BizReach licence with nobody working it is a pure loss.
- Flat-rate, time-boxed, paid in advance. The big four still sell four-week windows at ¥200,000 to ¥1.5M, with placement tiers determining how prominently you appear. There is no refund for zero hires. Western boards drifted to cost-per-click and subscription years ago; Japan's premium media are still selling display real estate, and Indeed PLUS is only now eroding it.
- Every major board owns an agency. Recruit, Persol, Mynavi and en all run a job board and a 30–35% placement business off the same candidate pool. LinkedIn does not own a search firm. Assume the conflict exists and price for it.
- Somebody else writes your job ad. Recruit, Mynavi and en send copywriters and photographers. Ads are long, narrative and image-heavy. en will publish the drawbacks of the job whether you like it or not. If you arrive expecting to paste in a global job description, you will not recognise what goes live.
- The salary on the ad is total annual cash (年収), not base. Japanese postings quote expected annual income (想定年収), which includes bonus. A foreign headquarters reading a ¥12M posting as base pay is one of the most common and most expensive misreadings in this market.
- Candidates are hiding, by design. Where being seen looking can still cost you standing with your current employer, the winning product feature is concealment: block-company lists, undisclosed postings (非公開求人), anonymised profiles. This is the single best explanation for why LinkedIn — a product built on public professional identity — has 5 million Japanese users while a site that lets you hide has 13 million.
- There are two separate hiring universes. Simultaneous mass new-graduate hiring (新卒一括採用) and mid-career hiring run on different platforms, different calendars, different pricing and different social rules. No Western market has this split.
- Boards are regulated employment businesses. The October 2022 amendment to the Employment Security Act (職業安定法) brought job media under a notification regime for specified job-information providers (特定募集情報等提供事業), with statutory obligations to keep listings accurate and current and to disclose the purpose of personal-data collection. Your posting is a regulated communication, not marketing copy.
The strategy
- Segment by seniority before you spend a yen. Below manager, in Japanese, locally mobile: boards fill the role, and paying 35% is a self-inflicted wound. Manager and above, bilingual, or narrow specialist: boards will not fill it, and a posting is brand spend. Draw that line explicitly in your recruiting plan and defend it against the vendor who tells you otherwise.
- Run one paid board, not four. Split budget across Rikunabi NEXT, doda, Mynavi and en and you buy four half-strength placements and four times the screening. Pick the one that matches your segment — Rikunabi NEXT for reach, doda for experienced professionals, Mynavi for regional, en for retention-sensitive junior — and buy a higher tier on that one.
- Make Indeed and the domestic search engines free infrastructure. Structure your careers page for indexing, feed Indeed, Kyujin Box and Stanby, and let organic carry the baseline. Layer paid on top of a working free foundation, not instead of one.
- If you buy a scout platform, staff it. BizReach and its peers pay back on outreach volume and message quality, not on licence ownership. A minimum viable programme is one person, named, with weekly outreach targets and templates in Japanese written by someone who has actually hired here. Unstaffed scout licences are the most common wasted line item in a Japan recruiting budget.
- Shift risk where you can. Green's success-fee model and Indeed's cost-per-click both transfer risk off you. Use them where they fit before committing to flat-rate four-week windows.
- Fix the job ad before you fix the channel. The Japanese posting is a different genre: concrete daily responsibilities, an explicit total-cash (年収) range, named team size, stated overtime and remote policy, an honest account of what is hard. Vague global job descriptions translated literally are the most common reason a well-chosen board underperforms. This is usually the cheapest fix available and almost nobody does it first.
- Accept that referrals are a channel here, not a perk. Referral hiring (縁故) fills 23.5% of mid-career hires — nearly four times what agencies do. A functioning employee referral programme in Japan outperforms a second job board almost every time, at a fraction of the cost.
- Reserve the agency budget for where the cliff is. 30–35% is the correct price for a role no board will fill, and a terrible price for one that three would. The discipline is knowing which is which before the requisition opens — not after ninety days of silence.
The boards are not the problem. Pointing them at the wrong third of your hiring plan is.
Sources
Entry-route shares are from the Ministry of Health, Labour and Welfare's Employment Trends Survey (雇用動向調査); the full entry-route table is published only in the e-Stat statistical tables and the most recent complete breakdown is FY2022, with the distribution moving only slightly year to year. The 2009 comparison and the analysis of route by employer size are from MHLW's own study of entry routes. Market size and growth are from Yano Research Institute's 2025 human-resources business survey; agency numbers from MHLW business reports. Platform figures are company-reported: doda member report (July 2025), en-japan (mid-2025), Mynavi (September 2025), Rikunabi NEXT, BizReach, Recruit Direct Scout, doda X, Green, Findy, Wantedly, Daijob and CareerCross. Indeed PLUS dates and participating sites are from Recruit and Indeed Japan press releases (January and October 2024). LinkedIn's Japan membership is NapoleonCat, May 2025. The US comparison is iHire's State of Online Recruiting 2025. Posting rates are list prices before agency discount and move frequently.