The statistician who was asked to count the agents

In the spring of 1996 I was asked to count the agents of a provincial life insurer in Anhui, and discovered that nobody could agree on what an agent was. I had joined the Anhui branch of Dongyuan Life in August 1994, straight from a statistics degree at Fenghe University, to prepare premium and claims returns in the planning department. When the branch launched an individual agency force, someone had to report its size to head office each month. Was an agent the person who had attended an induction, the person who held a contract or the person who had sold a policy in the last ninety days? The three numbers differed by a factor of four. The branch reported the largest, and I learnt early how much an organisation can want a figure to be big.

By 1999 the agency in Anhui had grown to almost 20,000 people, and the branch asked me to run agent training. I had no sales experience, so I spent my first months going on appointments with agents in Hefei, Wuhu and the villages around Lu'an, carrying their brochures and listening to what customers actually asked. The induction course at the time taught product features and recruitment. It taught almost nothing about explaining a participating policy's dividends honestly, which was what customers most often misunderstood. I rewrote it with two experienced agents and a claims manager, and added a final session in which new agents heard from policyholders whose claims had been declined. That course is the first piece of work I am still proud of.

In April 2004 I moved to Chengdu as director of human resources for the Sichuan branch, with 3,600 employees and more than 50,000 agents. It was my first general HR role, and I had to learn pay administration, labour contracts, social insurance and the workings of the employee representative congress quickly and in public. The work that shaped me most came in May 2008, after the earthquake in the north-west of the province. Several of our offices were damaged, colleagues had lost relatives and the branch faced an exceptional volume of claims. We moved staff from other cities into the affected counties for months at a time. I learnt then that a rotation plan is also a family plan, a lesson I have had to learn again since.

From January 2010 I was deputy general manager of the individual agency department at Dongyuan's head office in Beijing, responsible for recruitment standards, agent training and the development of agency managers across 31 provincial branches. These were the years of fastest growth in the agency model. I helped build a manager-development programme that served the company well, and I also helped run recruitment campaigns that I would not run today. Headcount was the measure everybody watched, including me. I completed a part-time MBA at Xiyuan Business School in 2011 with a dissertation on why agents left in their first year. The answer was mostly income, and the changes I recommended were considered and set aside.

In June 2015 I became head of human resources at Brennmark Xinghe Life, a joint venture in Guangzhou between a European insurer and a Chinese state-owned industrial group. Working for two shareholders with different instincts taught me to put a decision, its alternatives and the strongest objection on one page, because each side read a paper for what the other had left out. The joint venture's agency was small and productive, built on selection rather than volume, and it showed me that a different model could work in China. I also prepared papers for a board remuneration committee for the first time.

I joined Songbai in March 2020. By the end of that year our agency had fallen to 164,000 from 231,000, as agencies were falling across the industry, and the plan for 2022 assumed we would recruit our way back. I argued that we should not, and we did not. Since then my work has been to make Songbai's people figures describe the work honestly: an agency counted by the agents who earn a living from it, a performance system that does not punish accurate reporting, an actuarial bench built rather than bought, and provincial leadership that changes hands before it becomes personal.

I work best with colleagues who will disagree on paper. Every proposal I take to our president carries the strongest objection I have received, attributed with permission. I ask for cohort data rather than totals, recordings rather than completion rates and audit files rather than assurances. Colleagues would say that I am slow to call a programme a success, that I visit branches more often than a head-office executive needs to, and that I still read agents' income statements the way I first read them in Anhui.

Open to Board appointments · Advisory roles · Speaking invitations

Languages / Mandarin Chinese · English