“We need a survey” is a request for a method. It leaves open what the business needs to learn, why it matters and what anyone will do with the answer.
A better brief starts with a decision. The research design follows from what you need to know to make it.
Set out the objective
Separate the commercial goal from the research question. “Improve profitability” is a business objective; “understand why customers reject our premium range” is something research might help answer.
Make the decision explicit. Will you change the product, alter the price, focus on a different audience or stop investing? The answer determines what evidence is useful and how strong it needs to be.
Name the person or team who will act, as well as the constraints. A recommendation that cannot be funded, delivered or agreed may not help the business move forward.
Write down your hypotheses
A hypothesis is a proposed explanation you can investigate. It should make your current assumptions visible, not prescribe the answer you want the research to produce.
For an illustrative premium-range brief, competing explanations might include poor visibility, an unsuitable format, a price-value mismatch or weak demand for the proposition itself.
Ask what evidence would support each explanation and what would challenge it. Include the possibility that more than one explanation is operating at once.
Choose methods for the gaps
Check existing evidence first. Distribution, availability, transaction data and earlier research may answer part of the question before you commission anything new.
Observation can examine what happens in context. Interviews can explore experiences and possible explanations. Quantitative work can estimate how common a pattern is within the population the design supports. An experiment can test the effect of a particular change when the setup permits a credible comparison.
These methods are not interchangeable. Choose the combination that answers your questions, with a sample that fits the decision and an honest account of its limitations.
Agree what a useful answer looks like
Clarify the expected output before the programme starts. That might be a comparison of options, an explanation of barriers, a set of priorities or a recommendation with supporting evidence.
Also agree what the work cannot establish. A preference survey is not automatically a forecast of purchasing, and research about customers’ reactions is not a financial business case on its own.
Bring the team into the work at useful points so assumptions can be challenged while the research is still taking shape.
Put the brief on one page
Use four headings:
| Heading | What to establish |
|---|---|
| Objectives | The business decision and the questions it depends on |
| Hypotheses | Current explanations and evidence that could challenge them |
| Methods | The evidence needed, who to learn from and how |
| Expectations | Useful outputs, limitations, timing and who will act |
This structure is a starting point for discussion. It should make the brief clearer, not become another form completed without thought.
Try it with your team
Take a live request for research and complete those four headings. If you cannot name a decision or an owner, resolve that before debating sample size or the number of slides in the debrief.
Previous: Ask questions people can actually answer. Next: Make insight earn its commercial keep.
Have a perspective or a related question?
Discuss this piece