You Can’t Algorithm Your Way to Empathy

Every now and then, a new research tool comes along that promises to change everything. Lately we’re hearing more and more about synthetic data: information that’s created by algorithms to look and behave like real consumer responses, without needing to ask actual people anything at all. It’s not science fiction; it’s already being used to model behaviors, predict trends, and simulate feedback in place of actual conversation.

And sure, on paper, it sounds like a dream. It’s fast. It’s scalable. It doesn’t cancel last-minute or forget to complete a diary entry. Synthetic data can fill in gaps, simulate responses, and help model behavior across different market scenarios.

But here’s the thing: it’s not real. And it’s not enough.

As researchers, the work we do isn’t just about gathering data. It’s about noticing the unsaid. It’s hearing the hesitation in someone’s voice when they talk about price. It’s watching how they move through a store or light up when they describe the first thing they reach for in the morning. It’s the unexpected turn a conversation takes that reveals something no one thought to ask.

Synthetic data can only repeat what it’s been taught—it can’t surprise us with something we never recorded. It’s a mirror of the past, not a window into the emotional, messy, human complexity that drives real decisions.

Kantar cautions that synthetic data is only as good as the real-world data behind it, and if that foundation is flawed, biased, or incomplete, your outputs will be too. Ipsos goes a step further, urging research teams to use synthetic data responsibly and never in place of real consumer voices.

The best insights still come from human stories. And the best decisions still start with understanding the humans behind the behavior.

If you're in a rush to optimize, simulate, or scale your research, we get it. But don’t skip the part where you talk to people. That’s the whole point.