Six appointments, one question: who carries the risk?

Mandate

Reduce dependence on one currency and one client country, build a domestic business that earns its margin, and prepare the company for a listing.

Outcomes

• Raised revenue from RMB 3.05 billion in 2019 to RMB 4.36 billion in 2025, with Japan falling from 68 to 47 per cent of the total.
• Restored Japan gross margin from 23 per cent in the second half of 2022 to 28 per cent in 2025 through a shared currency clause.
• Opened the Ho Chi Minh City centre in 2023; it employed 1,050 people at the end of 2025.
• Grew domestic revenue to RMB 1.78 billion while holding domestic gross margin at 24 per cent.
• Listed the company in Hong Kong in November 2025, raising HK$1.08 billion.

Decisions in context

The board appointed me president and chief executive in April 2019, when Yanxiao had revenue of RMB 3.05 billion, 8,900 employees and two-thirds of its business in Japan. Qiao Dongsheng, who founded the company in 1995, stayed on as chairman. My first year was taken up by the conversion to a joint-stock company and then by 2020, when we moved most of our engineers to remote work under client rules that had never contemplated it. From 2021 the work changed character: the yen's fall, substitution programmes at home and the compliance demands of foreign clients each required decisions the company had not faced before. I lead an executive committee of eight, spend a week each month in Japan, and have been a director since the conversion, an executive director of the listed company since November 2025. By the end of 2025 we employed 11,800 people, and Japan provided 47 per cent of revenue.

Reason for the move

Current role. Not seeking an executive move; considering one independent directorship with the Yanxiao board's consent.

Mandate

Turn a loss-making domestic unit into a business that could stand beside the Japan business.

Outcomes

• Grew domestic revenue from RMB 290 million in 2015 to RMB 740 million in 2018.
• Opened the Chengdu centre in 2016; it reached 900 engineers by 2018.
• Moved the domestic unit from a 4 per cent operating loss to a 9 per cent operating margin.
• Signed the company's first two contracts to migrate state-enterprise systems to domestic databases, in 2018.

Decisions in context

Qiao Dongsheng asked me to return in January 2016 to build a business with Chinese clients, which then had RMB 290 million of revenue and lost money on most contracts. Chinese enterprises bought through tenders, wanted delivery in weeks rather than quarters and accepted more rework than any Japanese client would. I separated the domestic unit's sales from its delivery, stopped bidding for one-off government websites and concentrated on manufacturers, ports and utilities in the north-east and the south-west. We opened a centre in Chengdu in 2016, close to clients in Sichuan and Chongqing and to universities producing engineers who had not been drawn into the Japan business. By 2018 domestic revenue was RMB 740 million and the unit earned a 9 per cent operating margin. That year we also signed our first two contracts to migrate state-enterprise systems to domestic databases, the start of a much larger business.

Reason for the move

Appointed president and chief executive by the board.

Mandate

Decide which of the parent's work belonged in Japan, in its own centre or with outside vendors, and raise the centre's quality to the parent's domestic standard.

Outcomes

• Reduced defects found by the parent's teams by 46 per cent over three years.
• Introduced a vendor scorecard covering rework, speed of answers to specification questions and staff turnover as well as rate.
• Moved 180 roles of design work from Tokyo to Dalian with the agreement of the parent's project managers.

Decisions in context

In July 2012 Kurobane Systems, then one of Yanxiao's largest customers, asked me to help run its own development centre in Dalian, and I became its general manager in 2014. Leaving a vendor for its client was unusual, and Qiao Dongsheng agreed only after Kurobane undertook not to recruit from Yanxiao for two years. I asked in writing to be kept out of every decision on Yanxiao's share of the parent's vendor work, which the Tokyo division took directly. The centre employed 1,600 people and recommended where about ¥6 billion a year of further work should be placed. I learnt what clients actually compare: rate, defect counts and little else. I added rework hours, the speed of answers to specification questions and staff turnover to the vendor scorecard, and defects found by the parent's teams fell by 46 per cent. I also learnt why an integrator rarely tells a vendor how it decides what to send offshore.

Reason for the move

Invited by Yanxiao's founder to return and build its business with Chinese clients.

Mandate

Grow the financial-services delivery unit without losing quality as Japanese banks' and insurers' projects moved more work offshore.

Outcomes

• Grew the unit from 420 to 1,100 engineers and its revenue to RMB 380 million.
• Roughly halved client-found defects per thousand lines between 2008 and 2011.
• Built a training programme that qualified 140 bridge engineers in five years.

Decisions in context

I returned from Tokyo in April 2007 to run the unit that served Japanese integrators' projects for banks and insurers. It had 420 engineers and a reputation for good code and slow answers. Japan's financial institutions were modernising older systems, and integrators were sending more design and testing offshore. I reorganised the unit around client accounts rather than technologies, so that each account had a bridge team that knew its rules, and built a programme that trained 140 engineers as bridge engineers over five years. Client-found defects per thousand lines fell by about half between 2008 and 2011, and the unit grew to 1,100 engineers and RMB 380 million of revenue. The financial crisis slowed Japanese spending for a year, and I learnt to keep teams together through a downturn by moving them between accounts rather than releasing them.

Reason for the move

Invited by Kurobane Systems to run its Dalian centre, a chance to see offshore work from the client's side.

Mandate

Turn Japanese clients' requirements into work Dalian could deliver, and carry the clients' concerns back.

Outcomes

• Managed eleven projects for integrators serving regional banks and insurers.
• Reduced specification questions left open for more than three days from 38 to 9 per cent on my largest account.
• Wrote the bridge-engineering handbook that Yanxiao still uses in revised form.

Decisions in context

Yanxiao Japan opened in Tokyo in October 2000 with six people, and I was one of them. As a bridge engineer I attended requirements meetings, wrote Japanese questions for Dalian and Chinese answers for Tokyo, and spent many late evenings explaining why a design document meant something other than what it said. From 2003 I managed projects myself, eleven in all, most of them for integrators building systems for regional banks and insurers, with offshore teams of up to 160 engineers. On my largest account we reduced specification questions left open for more than three days from 38 to 9 per cent. I also wrote the handbook on bridge engineering that we still use in revised form. Six and a half years in Tokyo taught me how Japanese clients weigh a supplier: by whether it raises problems early and keeps its word afterwards.

Reason for the move

Asked to return to Dalian to run the financial-services delivery unit.

Mandate

Write and test programs to Japanese design documents for integrators' banking and insurance projects.

Outcomes

• Wrote and tested loan-servicing and policy-administration modules for four Japanese projects.
• Led a team of eight from 1999.
• Introduced a checklist of each client's coding conventions that cut review returns on the team's work.

Decisions in context

I joined Yanxiao in August 1997, a month after graduating, as one of 140 programmers in a company that did nothing but write and test code for Japanese integrators. The work arrived as Japanese design documents, and much of my first year went on learning the vocabulary of loan servicing and insurance administration in a language I had studied only as an elective. The 600 red-marked pages that came back from Tokyo in 1999 changed how I worked: I wrote a checklist of each client's coding conventions and made the team review against it before anything was sent. I led a team of eight from 1999 and studied Japanese in the evenings, passing the highest level of the proficiency test that December. In 2000 the company chose me as one of six people to open its Tokyo office.

Reason for the move

Selected for the new Japanese subsidiary as a bridge engineer.