people&proof.
Research practice

Good research starts with a decision, not a survey.

“We need to understand our customers better” is a reasonable ambition. It becomes a useful research brief when you identify what that understanding will change.

Good research starts with a decision, not a survey.

Name the choice you face.

A team deciding which market to enter needs different evidence from a team rewriting an onboarding flow. Start by writing the decision in one sentence. Include who will make it and when. This simple step helps prevent a study from collecting interesting answers that arrive too late or cannot influence the work.

List the assumptions underneath it.

Every proposed action carries a set of beliefs: that people recognise a problem, that they want the benefit, that they trust the promise, or that the experience is easy enough. Identify the assumptions that would change your direction if they proved wrong. Those are candidates for research.

Choose the people before the questions.

The easiest respondents to reach may not be the people whose decisions you need to understand. Current customers can explain their experience, but they cannot fully represent prospects who never converted or customers who have left. Define the relevant audience and be explicit about who the findings will not represent.

Let the question choose the method.

Interviews can help uncover how people describe a decision and which factors they consider. A well-designed survey can examine how responses vary across a defined sample. Usability observation can show where a task becomes difficult. No method should be the automatic answer to every brief.

Agree on how the findings will be used.

Before fieldwork, ask what evidence would support proceeding, changing the idea, or stopping it. This does not mean forcing a study into a binary verdict. It means making room for evidence to influence the plan, rather than asking researchers to validate a conclusion already reached.

Finish with action and uncertainty.

A decision-ready report explains what the evidence supports, what remains uncertain, and what to do next. It connects recommendations to findings and avoids presenting every observation as equally important. The measure of a useful study is the quality of the decision it helps a team make.

← Back to all insights

What would you like to understand?

Bring us the question. We’ll help you find the right way to answer it.

Start a conversation