Market research fails most often by confirming what the team already believes

If you walk in hoping the data will bless an existing plan, you will always find a way to read it as approval. That is not research at all, it is a search for permission wearing a lab coat.

Useful research is deliberately designed so it can prove you wrong. It asks questions whose answers could realistically kill the idea, not only flatter the version you already wanted to ship.

Small, honest samples beat large, biased ones, because a leading question scales its own distortion. Ten straight answers teach you more than a thousand nudged ones, and cost far less to collect and act on.

Takeaway: before collecting anything, write down the specific result that would make you cancel the project. Then go looking for it.