Japan's System Engineering Service sector is one of the largest IT labour markets in Asia by volume — tens of thousands of engineers are dispatched under SES or haken contracts at any given time, working across financial services, manufacturing, government systems, telecommunications, and an expanding range of platform and consumer tech companies. The model is well-established, legally defined under the Roudousha Haken Jigyou framework, and embedded in how many large Japanese enterprises consume engineering talent.
It is also, structurally, quite old. The core dynamics of multi-layer brokerage, undisclosed rates, and arms-length relationships between engineers and end clients were designed for a market where information moved slowly and talent supply was constrained in ways it no longer is. The question the industry is now working through — unevenly, with some resistance — is what the next version of this market actually looks like.
Where the SES model came from
Japan's IT staffing industry expanded rapidly through the 1980s and 1990s alongside the enterprise computing boom. Large system integrators — referred to colloquially as SIer — built complex delivery hierarchies for government and financial sector clients, subcontracting work through multiple layers of smaller agencies. The nimakedori (double-taking) structure, where each layer in the chain extracts margin before passing work down, became normalised.
For much of this period, the opacity was arguably sustainable. Engineers were relatively scarce in certain specialisms, client companies lacked the capability or HR infrastructure to manage contract staff directly, and the agency layer provided genuine services: screening, payroll administration, dispute resolution, and a degree of career continuity for engineers between engagements. The intermediary had a function.
The problem is that the function has partially decoupled from the structure. Many of the services that once justified the intermediary layer are now available through platforms, legal-standard contracts, and direct administrative relationships. What remains is often the margin — extracted from the gap between what clients pay and what engineers receive — without the full service stack that once backed it.
The generational shift in expectations
Engineers entering the Japanese IT market now are not the same cohort that the traditional SES model was designed around. Many have visibility into global talent markets through platforms and online communities. They have exposure to compensation discussions that are increasingly open — both within Japan and through the growing presence of foreign tech employers hiring locally. They have seen how freelance and direct-hire markets in other sectors work.
The result is a growing expectation of information parity. Younger engineers, and a significant portion of mid-career engineers, are no longer willing to accept the premise that the billing rate to the client is none of their business. They regard that information as directly relevant to evaluating whether the working arrangement is fair. And they are correct — it is.
This shift shows up in how engineers talk about their work experiences. Discussions in engineering communities about pinpane — the margin taken by SES agencies — have become more common and more specific. Engineers share benchmarks, compare notes on which agency structures are more or less transparent, and increasingly choose platforms and arrangements that give them more visibility. The cultural convention around not discussing compensation is eroding at the edges of the market most relevant to younger and higher-skilled talent.
What direct matching means structurally
Direct matching — the model Rezon is building toward — does not eliminate intermediation. It changes what intermediation means. In a direct-match model, the platform's role is to facilitate a genuine meeting between an engineer who understands the real terms of an engagement and a client who has written a brief specific enough to attract the right candidate. The platform earns a disclosed fee for that facilitation. The engineer knows what the client is paying. The client knows what the engineer is receiving. Neither party is operating in an information vacuum.
This is not primarily a technology problem. The data infrastructure to run such a matching process is straightforward. The harder work is on the structural and commercial side: getting clients to provide real project briefs rather than vague proxy descriptions, building the trust that enables disclosed rates to be a feature rather than a liability, and demonstrating to both engineers and clients that the transparent model produces better outcomes over time — not just fairer ones.
Early evidence from this model, observed in adjacent markets where rate transparency has been introduced — notably in some overseas engineering staffing platforms, and in Japan's growing freelance tech market — suggests that matching quality and retention improve meaningfully when both parties enter the relationship with accurate information. Engineers who accept projects knowing the real terms stay longer and perform better. Clients who attract engineers through genuine project specificity rather than opaque placement tend to get candidates who are intrinsically motivated by the work, not just the paycheck.
Where the model faces real friction
We should be honest about the genuine resistance the shift toward transparency encounters. It is not only structural inertia.
For established SES agencies, disclosed rates create competitive pressure that did not exist before. If engineers can see that a particular agency is extracting a 30% margin while a competing platform discloses a 10% fee, the agency has a problem it did not have when that comparison was impossible. The incentive to resist transparency is real and financially grounded, not merely habitual.
For some client companies, particularly those operating within large procurement frameworks or with longstanding relationships with specific SIer partners, changing how they access contract engineering talent is genuinely difficult. Internal approval processes, preferred-vendor lists, and legal compliance requirements all create friction. Even clients who would benefit from a more transparent model may find it operationally hard to access one.
And for engineers who have worked within the traditional system for many years, the shift is not automatically easier. Some have built their careers around agency relationships that work reasonably well for them. Transparency is not costless — it changes the negotiation dynamics and requires engineers to know and advocate for their own market value, which is a skill that not everyone has had reason to develop.
What this means for engineers working in Japan today
The shift from black-box to direct-match is happening, but it is happening unevenly and over a longer timeline than the more optimistic accounts suggest. For engineers navigating the market now, that means the old and new models coexist, and the right approach depends on where you are in your career and what you are optimising for.
For engineers with in-demand skills and the leverage to be selective — cloud-native infrastructure, fintech-grade backend engineering, data engineering pipelines — the market has already moved enough that demanding transparency on rates is not a radical act. The platforms and direct-hire arrangements that disclose this information are accessible, and the risk of missing out by asking is lower than it has ever been.
For engineers earlier in their careers, or in specialisms where the market is less tight, the calculus is different. Building a track record through traditional SES is not wrong. The advice is to stay conscious of the information asymmetry, calibrate your rate floor against market data rather than agency quotes, and treat the opacity as a known limitation to manage rather than a given to accept.
The direction of travel in the market is clearer than it has been at any point in the past decade. The generation of engineers entering the market now, combined with the platforms being built to serve them, will continue to push toward a model where what the client pays and what the engineer receives are known quantities to both parties. How fast that arrives is still being worked out. But the destination is not in doubt.
