Chief Data Officer Singapore / Southeast Asia
Learning the shop before learning the model

In 1997 I spent most of my first month with a clipboard in provision shops around Johor Bahru, copying shelf prices for a retail-audit panel. In one shop the label on a carton of drinks said one thing and the owner charged another, depending on whether the customer paid that day or ran a tab. My form had a box for the label. It had no box for the arrangement. I have thought about that form for nearly thirty years, because much of the data I now manage still asks for the label when the business runs on the arrangement.
I grew up in Singapore and completed a statistics degree at Kenarimas University in May 1997. Pelangiscan Retail Audit hired me two months later. I began as a field auditor and became a panel analyst, checking the returns from 1,150 provision shops and minimarts in Singapore and Johor. The work taught me sampling, and also humility: a shop that closed for a week of mourning or changed hands could distort a district's figures more than any formula.
In 2002 I moved to Jakarta to run sales information for PT Sumberlaras Distribusi, a family-owned distributor supplying packaged food and personal care to warung and small grocers in Greater Jakarta and West Java. Its salespeople carried invoice books, and much of the trade ran on cash and short credit. My team of fourteen keyed, checked and reported those books for nine branches. I learnt that the most useful report was often a list of anomalies taken back to the branch manager, not a chart for head office, and that a salesperson's explanation was evidence rather than an excuse.
Telagawangi recruited me to Klang in June 2008 as commercial analytics manager. The company made canned soya, barley and grass-jelly drinks and was beginning to deliver 19-litre water to homes. Van orders were still written in books and keyed at depots overnight. I built the first depot-level sales reporting that the sales director trusted, mainly by reconciling it every week with finance's cash records until the differences were small and explained.
From 2014 I led route-to-market analytics. We replaced the invoice books with handheld devices across 24 depots during 2015, and I made a mistake I still describe to new colleagues. We designed the screens at head office, tested them in an air-conditioned room and filled them with fields that salespeople skipped in a hurry. Completion of optional fields fell below a third within two months. We redesigned the screens after riding with crews, removing nine fields and making three mandatory. Since then I have spent at least one day a month in a van.
I became head of data and analytics, reporting to the chief financial officer, in January 2019. The Movement Control Order in March 2020 shut many of our outlets overnight while household water orders rose sharply. Our forecasts were built for neither. For several weeks the most reliable signal came from depot managers' daily calls, and we said so to the board rather than present a model we did not believe.
In July 2022 I was appointed chief data officer, reporting to the group managing director. My remit covers data engineering, forecasting, data quality, field systems and the data protection office, with 74 colleagues and an annual budget of RM31 million. Our data protection officer has direct access to the chair of the Audit and Risk Management Committee, and I present to that committee each quarter. I attend board discussions on data matters as an executive. I have never held a directorship.
The cases on this site are deliberately unglamorous. A shop-level forecasting module was declined in favour of route-day forecasts. Salespeople were asked to code every bonus carton, which slowed them down. Distributors were required to give up spreadsheets of household customers, and two left us. In each, the data improved only because someone in the field did something differently, and I try to give them both the credit and the explanation.
I lead a team that includes engineers who have never sold a carton and former salespeople who have never written code. I ask both groups to sign off the same definitions. Inside the company I am known, not always fondly, for asking to see the handheld screen before I look at the dashboard built from it.
I am now preparing for a first independent seat. I would bring an understanding of fragmented, cash-heavy distribution and of the personal records that consumer businesses accumulate through partners. I would need to learn a great deal about the duties of a non-executive director, and I would rather join a committee under an experienced chair than lead one.
Open to Board appointments
Languages / English · Malay · Bahasa Indonesia
A leadership conviction
A shopkeeper deciding how many cartons to take, and whether to pay in cash, creates most of the data my company relies on. I design our records around that moment rather than around the report that will summarise it later. When the counter and the dashboard disagree, I begin by assuming the counter is closer to the truth.
- Group revenue
- RM868m
- Year to December 2025; canned drinks, chilled soya and water delivery
- Outlets reached each month
- 61,000
- 23,500 by our own vans; the rest through 38 distributors
- Data colleagues
- 74
- Engineering, forecasting, data quality, field systems and data protection
Education & credentials
- Bachelor of science in statistics
- Kenarimas University · 1997
Perspectives from colleagues
Nisha cost my salespeople two calls a day and said so in the paper instead of hiding it. When her first report pointed at individuals, she accepted my objection and redrew it by who had approved each scheme.
We had kept customer lists on drivers' phones for years. Her team paid for the devices, trained our crews in Malay and Tamil, and checked the deletion with us in our own office rather than sending a form.
In Jakarta she brought me lists of problems, never accusations. Every anomaly came with the salesperson's own explanation, which is why our branch managers came to trust her reports.