Testing whether the evidence supports the decision
My contribution to a board
I would bring a practical understanding of how operational data becomes a board matter. Most boards see data through a security incident, a regulatory change or a technology budget. I have spent twenty-five years with records that people act on: gas alarms in a mine, fuel and emissions figures at a power plant, customer consumption in a spot market and charging sessions that reveal where drivers go. That experience lets me ask management four plain questions about any important dataset. Who owns it, what is it for, how long is it kept, and who has looked at it recently? I can also read a forecast and say what its range means for a financial exposure. My limits are equally clear. I am not a lawyer, a cybersecurity engineer or an accountant, and I have never been a director. On an audit or risk committee I would complement those disciplines rather than duplicate them, and I would expect to learn from colleagues whose training differs from mine.
Judgement in practice
01
Separating the value of a data asset from the right to sell it
The vehicle maker's offer was commercially attractive and the proposal arrived with a ready answer: remove the phone numbers and the records become anonymous. The board-level question was different. What would drivers reasonably expect, and what would the law require before their whereabouts left our company? On another board I would ask management the same thing about any data partnership: which personal information is involved, what consent exists for the new purpose, and what the company would say if the arrangement were described publicly. I would also ask what the company holds that nobody has decided to keep. In our case the offer exposed 260 million identifiable records retained by default. A director does not need to design the retention schedule, but should notice when a commercial opportunity is the first moment anyone has asked how long data should exist.
02
Recognising when data shows a pattern but not a conclusion
Before the Spring Festival peak, our linked records showed repeated charger alarms closed without cause, but they did not prove that any connector was faulty. The safety manager and the operations director drew opposite conclusions from the same evidence. My contribution was to say exactly what the data supported and where it was silent, including 31 connectors with telemetry too incomplete to judge. The general manager then made a proportionate decision and recorded the dissent. As a director I would look for that discipline in management papers. When a paper presents safety or operational data, I would ask what confidence it carries, what was not observed and who disagreed. I would not want a board to treat a dashboard as a conclusion, nor to dismiss a pattern because it has not yet produced an incident.
My first year on a board
In my first months I would ask to see the company's data the way its operations see it, not through a summary slide. I would request a list of its most consequential datasets, with their owners, purposes and retention, and compare it with what the business actually does each day. A site visit to a control room, a customer service centre or a maintenance depot would teach me more than an induction pack. During the first committee cycle I would mostly listen, and ask the chair privately where my questions helped and where they distracted. By the end of the year I would want to understand how management reports incidents involving personal information or safety data, whether near misses reach the committee, and how internal audit tests data controls. I would also want to know which records nobody has examined recently, because in my experience that list usually contains the next problem.
Conflicts and their management
My potential conflicts follow from where Hanxiyuan sits in the energy system. We buy electricity in the provincial market, sell it to 1,860 commercial customers, buy chargers from three manufacturers, use a maintenance contractor and sell aggregated charging data to a vehicle maker. Any of those parties, or their competitors, could appear in a prospective company's supply chain or customer base. Before accepting a role I would list those relationships for the nominating chair and our own legal department, and decline if a material one could not be managed. If a matter touching Hanxiyuan arose once appointed, I would declare it and leave the discussion. I would never carry a customer's consumption data, trading positions or a driver's information between companies. Time is a second constraint. Twenty days must cover preparation and committee work, and a spot-market crisis or a safety incident at Hanxiyuan would take priority, so I would say so before appointment.
My independence statement
I am a full-time executive of a state-owned group's subsidiary, so my independence has to be tested against that group's relationships, not assumed. Our customers, charger suppliers, maintenance contractors and the vehicle maker that buys our aggregated data would all need checking against any prospective company. I would set those relationships out in writing and accept the company's own assessment of whether I qualify.
Board and committee exposure
Presented to the Board or a committee
Hanxiyuan Energy Services
Board and its risk committee
March 2022–present
Presents quarterly on data security, personal information, the cost of forecast error and safety-data findings as chief data officer; no director appointment.
Reported to a committee
Taolingshan Thermal Power
Plant safety and environmental committee
January 2012–April 2016
Reported monthly reconciliations of emissions monitoring, fuel data and carbon-reporting findings as fuel and emissions data supervisor.