Growing up ethnic Chinese in Indonesia, I was blessed with several gifted primary school teachers who smoothed my path toward math and data science. I was the odd one who shined to equations like Euler’s Identity. From preschool onward, the best time of day was when my father arrived home with his laptop, and a strong interest in video games (yes, I was one of those kids) got me early into all coding and tinkering that a computer can endure.
Landing eventually at UC Berkeley, I pursued data science and was pulled into projects ranging from geolocation to entrepreneurial process analysis. These interests brought me to Global Key Advisors, where I helped initially to get a startup success/failure study off the ground. This effort has involved structuring data based upon materials produced by startups going through annual boot camp pitch competitions organized by the Cleantech Open, a nonprofit which is the world’s oldest and largest venture capital accelerator for companies focused on sustainability. This is distinct from Global Key’s core business, which is quantitative financial advisory work around publicly traded securities. Global Key does not invest in or make recommendations regarding privately held companies, but CEO Josh Schein and I were keen to run the study after coming to appreciate that quantitative finance can bring valuable perspectives not normally present in a startup and incubator/accelerator environment.
My personal startup experience began while watching my mother start several businesses, and I later took a UC Berkeley course called DATA-X where groups formed startups based on class content. I believe that nearly anyone can connect startups and private companies to personal life experiences, and I suspect that VC’s and investment bankers do this to the point of accumulating scar tissue on their souls. What is almost entirely lacking, however, is empirical data showing what makes startups succeed or fail. The most competent people best positioned to collect such data (e.g., entrepreneurs, VC’s, and investment bankers) lack the time, incentives, or organized means of doing so. Some of the most highly regarded statistical literature on startup success/failure involves datasets as small as 200 founders of 100 companies. Academic practitioners lack better options.
Our startup success/failure project addresses this shortfall. The project began with sixteen seasoned professionals with diverse expertise ranging from legal and intellectual property strategy to financial positioning, technology and market validation, branding, customer discovery, and venture capital. This group met for three afternoons over the course of several months and developed a broad VC-style checklist of 283 due diligence questions designed to be as objectively answerable as possible. Along with more than 200 college interns over time (this is a large-scale project), I joined initially to score startups against the checklist and ascertain the extent to which these startups later succeeded or failed. The project strongly clicked with me, and eventually I was promoted to co-running it.
Although many interns wanted to ‘analyze’ data from day one, the most valuable work at this stage involves building data to analyze. As of today, June 11, 2025, we have scored more than 800 companies twice, meaning at least two people have separately scored each of these companies. This will increase to more than 900 companies by the end of Summer 2025. Importantly, our data is now attracting significant academic engagement.
My involvement in this project quickly pulled me into the core work of Global Key where I have dived headfirst into ‘smart money’ research in a full-time permanent role. Most of what Global Key studies is non-overlapping with respect to the startup success/failure project, although interesting links arise. As one example, Global Key recently published a study showing that the frequency with which an insider trades correlates negatively to performance. Might the same be true of VC’s and private equity professionals making investment decisions on behalf of their funds? Is a classic VC fund (constant influx of investment ideas from outside) superior to the venture studio model (same people responsible for all idea generation)? Apropos to Global Key clients, should individual investors curtail personal trading except in rare circumstances?
I will have more to say as I get further into these challenges.