Why Simulation?
Understanding people must start before measurement begins
Project Brainstorm is a simulation-based market research and consumer intelligence platform designed to help users better understand how people evaluate products, concepts, and decisions. Its role is to help users explore questions, assumptions, tensions, and decision dynamics before committing significant resources to validation, experimentation, or execution.
The idea of using simulation for market research can feel counterintuitive. If you want to understand real people, why not just ask them?
It's a reasonable objection. After all, market research already has established ways of collecting insight directly from consumers through interviews, focus groups, surveys, and panels.
The problem isn't that these methods are wrong. It’s that they become difficult to use when the thing you're trying to understand doesn't yet fully exist, the context is still evolving, and you're not even sure what questions matter yet.
In those situations, the challenge is often less about measurement and more about exploration. You're not trying to validate an answer. You're trying to work out what the right questions are in the first place. That’s where simulation becomes useful.
Project Brainstorm was built around this idea. Not as a replacement for research, but as a way of exploring how people reason before committing time, resources, or brand reputation to a decision.
Three Decades of Watching the Same Mistake
As someone who has spent the better part of three decades moving between financial journalism, venture building, and investing across Asia-Pacific, I've gradually come to believe that understanding people is one of the most important, but most overlooked, parts of making good decisions.
After all, whether you're launching a product, entering a market, making an investment, or designing a campaign, the outcome ultimately depends on how people respond.
Yet time and again I've watched businesses, investors, and institutions make decisions based on assumptions about how people think rather than a genuine understanding of them.
What's always struck me is that this challenge isn't limited to outsiders. Even people who appear culturally adjacent often misread one another. Singaporeans can misunderstand mainland Chinese consumers. Mainland Chinese businesses can misunderstand Hong Kong audiences. Regional companies can struggle to understand neighbouring markets they've operated beside for years.
The result is a pattern that repeats itself over and over again. Products that don't resonate, campaigns that misfire, market entries that stall, and investments built on assumptions that don't hold.
That's ultimately what motivated me to build Project Brainstorm. Not to replace human judgment, but to make it easier to understand the people on the other side of the decisions we're making.
Simulation Is For Exploration, Not Measurement
Most research methods are built around answering questions. A survey assumes you already know what you want to measure. An experiment assumes you know what variable you’re testing. Even a focus group usually starts with a clearly defined concept, product, or campaign.
But many of the decisions that matter most don’t begin in such neat circumstances.
Often, the hardest part isn’t answering a question. It’s working out what the question is.
You’re trying to enter a new market. Or launch a new category. Or position a product differently. You know there’s a decision to make, but you don’t yet know which assumptions matter, what constraints people are operating under, or what might cause them to think differently.
In those situations, the challenge isn’t a shortage of data. It’s a shortage of understanding. That’s why I’ve come to think of simulation as something that sits before measurement. Its job isn’t to validate; its job is to explore.
The Cost of Getting Context Wrong
Unfortunately, this isn’t just a theoretical problem.
When assumptions about people are wrong, the consequences can be surprisingly expensive.
That becomes especially true in Asia-Pacific, where fragmented markets, multiple languages, cultural nuance, and political sensitivities make understanding people considerably harder than many outsiders assume.
We’ve seen this repeatedly over the years. Swatch lost roughly $300 million in market value after a culturally insensitive campaign in China. Dolce & Gabbana's now infamous "chopsticks" advertisements reportedly cost the company around $100 million in China sales. Walmart spent nearly two decades trying to make Japan work before ultimately exiting at a cost of between $1.6 and $2 billion.
And the examples keep coming. As recently as last month (June 2026), Lululemon found itself apologising after staging an event on China’s Great Wall featuring Japanese drummers.
The details differ, but the pattern is remarkably consistent.
Companies often assume they understand how people in a market will interpret a product, message, or decision.
Sometimes they’re right. When they’re not, the cost can be significant.
Why Context Matters More Than Preference
One reason understanding is so difficult is that people don’t make decisions based on fixed preferences.
Economics often talks about consumer preferences as though they're stable, lifelong things people carry around with them. But my own experience has made me sceptical of that idea. I've lived in multiple countries and watched my own preferences shift over time. Things that felt normal or desirable when I lived in the US or Canada sometimes felt strange in Asia. And even within Asia, I found myself thinking differently depending on whether I was in Hong Kong, Shanghai, Beijing, Singapore, or Manila. Sometimes the change was obvious. More often, it happened so gradually I barely noticed it.
That made me wonder how much of what we call "preference" is simply a response to context.
A person's finances, environment, social expectations, cultural norms, daily routines, recent experiences, and exposure to information all shape what feels sensible, desirable, or important.
That's why the same concept can succeed in one market and fail in another. It's also why two people looking at an identical product can arrive at completely different conclusions.
What feels sensible in Tokyo may feel strange in Jakarta. What appears affordable to one group may feel irresponsible to another. What sounds credible to one audience may feel suspicious to someone else.
The important question, then, is often not:
“What do people prefer?”
But:
“Why does that preference exist in the first place?”
That's a harder question to answer. And it's one of the main reasons I became interested in understanding context before trying to understand preference.
How Context Shapes Responses
Once you start looking at decisions this way, another problem emerges.
How do you accurately capture context?
While building Project Brainstorm, I kept running into a surprisingly persistent issue: the answers produced by an AI model were often shaped by the question itself.
Take something as straightforward as:
“Would Japanese consumers be interested in this premium tea brand?”
A language model will happily generate an answer. But small changes to the framing can produce very different reasoning. Emphasize sustainability and one set of arguments emerges. Emphasize price and another appears. Focus on health benefits and the discussion moves again.
The model isn't necessarily wrong. It's just responding to the context you've provided.
The challenge is that most of the relevant context has already been decided before the model starts reasoning.
Real customers encounter products within a much richer environment. They compare them to local alternatives. They evaluate them through the lens of social expectations, economic realities, cultural assumptions, and the information available to them.
The same product can therefore mean very different things to different people. A premium tea brand might be seen as a thoughtful gift in one market, an everyday indulgence in another, and an expensive product with unclear provenance in a third.
Nothing about the product has changed. What has changed is the context surrounding it.
The same logic applies to market fit.
We often talk about market fit as though it's a property of the product itself. But I suspect it’s better understood as a relationship between a product and a particular group of people at a particular moment, within a particular environment.
A product can resonate strongly in one market and struggle in another. It can appeal to one segment while being ignored by an adjacent segment. The product remains the same, but the conditions under which people evaluate it are different.
That’s why Project Brainstorm starts with the market rather than the product. Before introducing the thing being evaluated, it tries to understand the people doing the evaluating. Who exists in this market? How do they live? What constraints shape their decisions? What information are they likely to encounter? What feels normal or familiar to them?
Only after that foundation is established does the system introduce the thing being evaluated. The goal isn’t to simulate preferences. It’s to simulate the conditions under which preferences emerge.
In Closing
If context shapes reasoning this profoundly, then another question naturally follows. Modern AI models can already generate convincing conversations and plausible opinions. Why not simply ask ChatGPT what people think?
When I started thinking about this problem, I considered if that might be sufficient.
I don’t think it is. Understanding why became one of the most important design decisions behind the entire system.
This article was originally published in Substack (https://josephlo.substack.com/p/why-simulation?r=7fr2) on July 10, 2026.