17/09/2026
There is a common misconception that synthetic data means “made-up answers.”
It doesn’t.
At its best, synthetic data is not about replacing real people with fiction. It is about using real population data to create statistical representations of real segments — so researchers and businesses can explore questions at a scale, speed, and cost that conventional panels often cannot reach.
That matters for two reasons.
For research, synthetic populations can help academics test assumptions, repeat studies, model different scenarios, and understand how insights may shift across segments. Used carefully, synthetic data becomes a research instrument — not just a shortcut.
For business, it can significantly reduce the cost and time required for market research, while supporting faster, evidence-based decision-making. It allows teams to explore consumer preferences, test ideas, and identify patterns before committing major resources.
But synthetic data also needs scrutiny.
How is it built?
What real data informs it?
Where is it reliable — and where is it not?
How is it different from generic AI output, “farmed data,” or a robot pretending to be a respondent?
Synthetic data is not a magic answer. But used responsibly, it may become one of the most important tools for connecting academic rigour with real-world business insight.
Preferences AI, a leading synthetic data company for APAC, is collaborating with Prof. Eunyoung SONG at City U Department of Marketing, to answer those questions and more.